0148-396X/84/1502-0178$02.00/0 NEUROSURGERY Copyright € 1984 by the Congress of Neurological Surgeons Vol. 15. No. 2. 1984 Printed in U.S.A. Temporary Neurological Deterioration after Extracranial- Intracranial Bypass Roberto C. Heros, M.D., R. Michael Scott, M.D., J. Philip Kistler, M.D., Robert H. Ackerman, M.D., and E. Scott Conner, M.D. Harvard Medical School, Massachusetts General Hospital (R.C.H., J.P.K., R.H.A., E.S.C.), and New England Medical Center (R.M.S.), Boston, Massachusetts Five patients who experienced temporary neurological deterioration after extracranial to intracranial bypass procedures are reported in detail. These patients suffered transient ischemic attacks or more prolonged deficits usually of a different nature than the preoperative symptoms. All patients had a good outcome and the spells ceased; the neurological deficits improved within a maximum of 2 weeks. Obvious causes of deterioration such as intra- or extracerebral hematomas, occlusion of a previously stenotic vessel, or graft occlusion were ruled out by computed tomography and angiography in each case. Intraoperative causes of neurological deterioration such as anesthetic effect, hypotension, and temporary occlusion of the cortical vessel or sacrifice of its small branches were not likely to be the cause of the deficits because in each case. the patient awoke satisfactorily and deterioration occurred hours to days later. In each case, postoperative angiography showed good perfusion of at least one major division of the middle cerebral territory. Anticoagulation with heparin in three patients did not change the clinical course. In one patient who was not anticoagulated, embolism could have been responsible for a single prolonged ischemic event, but in the other patients thromboembolism does not seem likely to have been responsible for the deficits. The cause of the deterioration in these patients remains unexplained. We speculate that hyperperfusion of chronically ischemic brain tissue and shifts in the watershed region resulting from the new flow pattern after bypass grafting are two mechanisms that may have been of importance in the etiology of these deficits. (Neurosurgery 15:178-185, 1984) Key words: Angiography, Cerebral blood flow, Cerebral ischemia, Computed tomography, Extracranial-intracranial bypass graft, Postoperative neurological deficit, Transient ischemic attack In our relatively modest experience with extracranial to intracranial (EC-IC) procedures, we have been impressed by the frequency with which we have encountered temporary postoperative neurological deterioration. This deterioration has been in the form of reversible focal neurological deficits lasting up to a few days. We have reviewed our experience with 134 consecutive EC-IC procedures. We are reporting, in some detail, cases in which a definite temporary neurological deterioration occurred during the immediate postoperative period. We are excluding cases of serious permanent compli- cations, which have been discussed in previous publications (11-13, 25). In patients with permanent neurological morbid- ity, a clear-cut cause (intracerebral hematoma, subdural he- matoma, embolization, occlusion of a stenotic artery, aneu- rysmal rupture, graft occlusion, seizures, hypotension, etc.) could be found almost without exception. In contrast, the cause of temporary deterioration in the group of patients that we are reporting here remains unknown, and our attempts to explain the observed deficits are purely speculative. CASE REPORTS Case | This 69-year-old man presented initially in May 1981 with a l-month history of recurrent episodes of right facial weak- ness, right arm weakness, and tingling of the right lip and fingers. Each episode lasted 5 to 10 minutes. An arteriogram showed complete occlusion of the left internal carotid artery (ICA) and mild stenosis of the origin of the right ICA. The intracranial portion of the right ICA showed only mild irreg- ularity. The patient was placed on sodium warfarin therapy and had no further spells. Routine noninvasive studies in September 1982 suggested progression of the stenotic lesion 178 in the right ICA and, therefore, he was electively admitted for angiography. The angiogram showed that the left ICA remained com- pletely occluded. In addition, a very focal narrowing of the intracranial portion of the right ICA at the carotid siphon was now present. This area of narrowing measured about | mm in the lateral view. The left middle cerebral artery (MCA) territory was supplied mostly in an anterograde, although delayed, fashion by the right ICA through a small anterior communicating artery (AComA) (Fig. 1A). Regional cerebral blood flow (rCBF) shown by xenon inhalation revealed sig- nificantly lower flow with preserved CO) reactivity in the left hemisphere. Although the patient had remained asymptomatic on war- farin therapy, the intracranial lesion of the right ICA, which had progressed markedly in spite of anticoagulants, was thought to represent a major threat to the patient because both hemispheres were now being supplied by the right ICA. Hence, a left superficial temporal (STA) to MCA bypass was performed prophylactically to “protect” the left hemisphere in case of occlusion of the right ICA. The operation was accomplished without difficulty, and the patient was neuro- logically well when he awoke from anesthesia. However, the next day he developed