Cerebral hyperperfusion syndrome: A cause of neurologic dysfunction after carotid endarterectomy

Transcription

1 Cerebral hyperperfusion syndrome: A cause of neurologic dysfunction after carotid endarterectomy Martha M. Reigel, M.D., Larry H. Hollier, M.D., Thoralf M. Sundt, Jr., M.D., David G. Piepgras, M.D., Frank W. Sharbrough, M.D., and Kenneth J. Cherry, M.D., Rochester, Minn. Neurologic deficits evident when patients initially awaken from surgery are generally due to intraoperative embolization or inadequate cerebral protection in patients with marginal cerebral perfusion; neurologic deficits occurring in the immediate postoperative period are usually related to acute carotid occlusion or embolization. However, in a small subset of patients, transient postoperative neurologic dysfunction seems to be related to a syndrome of cerebral hyperperfusion rather than a lack of adequate cerebral blood flowi This study describes the courses of 10 patients with classic findings of cerebral hyperperfusion syndrome. Typically, this syndrome occurred in patients with longstanding severe chronic cerebral ischemia and occurred after correction of a very high-grade carotid stenosis. Intraoperatively, there was often a dramatic increase in xenon-labeled cerebral blood flows, with postocclusion flows sometimes attaining three to four times baseline levels. Postoperatively, the patients initially did well. However, over the next several days, many of them began to complain of unilateral headache on the operated side and subsequently had seizures. Electroencephalography obtained during this period uniformly revealed periodic lateraliz/ng epileptiform discharges on the side of the brain ipsilateral to the endarterectomy. Although neurologie dysfunction fully resolved in all of the patients in this group, it is possible that intracerebral hemorrhage may occur in some patients with hyperperfusion syndrome. The pathophysiology of this syndrome is believed to be related to preoperative loss of cerebral autoregulatory mechanisms caused by chronic cerebral ischemia. (J VAse SUgG 1987;5: ) Neurologic complications associated with carotid endarterectomy are of varied origins. Those complications occurring during operation are usually related to either embolization or inadequate protection in patients with marginal cerebral perfusion. Most complications that occur hours or days after operation are ischemic in nature, caused by either carotid occlusion or embofization. However, in a small subset of patients, postoperative neurologic dysfunction seems to be related to a syndrome of cerebral hyperpeffusion rather than lack of adequate cerebral blood flow. Sequelae of cerebral hyperperfusion during the postoperative period include unilateral headache (sometimes with migraine variants), seizures, and From the Departments of Vascular Surgery, Neurosurgery, and Neurology, Mayo Clinic and Mayo Foundation. Presented at the Tenth Annual Meeting of the Midwestern Vascular Surgical Society, Indianapolis, Ind., Sept , Reprint requests: Larry H. Hollier, M.D., Department of Vascular Surgery, Mayo Clinic, 200 First Street, SW, Rochester, MN transient neurologic deficits. In rare cases, this constellation of symptoms progresses to intracerebral hemorrhage with permanent neurologic deficit or death. METHODS In an effort to characterize this syndrome more clearly, records of 2439 patients who underwent carotid endarterectomy at the Mayo Clinic between 1972 and 1985 were reviewed. All patients underwent perioperative electroencephalography (EEG). Thirty-two patients, or 1.3% of the total, had periodic lateralizing epileptiform discharges (PLEDs) on EEG during the postoperative period. PLEDs, as shown in Fig. 1, are periodic discharges on one side of the brain and denote an acute, localized cerebral focus of irritability. PLEDs are usually transient and resolve fully; they are a sensitive but nonspecific objective indicator for the constellation of symptoms that comprises hyperperfusion syndrome after carotid endarterectomy. Five patients who had emergent endarterectomy

