J Neurosurg 75:891-901, 1991

Disabling cerebral transient ischemic attacks PATRICK W. McCoRMICK, M.D., FR,~NK J. TOSt~:CE~, M.D., JEAN MCKINNEV, R.N., AND JAMES I. AUSMAN, M.D., PIhD.

D~Tarrmenl r?[Neuroloq,ical St#~'eO'. lh'm3' Ford Netm)sto?gical lnstilute, Detroit, Michigan u- The surgical management of an emerging clinical entity, namely disabling transient cerebral ischemic attacks, is described. A series of 19 patients treated in a 2-year period (12 with anterior circulation dysfunction and seven with posterior insufficiency) met the following criteria: 1) stereotypical recurrent episodes of transient neurological dysfunction related to the anterior or posterior circulation distribution: 2) failure of maximum medical therapy to control the transient neurological dysfunction: 3) four-vessel cerebral angiography demonstrating an isolated vascular territory corresponding to patient symptoms: 4) inhalation xenon cerebral blood flow studies with at least three of eight probe-pairs showing significant asymmetries in the initial slope index, localizing an area of relative oligcmia to the symptomatic hemisphere (anterior circulation only); and 5) severe restriction of lifestyle due to transient ischcmic attacks (TIA's). Seventeen patients unde~,ent surgical bypass therapy: deep sylvian superficial temporal artery (STA)-middle cerebral arte~, (MCA) bypass in nine; surface STA-MCA bypass in three: STA-superior cerebellar artery bypass in three: STA-posterior cerebral artery bypass in one: and aorta-carotid arte~' bypass in one. There was one perioperative death and four perioperative strokes (two ipsilateral and two contralateral to the operated side). The average follow-up period was 14 months. Of the 16 surviving surgically treated patients, 13 (81%) have had an excellent to good outcome with complete resolution of TIA's and minimal neurological deficits. Three patients had a poor outcome with eilher a significant persistent neurological deficit or continued TIA's. The two patients not treated surgically continue to have vertebrobasilar insufficiency episodes while receiving oral anticoagulation medication. The overall mortality rate (5.5%) and stroke morbidity rate (22.2%) of surgical therapy for disabling TIA's are high in this neurologically unstable group of patients, but are associated with an 81% excellent to good response. Although the natural history of disabling T1A's is not known, these patients present with significant to total disability due to their symptoms. It is concluded that disabling TIA's respond to surgical revascularization and may represent an indication for cerebral revascularization surge~'. KEY WORDS

revascularization

ischemia vertebrobasilar insufficiency extracranial-intracranial bypass

p

RiOR to 1985, patients with transient neurological dysfunction in a compromised vascular distribution of the brain were commonly treated surgically, often with extracerebral-to-intracerebral (ECIC) bypass procedures. In that year, an international cooperative study s was published that randomly assigned 1377 patients to receive "best medical" or "best medical" plus superficial EC-IC bypass therapy. The data were interpreted using the endpoints of fatal and nonfatal stroke. The results demonstrated that for the population studied and in the manner evaluated, there was no apparent benefit from superficial EC-IC bypass surgery over "best medical" therapy. Furthermore, the patients enrolled were subdivided into angiographic subgroups, none of which benefited from superficial

J. Neurosurg. / Vohone 75/De~vmber. 1991

. transient ischemic attack

EC-IC bypass surgery. The population in that study included patients who had experienced one or more clinically diagnosed transient ischemic attacks (TIA's) referable to the anterior circulation, who were enrolled within 3 months of presentation, and who had an angiographically documented ipsilateral distal internal carotid artery (ICA) or middle cerebral artery (MCA) stenosis or occlusion, s9 Despite the heterogeneous nature of the patients' clinical presentation, little effort was made to select out and study clinical (not radiographic) subgroups to determine if one category existed for which surgical therapy was beneficial. One group, including patients with three or more TIA's, was looked at independently; as an aggregate, this group did not benefit from surgery. ~ 891