moderately severe pure motor hemi- paresis and expressive speech difficulty. A computed tomo- graphic (CT) scan was normal. An arteriogram showed a widely patent anastomosis (Fig. 1B). The STA graft now filled the left MCA trunk completely in a retrograde fashion (Fig. 1C). A right carotid injection showed that the right ICA lesion at the siphon was unchanged. The right ICA now filled only the left A-1, and the dye column tapered off at the origin of the left MCA (Fig. 1D). The patient’s hemiparesis and expressive aphasia resolved August 1984 He 3 D * ». Fic. I. DETERIORATION AFTER EC-IC BYPASS 179 m. — tN 4 Me ws Case 1, A, preoperative anteroposterior (AP) right carotid injection shows delayed filling of the left MCA territory. B. postoperative left common carotid injection shows a large STA graft filling most of the left MCA territory. C. postoperative AP left common carotid injection shows retrograde flow from the STA into the proximal MCA trunk. D. postoperative AP right carotid injection shows that there is no filling of the left MCA territory and that the contrast column tapers off at the origin of the left MCA. spontaneously over the next 48 hours. He was last seen in June 1983 and at that time had remained asymptomatic. The etiology of this serious, although temporary, deficit remains unclear. Had this been an embolic event. we would not have expected such complete resolution from a profound deficit involving face, arm. leg. and speech. Could the reversal of flow in the left MCA trunk (initially anterograde; retrograde after the bypass) have been responsible? His deficit could conceivably be attributed to relative ischemia in the region supplied by perforators originating from the MCA trunk. After bypass grafting. this area was indeed the “watershed” or “distal field” of the left MCA. which was now supplied in a retrograde fashion by the bypass graft (8). The resolution of the deficit may have been due to gradual improvement of flow to the new watershed area. Alternatively, the problem could be explained by the phenomenon of hyperperfusion to a chronically ischemic and therefore dysautoregulated area of the brain (27). Case 2 This 56-year-old man presented because of several spells of weakness and numbness of the right arm, leg. and face. Each spell lasted for about 3 to 4 minutes. Despite aspirin and dipyridamole, he continued to have spells. On admission, his 180 HEROS et al. neurological examination was completely normal. An arter- iogram showed that he had a complete occlusion of the left ICA. The right ICA filled the left hemisphere through a large AComA (Fig. 2.4). In August 1982. a left STA-MCA bypass graft was performed without difficulty. The patient was well when he awakened from anesthesia. Over the next several days. however. he had several spells of expressive aphasia and right hand weakness. These spells lasted longer than the preoperative spells and, in addition. they differed from the preoperative spells in that now speech was involved. A CT scan was normal. A postoperative angiogram, during the time when the patient was having the spells, showed excellent filling of the posterior division of the MCA by the graft. The anterior division of the MCA was also filled by the graft, but the filling was delayed and poor (Fig. 2B). The middle cerebral trunk on the left was now filled in a retrograde fashion by the graft (Fig. 2C). Heparin treatment did not influence the frequency or severity of his spells. His spells cleared without further specific treatment after 10 days, and he has remained asymp- tomatic. Detailed study of the arteriograms reveals that, whereas preoperatively both divisions of the left MCA filled simulta- neously from the right carotid injection, now there was de- layed filling of the anterior division of the left MCA. The recipient vessel was a branch of the posterior division, and it filled the MCA trunk in a retrograde fashion. The anterior Neurosurgery, Vol. 15, No. 2 division was then filled in an anterograde but delayed fashion. The anterior division had in effect become the “distal field” or “watershed” area of the ieft MCA territory now supplied by the graft. We think that this patient’s postoperative spells may have been due to a relative “low flow” situation in this new distal field and that the spells ceased as flow through the graft gradually increased. Case 3 This 60-year-old man presented initially in February 1983 with a 3-month history of recurrent transient ischemic attacks (TIAs) consisting of clumsiness of the right arm and hand and expressive speech difficulty. The spells increased in frequency until, at the time of admission, he was having several spells a day. Each spell lasted from seconds to a few minutes. Exam- ination at the time of admission revealed that muscle stretch reflexes were increased on the right side, with mild clumsiness of the right hand. The patient’s speech was hesitant. with a mild expressive aphasia. rCBF studies showed decreased flow in the left hemisphere. An arteriogram showed complete occlusion of the left ICA. In addition, there was severe stenosis of the origin of the left external carotid artery (ECA) and a large ICA stump. There was some filling of the left MCA territory via external carotid to ophthalmic collaterals (Fig. 3A). The right carotid injection filled the anterior cerebral shows retrograde filling of the left MCA trunk by the STA graft. Case 2. 