2 Volume 5 Number 4 April 1987 Hyperperfusion syndrome 629 Fpi- F 7 F 7-T 3 T3-T 5 Age: 54 yrs PLEDs T 5-0 i Fp2- F 8 F 8 -T 4 T4-T6 T6-O 2 (i), T 5o,v t sec Age: 54 yrs FpI-F 7 _..... _.. - _. -. ~ _ _.... F 7- F 8 F8-T 4 - :.. _.,,......_. - T4_O 2 -. _ -._ Fp2- F _ F s - F 4 F4-T 6 T 6-0 Tso,v 'Isee Fig. 1. Periodic lateralizing epileptiform discharges (PLEDs). A, EEG tracings demonstrate PLEDs in left temporal area (upper four tracings) and right temporal area (lower four tracings). Lowest tracing on the right shows crossover effect from the left hemisphere. B, EEG tracings in the same patient 5 days later show resolution of PLEDs. for progressive stroke, either on original presentation or after cerebral angiography, were not considered. One patient with postoperative seizure activity shown by EEG that originated from an anatomically distant vascular distribution (vertebral artery distribution in the absence of a patent posterior communicating artery) was also excluded. Two patients who had a simultaneous unrelated medical reason for neurologic symptoms----one halothane hepatitis with asterixis, and one alcohol withdrawal--were not included. Finally, 14 patients who subsequently had intracerebral hemorrhage, although: they may have had preliminary symptoms compatible with hyperperfusion syndrome, were excluded. Thus, we identiffed 10 patients who had unequivocal and uncomplicated hyperperfusion syndrome in the postoperative period. Preoperative characteristics. Of the 10 patients with classic cerebral hyperperfusion syndrome, seven were female and three male. Ages ranged from 20 to 73 years (median age 60 years). Eight of the 10 patients had longstanding hypertension. Four patients had had a previous stroke, all of them ipsilateral. No patient had a history of previous migraine

3 630 Re~qel et al. Journal of VASCULAR SURGERY Table I. Xenon-labeled cerebral blood flows Patient Baseline Postocclusion * NOTE: Data are expressed in milliliters per 100 gm of brain tissue per minute. *Ipsilateral carotid occlusion with unmeasurable flow. or seizure disorder; no patient had diabetes meuitus, two had hyperlipidemia, and one had coronary artery disease. Indications for carotid endarterectomy were transient cerebral ischemic attacks in six patients, amaurosis fugax in one patient, and asymptomatic highgrade carotid stenosis documented by angiography in three patients. Four patients had stasis retinopathy, and one had a contralateral retinal embolus. Retinal artery pressures were obtained in 9 of 10 patients; seven of the nine had a preoperative systolic retinal artery pressure on the operated side of less than 50 mm Hg. In nine of the ten patients, carotid stenosis was due to atherosclerosis; the other patient, a 20-yearold woman, had Takayasu's arteritis. Five of the 10 patients had a unilateral internal carotid artery stenosis of at least 98%, and three additional patients had bilateral high-grade stenoses. Two patients had unilateral stenosis with contralateral occlusion. Intraoperative characteristics. Nine patients underwent unilateral carotid endarterectomy; the patient with Takayasu's arteritis had simultaneous bilateral revascuiarization with a bifurcated graft from the ascending aorta. All 10 patients had intraoperative EEG monitoring. Three patients had no EEG changes with carotid clamping and were not shunted. Seven patients were shunted: five because of changes in the EEG with clamping and two because of previous ipsilateral stroke without EEG changes. Seven of the ten patients had xenon-labeled cerebral blood flow measurements intraoperatively. An aliquot of 200 to 300 ~Ci of xenon 133 was diluted with saline solution to a total volume of 0.2 to 0.3 ml and injected into the common carotid artery with the external carotid artery temporarily occluded. Regional cerebral blood flow was calculated from clearance curves on the basis of extracranial detection of the injected radioisotope, with methods described previously.l,2 In two of the seven patients, postocclusion flows were three times baseline flows; in two other patients, flows after endarterectomy were four times baseline flows. Absolute increases in xenon-labeled blood flow ranged from 8 ml/100 gm of brain tissue/min to 94 ml/100 gm/min, with a mean increase of 54 ml/100 gm/min. Baseline and postocclusion xenon-labeled blood flows for the seven patients are shown in Table I. In four patients a saphenous vein patch was used for endarterectomy closure; in one patient, a Dacron patch was used. Heparin was reversed in two of the ten patients. All of the operations were done with the patient under general anesthesia: four under enflurane (Ethrane), three with isoflurane, and three early in the series with other agents (ketamine, fluothane, and droperidol and fentanyl citrate [Innovar]). One perioperative complication unrelated to cerebral hyperperfusion occurred, a thrombosis of the left external carotid artery. Postoperative course. Retinal artery pressures were obtained postoperatively in all 10 patients. In the nine patients with preoperative retinal artery pressures, increases in systolic pressure ranged from 5 to 74 mm Hg (mean 47 mm Hg). In the postoperative period, seven of the ten patients complained of unilateral headache on the side of the carotid endarterectomy. Headache appeared between the first and seventh day postoperatively, with a mean interval of three days. Nine of the ten patients had postoperative focal seizures contralateral to the side