P. W. McCormick, el al. A question remains whether surgical therapy in gcncral, and EC-IC bypass specifically, is of benefit to a more select clinical subgroup of patients. Patients who fail medical therapy have been suggested as constituting such a subgroup." Clinical Material and Methods

Palient Population

From November, 1987, to April, 1990, approximately 300 patients were evaluated for symptomatic cerebrovascular occlusive disease, of whom 19 fit the criteria for disabling TIA's as outlined below. Two patients who fit the criteria elected not to undergo surgery. Anterior circulation ischemia was responsible for the disabling TIA's in 12 patients (63%), and vertebrobasilar insufficiency (VBI) was present in seven (35%). Seven patients (35%) suffered at least one cerebral infarct, usually followed by fluctuating neurological symptoms. Twelve patients were surgically treated with superficial temporal artery (STA)-MCA anastomosis (involving deep sylvian MCA branches in nine), three with STAsuperior cerebellar artery (SCA) anastomosis, and one patient with STA-posterior cerebral artery (PCA) anastomosis. One patient was treated with an aorta-left common carotid artery and right subclavian artery bypass procedure. One patient required a second STAMCA anastomosis, making a total of 18 procedures in 17 surgically treated patients. Patient Selection Criteria

Five criteria were employed to select the study population. The first related to recurrent severe episodes of transient neurological dysfunction diagnosed as TIA's or VBI. The diagnosis of TIA depended on the history of the episode and the patient's clinical presentation. This series followed the definition of TIA outlined initially by the Classification of Cerebrovascular Diseases (CCVD) II and more recently by CCVD IlI. ~2' A patient is not considered to have had a TIA unless the episode had a rapid onset (less than 2 minutes) and a duration ranging from a few minutes to no longer than 24 hours without related persistent neurological deficit. A diagnosis of TIA's related to the anterior circulation required one or more of the following symptoms: 1) motor dysfunction (expressive dysphasia or paralysis/ clumsiness) of the extremities and/or face on one side; 2) loss of vision in the pattern of homonymous hemianopsia; 3) numbness, including loss of sensation or paresthesia involving the upper and/or lower extremity and/or face on one side; and 4) receptive dysphasia or aphasia (if the dominant hemisphere is involved). Diagnosis of VBI related to the posterior circulation required three of the following symptoms: 1) motor dysfunction in any combination of upper and lower extremities and face, on the left and/or right side; 2) sensory symptoms involving the left, right, or both sides; 3) loss of vision in one or both homonymous visual fields; 4) loss of balance, vertigo, unsteadiness, 892

or dysequilibrium, diplopia, perioral numbness, or dysarthria in any combination, but not when any one of these symptoms occurred alone. The second criterion required cerebral angiography (four-vessel selective injections with magnified subtracted views) demonstrating occlusion of large-caliber vessels perfusing the symptomatic vascular territory with collateral flow routes compensating. The third related to regional cerebral blood flow (CBF) (quantified by the ~3~Xe inhalation technique ~~ documenting unilateral reduced CBF in the symptomatic vascular territoD'; at least three significant interbemispheric asymmetries of the initial slope index must have been recorded (anterior circulation only). Failure of aggressive medical management, as outlined below, to decrease or ameliorate the patient's TIA's or VBI was the fourth criterion, and the fifth was medical inability to work and severe restriction of lifestyle caused by TIA's. Patient Management The evaluation of each patient included a computerized tomography (CT) scan and/or magnetic resonance (MR) imaging. These diagnostic methods were used to document the presence of an infarct and to rule out a hemorrhage or mass lesion. Admission studies included complete differential blood cell and platelet counts, a coagulation test and blood profile, urinalysis, chest x-ray studies, and serum electrolyte profiles including glucose and fasting lipid levels. All patients received a cardiac work-up consisting of at least an electrocardiogram, an echocardiogram, and 24-hour ambulatory (Holler) monitoring to evaluate cardiac function and to identify coexisting cardiac problems. If a high suspicion of cardiac etiology was present, a transesophageal cardiac echogram was included. Patients with a negative cardiac work-up underwent four-vessel cerebral angiography, utilizing magnification and subtraction techniques to assess collateral circulation. Angiography also documented the presence ofaneurysms or arteriovenous malformations (AVM's), which represented relative contraindications to blood pressure manipulation and anticoagulation. If the angiograms showed normal cerebral vasculature, an interdisciplinary work-up was performed to rule out other causes of TIA's, VBI, or cerebral infarction. Regardless of the angiographic pattern, patients with TIA's or VBI were initially treated medically. Medical therapy included administration of antiplatelet and/or anticoagulant agents, blood-pressure manipulation, phlebotomy, and hospitalization for volume loading. This protocol for medical management of TIA's was similar to that outlined elsewhere. ~2~3~5 Aspirin (325 rag) was given daily; anticoagulation medication was considered therapeutic when partial thromboplastin time or prothrombin time was 1.5 to 2.0 • control values. Medications to control blood pressure and dosage schedules were manipulated to avoid episodes of relative hypotension (diastolic pressure < 85 torr). PhleJ. Neurosurg. / Volume 75 / December, 1991