4, preoperative AP right carotid injection shows rapid filling of the left MCA territory. B, postoperative lateral left common. tion shows a large STA filling preferentially the posterior division of the left MCA. C, postoperative AP left common carotid injection Fic. 3. Case 3. 4, preoperative lateral left common carotid injection shows a large ICA stump. a tight stenosis at the origin of the left ECA, retrograde filling of the intracranial portion of the ICA by the ophthalmic artery, and a small left STA. B. preoperative AP right carotid injection shows no filling of the left MCA territory. C. postoperative lateral left common carotid injection shows an enlarged STA supplying mostly the inferior division of the MCA. August 1984 territory of each hemisphere and. in addition, the entire upper basilar territory. The left MCA did not fill (Fig. 3B). At fluoroscopy, one vertebral artery was thought to be occluded and it was very difficult to catheterize the other narrow vertebral artery: therefore. no vertebral study was obtained. The “stump” of the ICA was rather large and its appearance, in addition to the presence of a distal thin column of contrast agent that was presumed to be in the ICA, suggested that this was an example of “pseudoocclusion” and that perhaps the ICA could be opened successfully by endarterectomy (26). Therefore, a carotid endarterectomy was attempted. At oper- ation, we were unable to obtain good backflow from the ICA, even with the use of the Fogarty catheter. Therefore, the ICA was ligated and an angioplasty was performed at the bifurca- tion to enlarge the origin of the ECA to enhance the collateral flow through the ophthalmic artery. A postoperative digital subtraction angiogram showed patency of the ECA. In addi- tion, the pulse of the STA, which had been weak preopera- tively, was now quite strong. We contemplated proceeding with an STA-MCA bypass procedure but, because the fre- quency of the TIAs diminished markedly and the patient thought that his speech and right hand use were improved, we elected to treat him conservatively with antiplatelet drugs. During the next 3 months, he had several TIAs. By March, the frequency of his TIAs was increasing. We decided to proceed with an STA-MCA bypass graft. The operation was accomplished without difficulty. During the Ist postoperative week, however, the patient had several episodes of marked expressive aphasia with numbness and drooping of the right side of the mouth. In addition, he had several spells of right- sided weakness involving the face. arm, and leg. These spells differed from the preoperative spells in that they were of longer duration. were more severe, and consisted of a different constellation of symptoms. His preexisting persistent speech hesitancy and right hand clumsiness had definitely worsened. A CT scan showed only a small amount of subdural air in the operative site and no evidence of cerebral infarction. A nuclear magnetic resonance (NMR) scan suggested an area of ischemia in the left posterior inferior frontal region. An elec- troencephalogram (EEG) showed some theta slowing on the anterior temporal and frontal leads on the left side. There was no evidence of seizure activity. rCBF studies showed decreased flow only in the frontal region. and this was confirmed by positron emission tomography (PET) which, in addition. showed that O2 metabolism was decreased proportionally to blood flow in this area. He was treated with phenytoin and heparin without obvious benefit. A postoperative arteriogram showed that the graft filled very rapidly the posterior division of the MCA (Fig. 3C). Filling of the anterior division of the middle cerebral territory was considerably delayed. The an- DETERIORATION AFTER EC-IC BYPASS 181 ms i at Fic. 4. Case 4. A, preoperative left carotid injection shows occlusion of the ICA just beyond the ophthalmic artery. B, preoperative right carotid injection shows good filling of the left MCA territory. C, postoperative left common carotid injection shows filling of the entire MCA territory. terior division seemed to be filled partially by the graft and partially by external collaterals through the ophthalmic artery. The patient's spells resolved over a 2-week period without further specific treatment. Follow-up examination revealed that his mild neurological deficit had significantly improved, and he had had no further TIAs. This man presented clear clinical, arteriographic, and blood flow evidence of “low flow” in the left hemisphere. In addition to the multitude of brief stereotyped TIAs he had initially, he presented with a mild, progressing neurological deficit. which was referred to as a “slow stroke” by Sundt et al. (27). We suspect that he had significant vertebrobasilar occlusive dis- ease and, therefore, had a limited capacity for collateral flow to the left MCA territory from the posterior circulation. In addition, he had an atretic or nonexistent left A-1. His only source of anterograde flow to the left MCA territory was through the ophthalmic artery. His condition improved tem- porarily after the external endarterectomy, probably as a consequence of increased flow through the ECA to ophthalmic collaterals. The STA-MCA bypass supplied primarily the posterior territory of the MCA. His clinical symptoms postoperatively suggested ischemia in the anterior (frontal) distribution of the MCA. This was confirmed by NMR, rCBF, and PET studies as well as by the EEG. It is likely that the graft, by reversing flow patterns, created a situation of relative hypoperfusion in the anterior territory of the left MCA that improved gradually as flow through the graft increased. Case 4 This 67-year-old man was initially seen in an outside hos- pital complaining of a 6-day history of frequent episodes of slurred speech and numbness and weakness of the right face and right upper extremity. A CT scan was normal. An arter- iogram on May 22. 