of the endarterectomy and secondarily generalized convulsions. The patient who did not manifest seizure activity clinically had evidence of subclinical seizure activity on EEG ipsilateral to the endarterectomized carotid vessel. EEG was obtained routinely during the postoperative period in all ten patients. In all 10 patients the EEGs were read as showing PLEDs. All 10 patients were treated with phenytoin (Dilantin) for their seizures, which were frequently difficult to control; many of them also required diazepam (Valium) and phenobarbital. Three of the ten patients exhibited an acute organic brain syndrome characterized by paranoid ideation and required treatment with chlorpromazine (Thorazine) and haloperidol (Haldol). One patient exhibited severe clinical depression after the seizures. Seven of the patients were evaluated postoperatively with CT scanning. In six patients, the CT scan was normal or unchanged from preoperative scans. In one patient, the CT scan showed patchy edema, which was thought to be consistent with hyperperfusion (Fig. 2). Three patients underwent postoperative cerebral angiography to assess the possibility of technical problems at the endarterectomy site caus-

4 Volume 5 Number 4 April 1987 Hyperperfusion syndrome 631 Fig. 2. Series of head CT scans in patient with cerebral hyperperfusion syndrome. A, Nine days before carotid endarterectomy, normal scan. B, Five days after endarterectomy at time of seizure, showing diffuse patchy edema consistent with hyperperfusion. C, Eight weeks after endarterectomy, with resolution of edema. ing neurologic dysfunction; all angiograms confirmed a patent internal carotid artery and none showed evidence of thromboemboli. The angiogram of one patient had unilateral increased vascularity consistent with cerebral hyperperfusion (Fig. 3). Five patients had transient postictal paresis and one patient had a visual field cut contralateral to the operated carotid artery. One patient had a transient neurologic deficit as the first sign of postoperative difficulty. All neurologic deficits resolved fully by the time of dismissal. Length of hospital stay ranged from 12 to 29 days, with a mean stay of 17.8 days. DISCUSSION In 1968, Waltz 3 reported the effects of changes in systemic blood pressure on blood flow in ischemic and nonischemic cerebral cortex. He anesthetized 21 cats and occluded the middle cerebral artery on one side in each animal. He then manipulated systemic blood pressure with nitroprusside and measured cortical blood flow. In nonischemic cortex, cortical blood flow remained constant despite changes in systemic blood pressure, whereas in ischemic cortex, cortical blood flow varied directly with systemic blood pressure from 35 to 120 mm Hg. Above 120 mm Hg, there was usually no continued increase in cortical blood flow. However, in three animals increased systemic blood pressure did produce a continued increase in cortical blood flow. Waltz hypothesized that cerebral ischemia produces an impairment of autoregulation of cerebral blood flow. In 1975, Leviton, Caplan, and Salzman 4 described headaches occurring after carotid endarter- ectomy and postulated that they might be related to loss of autoregulation. In that same year, Sun&, Sandok, and Whisnant 5 reported six patients who had seizures after carotid endarterectomy and related their postoperative neurologic dysfunction to cerebral hyperperfusion. In 1981, Sundt et al.6 reviewed cerebral blood flow and EEG changes in 1145 carotid endarterectomies performed from 1972 to They found the most common causes of perioperative neurologic complications to be intraoperative embolization and postoperative hyperperfusion syndromes. Six patients had severe unilateral headache postoperatively (mean increase in cerebral blood flow was 40 ml/100 gm/min). Two of the seizure patients early in the series were treated with heparin and sustained intracerebral hemorrhage. Three additional patients had intracerebral hemorrhage postoperatively, with two of the three patients having received prophylactic aspirin therapy. Mean increase in cerebral blood flow for the five patients with intracerebral hemorrhage was 47 ml/100 gm/min. By comparison, mean increases in cerebral blood flow for all other patients without these postoperative neurologic complications was 1 to 5 ml/lo0 gm/min. Sundt et al.6,7 concluded that patients with a high-grade carotid stenosis and marked increases in cerebral blood flow perioperatively were at risk for postoperative neurologic dysfunction. Several authors published case reports of a total of seven patients who manifested perioperative courses characteristic of hyperperfusion syndrome: preoperative high-grade carotid stenosis, postoper-