D i s a b l i n g cerebral t r a n s i e n t i s c h e m i c a t t a c k s

FIG. 1. Flow chart showing the treatment paradigm used to evaluate and treat patients with persistent transient ischemic attacks (TIAs). CT = computerized tomography; MRI = magnetic resonance imaging; CBF = cerebral blood flow: BP = blood pressure.

botomy was used to maintain hematocrit in the low normal range, and fluid loading was effected with saltpoor albumin and hemodynamic monitoring. Surgery was only considered if symptoms were progressive and disabling in spite of several weeks of medical management. The type of surgery was determined by the anatomy of the donor and recipient vessels. The anesthesia and surgical techniques for anterior and posterior EC-IC cerebral bypass have been described elsewhere. 23~ The management protocol for the patients described here is outlined in Fig. 1. Evaluation of Patient Outcome

A patient who was neurologically intact with complete resolution of TIA's was considered to have had an excellent outcome. A good outcome was defined as a marked reduction in the number of TIA's and/or a minimal neurological deficit. A poor outcome was defined as persistence of TIA's and/or significant postoperative neurological deficit.

Illustrative Cases Case 1

This 58-year-old man had retired from a military career and worked as an independent consultant. He had essential hypertension controlled with medication J. Neurosurg. / Volume 75 / Da'ember, 1991

and a history of tobacco use. In 1989, he experienced a series of left hemispheric TIA's manifesting as "confusion," expressive and receptive dysphasia, and right hemiparesis. This culminated in a left hemispheric cortical infarct with a mild residual expressive dysphasia and right hemiparesis. He was placed on a course of aspirin (325 rag/day) and Persantine (dipyridamole) (50 nag three times daily). An intra-arterial digital subtraction angiogram demonstrated bilateral ICA occlusions. Three months later, the patient presented for neurosurgical evaluation because of episodic left hemispheric events manifested as dysphasia (receptive greater than expressive) and numbness of the fight hand and face. At this time, the neurological examination demonstrated fine-motor weakness on the right, no cortical sensory deficits, and mild dysphasia (expressive greater than receptive). Cerebral angiography disclosed bilateral ICA occlusions with bilateral small posterior communicating arteries (PCoA's) (Fig. 2). The predominant vascular supply to the anterior circulation was through pial and periorbital collateral vessels (Figs. 2 and 3). The antihypertensive medication regimen was changed, and the patient stabilized with infrequent TIA's over a period of 289 months. When his attacks increased in frequency, he stopped working and confined himself to his house. Oral anticoagulant drugs did not improve the clinical 893