1981, showed total occlusion of the left ICA just above the ophthalmic artery (Fig. 44). The right carotid injection showed that the left MCA territory was filled in an anterograde fashion through the circle of Willis (Fig. 4B). A left STA-MCA was performed on June 24, 1981. Initially, the patient did well after operation, but the next day he developed slurred speech. weakness of the right upper extrem- ity, and confusion. A CT scan showed no explanation for this deficit. An arteriogram showed wide patency of the bypass with filling of the entire MCA circulation on the left side (Fig. 4C). Over the next 24 hours, the deficit gradually cleared and the patient has remained asymptomatic. Although there is no preoperative rCBF study available, the 182 HEROS et al. arteriogram suggests that there was good flow to the left hemisphere from the right ICA via the AComA. Therefore. this is not likely to have been a chronically hypoperfused hemisphere and the syndrome of postoperative “hyperperfu- sion” does not seem to have been operative in the etiology of this patient’s postoperative spell. In addition. the entire MCA territory filled well from the graft and there were no definite areas of delayed poor perfusion. It is possible that thrombo- embolism accounted for the transient deficit. This could have been related to a postoperative hypercoagulable state or to dislodgement from the distal clot on the ICA (14). It could be presumed that the ICA occlusion occurred only | month preoperatively when he had the initial cluster of TIAs and, if such was the case, the distal thrombus on the ICA may still have not been organized. Alternatively, this could have been another instance of relative stasis in the proximal MCA trunk, as discussed in Case |. Case 5 This 65-year-old gentleman with angina at rest and severe peripheral vascular disease had several episodes of right arm weakness and speech difficulty. He had one more spell despite heparin therapy, which was followed by a significant gastroin- testinal hemorrhage requiring a transfusion of 6 units of blood. He had two additional TIAs after discontinuation of the heparin, resulting in a fixed neurological deficit consisting of slurred speech and mild right upper extremity weakness. An arteriogram showed complete occlusion of the left ICA. The right ICA did not fill any of the left middle cerebral distribution. After the arteriogram, heparin therapy was re- started and the patient was transferred to the Massachusetts General Hospital. Soon after arrival, he had another episode of profound right hemiparesis and aphasia. This episode cleared after several hours, but he continued to show evidence of persistent right- sided weakness, slurring of speech, and confusion. A CT scan showed evidence of recent infarction in the frontal “wa- tershed” region. The last spell occurred in spite of adequate anticoagulation. In view of this situation, we decided to proceed with a left STA-MCA bypass graft. Postoperatively, the neurological examination showed no change from his preoperative condition. Two days after op- eration, he had another spell of profound right hemiparesis and marked dysphasia. Heparin was restarted: however, on the following day he had another similar episode. In addition, his persistent neurological deficit was found to be slightly worse than it had been immediately postoperatively. In par- ticular, his confusion, as well as the slurred speech, was worse. ACT scan once again showed the area of watershed infarction seen preoperatively, but there appeared to be more edema than in the preoperative scan. He underwent postoperative arteriography. which showed good filling of all of the left MCA territory by the bypass graft, but also some mass effect with depression of the sylvian triangle, particularly in the parietal region (Fig. 5). The heparin was discontinued. His TIAs stopped after about 2 weeks and his neurological deficit has improved gradually to the point that his condition is better than preoperatively. This man clearly was in an unstable neurological situation preoperatively and he was worsening in spite of heparin treatment. His clinical course, watershed infarction on CT scan, and lack of anterograde filling of the left MCA territory demonstrated angiographically indicate that he had a true “low flow” situation before operation. He awoke from anes- thesia unchanged and then worsened in spite of an adequate bypass graft. which should have improved the blood flow to Neurosurgery, Vol. 15, No. 2 Fic. 5. Case 5. Postoperative left common carotid injection shows a large posterior branch of the STA supplying all of the MCA territory. Note the depression of the sylvian triangle by a posterior frontal- parietal mass effect. the left hemisphere. Increased edema and mass effect from the area of preoperatively demonstrated infarction was shown by CT scanning and angiography. We postulate that, in this patient, the