5 632 Re~el et al. Journal of VASCULAR SURGERY Fig. 3. Cerebral angiograms in patient with cerebral hyperperfusion syndrome. A, Five days before carotid endarterectomy. B, Eight days after endarterectomy at time of seizure, showing increased cerebral vascularity ipsilateral to carotid endarterectomy. arrive headache, and postoperative seizures on the side of the body contralateral to the carotid endarterectomy. 8-1 Seizures were often difficult to control and in one case progressed to ipsilateral cerebral hemorrhage and death. On postmortem examination, there were no intravascular thromboemboli and pathologic changes resembled those seen in the brain in acute severe hypertension. Bernstein, Fleming, and Deck 11 in 1984 described the case of a 56-year-old man with a very high-grade left carotid stenosis who had left carotid endarterectomy. Postoperatively he received 1300 mg of aspirin per day. On the first postoperative day he complained of a severe left-sided headache. On the second postoperative day he had grand mal seizures. On the sixth postoperative day he suffered right hemiparesis and died of a left intracerebral hemorrhage. At postmortem examination, small arteries and arterioles throughout the left cerebral cortex showed reactive changes consisting of swelling and hyperplasia of endothelial cells, extravasation of erythrocytes, and some fibrinoid necrosis; there was marked edema of the adjacent brain. These features of altered vascular permeability are similar histologically to those seen in the brain in cases of malignant hypertension. The small blood vessels of the opposite hemisphere were normal. The authors concluded that chronic cerebral ischemia distal to the high-grade carotid stenosis had led to chronic vasodilatation, loss of autoregulation, and a consequent absence of arterial vasoconstriction to protect the capillary bed. Schroeder, Holstein, and Engell ~2 in 1984 described the postoperative course of a 55-year-old man with a high-grade left carotid lesion and an occluded right carotid artery who underwent left carotid endarterectomy. On the first postoperative day, the patient complained of a headache. Sequential cerebral blood flow measurements with the use of intravenous xenon revealed an initial increase of greater than 200%; values gradually returned to normal by the seventh postoperative day. When these studies are considered in conjunction with the current description, a clear picture emerges of cerebral hyperperfusion syndrome. Patients with this disorder often have a history of systemic hypertension and undergo carotid endarterectomy to relieve severe chronic cerebral ischemia. Cerebral blood flow in this disorder usually increases markedly after endarterectomy. Postoperatively the patients often complain of unilateral headache, followed in some cases by ipsilateral cerebral neurologic deficit and/or seizures; intracerebral hemorrhage may occur. CT scans during this period are normal or show patchy edema over the affected hemisphere. Angiograms are normal or show increased perfusion of the affected hemisphere. Pathologic examination of postmortem specimens is consistent with changes seen in malignant hypertension. Cerebral blood flows measured postoperatively confirm continued hyperperfusion for days after surgery, with eventual normalization of cerebral blood flow as autoregulation returns. Although this syndrome has been poorly recognized in the past, it is evident that many vascular surgeons have seen one or more patients whose clinical course is compatible with hyperperfusion. Unfortunately, these cases have often been thought to represent microembolization and have been treated with heparin, sometimes with subsequent intrace-

6 Volume 5 Number 4 April 1987 Hyperperfusion syndrome 633 rebral hemorrhage. 131s In addition, since most patients are dismissed on the third to fifth day after carotid endarterectomy, some of these patients have their complication only after dismissal. Thus, guidelines for earlier diagnosis and management would be helpful. Careful preoperative examination and thorough evaluation of the cerebral angiograms will help to identify those patients who have findings suggestive of severe cerebral ischemia. Patients with severe stasis retinopathy, bilateral internal carotid artery stenosis greater than 95%, unilateral carotid occlusion with contralateral high-grade carotid stenosis, and patients with unilateral very high-grade carotid stenosis with poor collateral cross-filling of that hemisphere should be recognized as potentially at risk for development of hyperperfusion syndrome. If perioperative use of antiplatelet agents or anticoagulants is indicated, they should be used with caution. Headaches are known to occur quite frequently in patients after carotid endarterectomy. 