graft produced a state of hyperperfusion that exacerbated edema related to a recent infarction. DISCUSSION Several authors have indicated that temporary neurological deterioration can occur after EC-IC procedures (3, 5, 17-23, 28-30). Such transient postoperative symptoms are said to occur in approximately 4 to 20% of the patients (3, 17, 20- 22, 29, 30). The nature of these symptoms is not discussed in detail in the literature. Popp and Chater state that, of their 110 patients, 6 had transient speech difficulty and 4 had transient motor difficulty postoperatively (18). Four of 51 patients of Deruty and colleagues suffered transient aphasia postoperatively (5). Similarly, 3 of 31 patients in another series suffered transient postoperative expressive aphasia (23). Reichman reported | patient with hemiparesis and aphasia and 2 patients with hemiparesis in a series of 70 patients (19). The etiology of these transient postoperative deficits re- mains unknown. Robertson and Watridge suggested that they occur as a result of temporary occlusion of the cortical vessel (20). Reichman thinks that the deficits may be related to sacrifice of penetrating branches from the recipient cortical vessels (19). In a study of serial EEGs and blood flow meas- urements before and after EC-IC procedures, DeWeerd et al. found that the EEG frequently appeared worse during the immediate postoperative period (6). They attributed this de- terioration to the effects of anesthesia because eventually the EEG returned to base line. They did not comment on whether these patients had any clinical evidence of neurological dys- function when the EEG was abnormal. The cause of deterioration in our group of patients is not obvious. All awakened from anesthesia satisfactorily and then deteriorated, making it unlikely that the deficits were related to the effects of anesthesia. Some had a small subdural collec- tion of air or fluid shown on the postoperative CT scan, but in none was there a sizable subdural or epidural hematoma. All of the patients had postoperative CT scans and in none was there an intracerebral hemorrhage or evidence of new infarction. Because none of the deficits were evident when the patient first awakened, we cannot attribute the deficits to temporary intraoperative occlusion of the cortical vessel or to sacrifice of perforators from these vessels. In addition, in most patients the deficits were too diffuse (face, arm, and leg or speech and hand) to be explained by this mechanism or even August 1984 by permanent occlusion of the cortical artery. All of the grafts were patent postoperatively: in fact, in all five cases we would classify the degree of cerebral irrigation by the graft as good to excellent. None of the clinical syndromes observed were characteristic of seizure activity, and three patients were tak- ing anticonvulsants without influence on their symptoms. In addition, postoperative EEGs in three patients showed no evidence of sharp activity. All of our patients had occlusion of the ipsilateral ICA and, therefore, the possibility of occlu- sion of a previously stenotic vessel as a consequence of the bypass procedure (2, 7, 8) does not apply to our cases. There were no significant episodes of intraoperative or postoperative hypotension in our patients. We cannot rule out the possibility that some of the events, particularly the single prolonged event in Case 4, were due to embolism from the “stump” of the ICA either through the ophthalmic artery (1) or through the graft (4) or from a distal thrombus in the occluded ICA (14). Four patients were re- ceiving antiplatelet drugs postoperatively, which, of course, does not eliminate completely the risk of thromboembolism. However, full heparinization should have improved the thromboembolic symptoms and it did not change the clinical course in the three patients who were fully anticoagulated postoperatively. Two patients (Cases | and 4), although symptomatic in the past, were free of symptoms immediately before operation. The rest of the patients had symptoms during the few days before operation, and we thought that the symptoms were hemodynamic in nature rather than embolic. It is, of course, difficult to make this judgment but. by using clinical and angiographic data plus cerebral blood flow and PET studies when possible, we attempt to select for EC-IC grafting patients who seem to be having TIAs on a hemodynamic basis (9, 10, 14-16, 24). Two of our patients had an “isolated hemisphere” angiographically. In these patients, the circle of Willis was functionally incompetent and there was no demonstrable anterograde flow to the involved MCA territory when the contralateral carotid or either vertebral artery was injected. The flow to the MCA territory in these patients is by external collaterals through the ophthalmic artery and by leptomen- ingeal collaterals from the anterior and posterior cerebral arteries. They are then, presumably. in a situation of chronic hypoperfusion. It is in this group of patients that Sundt and coworkers have encountered the syndrome of “hyperperfu- sion” after carotid endarterectomy (27). By correlating the clinical syndromes with intraoperative EEG and blood flow determinations, they concluded that several of their postop- erative complications were due to “hyperperfusion.” These complications ranged from transient focal neurological symp- toms to large intracerebral hemorrhages. These patients usu- ally had tight stenosis with insufficient collateral supply and low blood flow values before endarterectomy. After endarter- ectomy, the flow increased markedly and led to what the authors described speculatively as a “unilateral hypertensive encephalopathy.” Their claim is that these chronically hypo- perfused hemispheres become dysautoregulated and unable to handle physiologically the increased flow. It is possible that some of the deficits observed in our patients were related to this mechanism of abruptly increasing flow in a chronically hypoperfused and presumably dysautoregulated region of the brain. A second mechanism that may account for some of the deficits relates to a shift in flow patterns, In all of our cases, retrograde flow into the MCA trunk from the STA was demonstrated after the bypass. The postoperative angiograms were performed by common carotid injection rather than DETERIORATION AFTER EC-IC BYPASS 183 selective STA injection; therfore, we presume that the ob- served flow pattern was indeed physiological rather than a temporary alteration related to the pressure of the injection. This reversal of flow in the MCA trunk tends to shift phys- iologically the “distal field” or “watershed” area to the perfor- ators coming off the MCA trunk, as has been suggested in other situations (8), and to the other divisions of the MCA. In other words, flow now comes from the division receiving the graft in a retrograde fashion to the MCA trunk and then, in anterograde fashion, the other major divisions of the MCA are irrigated. In fact, delayed filling of the anterior division by the graft was demonstrated arteriographically in Cases 2 and 3; in these patients, the symptoms could be explained by hypoperfusion in that territory. In conclusion, the mechanisms responsible for temporary deterioration in the patients presented remain unknown. None of the obvious causes of postoperative deterioration such as intra- or extracerebral hemorrhage, graft occlusion, or new cerebral infarction were demonstrated in these cases. Anesthesia effect, intraoperative hypotension, and temporary occlusion of the cortical vessel or sacrifice of its small branches are unlikely to have been responsible because the deficits were not obvious during the immediate postoperative period. We cannot rule out thromboembolism, but the clinical syndrome, angiographic picture, and lack of response to anticoagulation did not suggest an embolic etiology except perhaps in one of the cases (Case 4), We speculate that hyperperfusion of chron- ically hypoperfused regions of the brain and shift in flow patterns that effectively changed the “watershed” of the MCA territory after the bypass may have been two of the mecha- nisms responsible for these complications. The outcome was good in all cases, and the deficits resolved spontaneously within 2 weeks; therefore, it seems that patients with this syndrome have a benign outcome if obvious causes of dete- rioration such as intra- or extracerebral hematomas, occlusion of a previously stenotic vessel, or graft occlusion are excluded by CT scan and arteriography. Received for publication, January 20, 1984; accepted, March 31, 1984, Reprint requests: Roberto C. Heros, M.D., Director of Cerebro- vascular Surgery, Massachusetts General Hospital, Boston, Mas- sachusetts 02114. REFERENCES 1. Barnett HJM, Peerless SJ, Kaufmann JCE: “Stump” of internal carotid artery—a source for further cerebral embolic ischemia. Stroke 9:448-456, 1978. . Bodosi M, Gacs G, Merei FT: Stenoses of the distal segments of the internal carotid artery. Surg Neurol 16:109-116, 1981. 3. Chater N, Popp J: Microsurgical vascular bypass for occlusive cerebrovascular disease: Review of 100 cases. Surg Neurol 6:115- 118, 1976. 4. Conley FK: Embolization of a superificial temporal artery to middle cerebral artery bypass: Case report. Neurosurgery 12:342- 345, 1983. . Deruty R. Lecuire J, Bret P, Capdeville: Results in 51 patients with EC-IC bypass: Technical and clinical considerations, in Peerless SJ, McCormick CW (eds): Microsurgery for Cerebral Ischemia. New Y ork, Springer-Verlag, 1980, pp 344-349. 6. DeWeerd AW, Veering MM, Mosmans PCM, van Huffelen AC, Tulleken CAF, Jonkman EJ: Effect of the extra-intracranial (STA-MCA) arterial anastomosis on EEG and cerebral blood flow: A controlled study of patients with unilateral cerebral ischemia. Stroke 13:674-679, 1982. 7, Furlan AJ, Little JR. Dohn DF: Arterial occlusion following anastomosis of the superficial temporal artery to middle cerebral artery. Stroke 11:91-95, 1980. is) w 184 HEROS et al. 8. Gumerlock MK, Ono H. Neuwelt EA: Can a patent extracranial- intracranial bypass provoke the conversion of an intracranial arterial stenosis to a symptomatic occlusion? Neurosurgery 12:391-400. 1983. 9. Halsey JH Jr, Morawetz RB. Blauenstein UW: The hemody- namic effect of STA-MCA bypass. Stroke 13:163-167, 1982. 10. Heilbrun MP: Overall management of vascular lesions considered treatable with extracranial-intracranial bypass: Part 1. Internal carotid occlusion. Neurosurgery 11:239-246, 1982. 11. Heros RC: Thromboembolic complications after combined in- ternal carotid ligation and extra- to intracranial bypass. Surg Neurol 21:75-79, 1984. 12. Heros RC, Nelson PB: Intracerebral hemorrhage after microsurg- ical cerebral revascularization. Neurosurgery 6:371-375, 1980. 