16~8 Mild headaches are usually managed easily with mild analgesics. However, severe unilateral headaches deserve further evaluation and careful management. Blood pressure must be meticulously controlled and maintained within a normal range. EEG is indicated to detect the presence of PLEDs, which are indicative of an acute, localized cerebral focus of irritability. Clinical judgment may also suggest CT scan of the head to rule out intracerebral hemorrhage. If PLEDs are noted on EEG, or if seizures occur, anticonvulsant medication should be started. Once seizures occur, they may be particularly difficult to control and may require such heavy, sedation that assisted ventilation is required. After resolution of thc seizures, anticonvulsants are generally continued for at least 3 to 6 months, with repeat EEG examination before the medication is discontinued. Transient neurologic deficits, which frequently occur in this syndrome, are usually evident postictally. However, in some patients, a lateralizing deficit may be the first sign of hyperperfusion syndrome. An angiogram is usually obtained in any postendarterectomy patient if a neurologic deficit occurs after the first 24 hours following operation. (Neurologic deficits occurring within the first 24 hours are more often technical in origin and usually warrant immediate reexploration.) In some patients, seizure activity is so difficult to control that angiography must be deferred until the patient can be stabilized and the blood pressure can be brought within normal range. Intracerebral hemorrhage can be a sequela of this syndrome and may occur in more than 40% of these patients. CT is indicated in most patients who have neurologic deficits or seizures. Obviously, anticoagulants should be avoided in these patients unless specific indications mandate their use. SUMMARY Cerebral hyperperfusion is an uncommon and potentially lethal syndrome that may occur after carotid endarterectomy. The development ofipsilateral headache during the postoperative period is not a benign phenomenon but a possible harbinger of seizures, transient neurologic dysfunction, or intracerebral hemorrhage with permanent neurologic deficits and death. Patients who complain of ipsilateral headache should have careful control of systemic blood pressure. Those patients who have seizures without evidence of focal ischemia on CT scanning should be treated with anticonvulsants. Anticoagulants should be avoided. Perioperative antiplatelet agents should be used cautiously, if at all, in patients with highgrade carotid stenosis who are at risk for the development of cerebral hyperperfusion syndrome. REFERENCES 1. Anderson RE, Sundt TM Jr. An automated cerebral blood flow analyzer: concise communication. J Nucl Med 1977; 18: Waltz AG, Wanek AR, Anderson RE. Comparison of analytic methods for calculation of cerebral blood flow after intracarotid injection of ls3xe. J Nucl Med 1972;13: Waltz AG. Effect of blood pressure on blood flow in ischemic and in nonischemic cerebral cortex. Neurology 1968;18: Leviton A, Caplan L, Salzman E. Severe headache after carotid endarterectomy. Headache 1975;15: Sundt TM Jr, Sandok BA, Whisnant JP. Carotid endarterectomy: complications and preoperative assessment of risk. Mayo Clin Proc 1975;50: Sundt TM Jr, Sharbrough FW, Piepgras DG, Kearns TP, Messick JM Jr, O'Fallon WM. Correlation of cerebral blood flow and electroencephalographic changes during carotid endarterectomy. Mayo Clin Proc 1981;56: Sundt TM Jr. The ischemic tolerance of neural tissue and the need for monitoring and selective shunting during carotid endarterectomy. Stroke 1983;14: Dolan JG, Mushlin AI. Hypertension, vascular headaches, and seizures after carotid endarterectomy. Arch Intern Med 1984; 144: Wilkinson JT, Adams HP Jr, Wright CB. Convulsions after carotid endarterectomy. JAMA 1980;244: Youkey JR, Clagett GP, Jaffin JH, Parisi JE, Rich NM. Focal motor seizures complicating carotid endarterectomy. Arch Surg 1984;119: Bernstein M, Fleming JFR, Deck JHN. Cerebral hyperperfusion after carotid endarterectomy: a cause of cerebral hemorrhage. Neurosurgery 1984;15: Schroeder T, Holstein PE, Engell HC. Hyperperfusion following endarterectomy (Letter). Stroke 1984;15:758.

7 634 Reigd et al. Journal of VASCULAR SURGERY 13. Wells CE. Cerebral embolism. Arch Neurol Psychiatr 1959;81: Mohr JP, Caplan LR, Melski JW, et al. The Harvard Cooperative Stroke Registry: a prospective registry. Neurology, (Minneap) 1978;28: Cocito L, Favale E, Reni L Epileptic seizures in cerebral arterial occlusive disease. Stroke 1982; 13: Pearce J. Headache after carotid endarterectomy. Br Mcd J 1976;2: Messert B, Black J. Cluster headache, hemicrania, 0aid other head pains: morbidity of carotid endarterectomy. Stroke 1978;9: Appenzeller O. Cerebrovascular aspects of headache. Med Clin North Am 1978;62:467-9.