13. Heros RC, Nelson PB. Ojemann RG, Crowell RM, DeBrun G: Large and giant paraclinoid aneurysms: Surgical techniques, complications, and results. Neurosurgery 12:153-163, 1983. 14. Heros RC, Sekhar LN: Diagnostic and therapeutic alternatives in patients with symptomatic “carotid occlusion” referred for extracranial-intracranial bypass surgery. J Neurosurg 54:790- 796, 1981. 15. Laurent JP, Lawner PM, O’Connor M: Reversal of intracerebral steal by STA-MCA anastomosis. J Neurosurg 57:629-632, 1982. 16. Levinthal R, Stern WE: Effect of proximal occlusion on anasto- mosis patency in dogs. Neurosurgery 5:596-597, 1979. 17. Mehdorn HM, Hoffman WF, Chater NL: Microneurosurgical arterial bypass for cerebral ischemia: The San Francisco experi- ence, in Peerless SJ. McCormick CW (eds): Microsurgery for Cerebral Ischemia. New York, Springer-Verlag, 1980, pp 350- 356. 18. Popp AJ, Chater N: Extracranial-to-intracranial vascular anas- tomosis for occlusive cerebrovascular disease: Experience in 110 patients. Surgery 82:648-654, 1977. 19, Reichman OH: Complications of cerebral revascularization. Clin Neurosurg 23:318-335, 1976. 20. Robertson JT, Watridge CB: The surgical management of extra- cranial and intracranial occlusive disease. Med Clin North Am 63:68 1-693, 1979. 21. Samson DS, Boone S: Extracranial-intracranial (EC-IC) arterial bypass: Past performance and current concepts. Neurosurgery 3:79-86, 1978. 22. Samson DS, Hodosh RM, Clark WK: Microsurgical treatment of transient cerebral ischemia: Preliminary results in 50 patients. JAMA 241:376-378, 1979. 23. Samson D, Watts C, Clark K: Cerebral revascularization for transient ischemic attacks. Neurology (NY) 27:767-771, 1977. 24. Schmiedek P, Gratz] O, Spetzler R. Steinhoff H, Enzenbach R, Brendel W, Marguth F: Selection of patients for extra-intracranial arterial bypass surgery based on rCBF measurements. J Neuro- surg 44:303-312, 1976, 25. Scott RM, Liu H-C, Yuan R, Adelman L: Rupture of a previously unruptured giant middle cerebral artery aneurysm after extra- cranial-intracranial bypass surgery. Neurosurgery 10:600-603, 1982. 26. Sekhar LN, Heros RC, Lotz PR, Rosenbaum AE: Atheromatous pseudo-occlusion of the internal carotid artery. J Neurosurg 52:782-789, 1980. 27. 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: With results of surgery and hemodynamics of cerebral ischemia. Mayo Clin Proc 56:533-543, 1981. 28. Sundt TM Jr, Siekert RG, Piepgras DG, Sharbrough FW. Houser DW: Bypass surgery for vascular disease of the carotid system. Mayo Clin Proc 51:677-692, 1976. 29. Tew JM Jr: Reconstructive intracranial vascular surgery for prevention of stroke. Clin Neurosurg 22:264-280, 1975. 30. Tice LD, Chater NL: Microvascular bypass for stroke: Review of 203 patients. Angiology 31:6-10. 1980. COMMENTS Dr. Heros and his colleagues have provided an interesting documentation and discussion of five patients who experi- Neurosurgery, Vol. 15, No. 2 enced delayed transient postoperative neurological deteriora- tion after successful extracranial-intracranial arterial bypass. In each instance. rather extensive investigation failed to iden- tify a causative agent, although the more routine postoperative complications such as intra- or extraaxial hematoma, vaso- spasm. and graft occlusion were demonstrably not present. In their thoughtful discussion, the authors postulate that these unusual and fortunately reversible deficits may have resulted either from hyperperfusion of chronically ischemic vascular territories or from a shift in middle cerebral artery flow patterns produced by distal graft input. If the five cases are examined in a purely clinical light, three seemingly distinct patterns seem apparent. Two patients suf- fered what seemed to be exacerbation of previous transient ischemic attack patterns with complete resolution within the initial postoperative week, a worrisome but not uncommon occurrence after successful bypass operation. Two additional patients had a single rather protracted episode of neurological deterioration which, at least anatomically, involved a different or greater extent of the middle cerebral distribution than had been previously symptomatic. The deterioration of the fifth patient, operated on acutely in the face of new ischemic deficit and CT documentation of fresh infarction, is the third pattern and has been attributed by the authors to exacerbation of vasogenic edema in the area of infarction with probable associated progression of local ischemia. It is the second group of patients, those with a single protracted episode of neurological deterioration, that I find to be of most interest. In a much different but perhaps analogous clinical setting, we recently experienced similar episodes of delayed reversible postoperative deficit perhaps related to the sudden restoration of normal perfusion to chronically is- chemic areas of cortex. Four patients undergoing arteriove- nous malformation resection were studied during pre- and postoperative periods with regional cerebral blood flow deter- minations; in the preoperative study each demonstrated areas at some distance from the area of the arteriovenous malfor- mation that were hypoperfused by regional cerebral blood flow criteria, probably on the basis of a steal phenomenon. In each patient, focal neurological deficit, developing 24 to 72 hours after an angiographically documented complete arteri- ovenous malformation removal, implicated dysfunction of these cortical regions now demonstrated by postoperative cerebral blood flow studies to have increases in local flow to but not above normal levels, Other investigations (CT, repeat arteriogram) failed to identify any causative etiology for these neurological deficits, which reversed over 12- to 48-hour periods. These cases, like those reported by Heros et al. and Scott et al.. suggest that perhaps the sudden restoration of even normal flow to chronically ischemic, albeit asympto- matic, areas of brain may be attended by focal neuronal dysfunction of undetermined etiology. There are obvious parallels between these three situations and those reported by Wilson and Spetzler in their hypothesis of the “normal per- fusion pressure breakthrough” problem. These similarities offer an attractive area for further investigation. Duke Samson, M.D. Dallas, Texas We have had the opportunity to encounter a similar num- ber of patients who have developed transient neurological deficits after cerebral revascularization. Among our 320 pa- tients who underwent anterior circulation bypass operations at Henry Ford Hospital within the last 5 years, 16 had similar symptoms. In investigations of the same nature as those of Heros et al.. we found no CT evidence of recent hemorrhage August 1984 or infarction, angiographic patency with filling of the middle cerebral tree. and no occlusion of the vessel or new area of abnormality. As Heros et al. indicate, because these events develop sometime after operation. they are probably not related to a transoperative problem such as hypotension. temporary occlusion of the vessel, or transection of the small cortical perforators. We agree that an embolic source could be the cause of the problem. Our angiographic sophistication is not such that we could detect microemboli, which could originate anywhere in the extracranial circulation and which could be the cause of these symptoms. Local problems at the anastomotic site, such as small intimal tags left within the lumen of the vessel or small areas of endothelial sluffing, could be the source of local emboli. Other external sources of emboli certainly cannot be excluded if these are of a microembolic nature. The postulate of dysautoregulation with secondary hyper- perfusion in an area of previous hypoperfusion is certainly very interesting. It is conceivable that a local steal phenome- non could develop as the bypass is completed, and either an area of watershed ischemia or a definite steal phenomenon could account for the development of transient symptoms. Once cerebral autoregulation were reestablished and accus- tomed to the recent entry of new flow into the area, this symptomatic phenomenon would tend to disappear. It is, however. somewhat disturbing that, if this is a dysautoregu- latory problem, the deficits aren’t longer-lasting. It is possible that the dysautoregulatory problem could vary over time and could be more pronounced depending on the patient’s activity and the activity of the brain; however, if the patients are kept in the hospital in base line activity conditions, this may not entirely explain the problem. The other mechanism postulated by Heros, the reorgani- zation or shift of the flow by the entry of the new collateral flow brought in by the anastomosis and the possible eddy current that it could create on the orthograde flow entering DETERIORATION AFTER EC-IC BYPASS 185 from the contralateral circulation, is also interesting. One could expect that eddy currents could create regions of stasis where the two opposing pressures meet and that perhaps zones of relative ischemia could develop in these areas of stasis. Unfortunately, due to our limited ability to study the flow patterns of the intracranial circulation, we really don’t know whether this postulate is indeed a fact or whether actually both circulatory patterns complement each other. Intuitively, it would make sense that areas of microstatis may be present: however, we cannot currently answer this question. One of the possibilities that was discarded by the authors as not being the likely source of the problem is that of seizure events occurring during the postoperative period. We have found in our patients a 5% incidence of postoperative seizures. These seizure events in general tend to be focal, can be repetitive, are motor in most cases, but can also be sensory seizures or affect the speech. Even though the EEG may be negative and the initial response to anticonvulsants may not be instantaneous, it is quite possible epileptogenic area could be controlled. The theory of hyperperfusion that Heros et al. have postulated certainly would go along with this postulate. It is possible that the hyperperfusion could cause an area of relative steal, but it is also possible that this area of hyperper- fusion with local vasodilation and local mass effect could create an epileptogenic focus that could lead to the develop- ment of seizures. Until now, we have not been able to prove the focality of the seizures with EEGs. We agree with Heros et al. that dysautoregulation and hyperperfusion could be the most important reason for the development of symptoms after a cerebral revascularization procedure. In general, we concur with the authors that there is no clear-cut explanation for this problem, and we congratulate them for their clear analysis and excellent presentation of these five cases. Fernando G. Diaz, M.D., Ph.D. Detroit, Michigan