P. W. McCormick, el al.

Flo, 2. Case 1. L~Ji:Composite drawing summarizing the preoperative cerebral angiograms. The internal carotid arteries (ICA's) are occluded bilaterally and the posterior communicating arteries arc small in caliber. The ICA's are perfused by periorbital branches of the external carotid arteries (arrow~s). The verlebral arteries demonstrate origin stenosis on the right (RVS) and left (LVS): 60% and 80%, respectively. ('enlerand Rigtll: Preoperative left (center) and right (right/common carotid angiograms, lateral views, showing opacification of the ICA's via collateral vessels.

picture, and a xenon CBF study demonstrated significant asymmetry in five of eight probes lateralizing to the left hemisphere. The patient was hospitalized, and hypervolemic therapy reduced the intensity, but not the frequency, of his TIA's. He then underwent a left STAdeep MCA bypass procedure, and postoperative angio-

FIG. 3. Case 1. Left vertebral angiogram, lateral view, demonstrating collaterals from the posterior cerebral artery, predominantly through small pial vessels (pial collaterals). 894

grams demonstrated excellent flow through the bypass (Fig. 4). He has not experienced further TIA's, and is currently leading an active life outside his home; his neurological condition remains unchanged. Case 2 This 47-year-old woman had a medical history of hypothyroidism and essential hypertension, both well controlled with medication. She had a history of tobacco use. Initial fasting blood work demonstrated a total cholesterol level of 319 mg/dl. In 1988, she developed a sudden right hemispheric cerebral infarct manifesting as persistent mild left hemiparesis involving the face and upper extremity more than the lower extremity. Her physical examination demonstrated 4+/5 leftsided hemiparesis with normal cortical sensation and no visual field deficits. An angiogram demonstrated right ICA occlusion with collateral flow via periorbital vessels on the right, the anterior communicating artery, and pial collateral flow from the right PCA (Figs. 5 and 6). A course of aspirin (325 mg/day) and Persantine (50 mg orally three times daily) was initiated and she was discharged. Two months later, the patient experienced right hemispheric YIA's manifesting as increased left hemiparesis and left upper-extremity numbness. In October, she developed a high fever with profuse vomiting, and after 2 days of self-treatment she developed a severe left hemiparesis which lasted several hours. She underwent rehydration therapy at her local hospital, and her deficit returned to baseline. The patient continued to have right hemispheric J. Neurosurg. / V~dume 75/December, 1991

Disabling cerebral transient ischemic attacks

FIG. 4. Case 1. L~[i. Composite drawing summarizing tile postoperative cerebral angiograms. An additional source of collateral flow to a deep middle cerebral branch has been created. Center and Ri,ghl: Postoperative left common carotid angiograms, anteroposterior (center) and lateral fri,ght) views, showing rapid and extensi,;e filling of the middle cerebral and anterior cerebral arteries.

TIA's and was placed on Coumadin (crystalline warfarin sodium) therapy. This did not change the pattern of her TIA's which became increasingly severe, occurring several times per day. She placed herself on strict bed rest because she fell with almost every TIA and had sustained several large bruises. At this time, a xenon CBF study demonstrated significant oligemia of the right hemisphere, and a sur-

face STA-MCA bypass was performed. There were no surgical complications, and postoperative angiography demonstrated patency of the anastomosis with feeding of the middle cerebral vessels on the right (Fig. 7). Postoperatively, the patient was placed on a course of aspirin (325 mg/day) and has not experienced any further TIA's; however, a mild left hemiparesis persists.

FIc. 5. Case 2. Left: Composite drawing summarizing the preoperative four-vessel angiograms. The right internal carotid artery (ICA) is occluded and the right posterior communicating artery, is not visualized. The anterior communicating arte~ (ACoA) is the major collateral vessel. Center: Left common carotid angiogram, anteroposterior view, demonstrating collateral flow from the ACoA to the right hemisphere. Right: Right common carotid angiogram, lateral view, demonstrating ICA occlusion with periorbital collaterals. J. Neurosurg. / Volume 75 /December, 1991

895

P. W. McCormick, eta/.

Fl~;. 8. Preoperative axial magnetic resonance images in Case 3. The T~-weighted image Neft) demonstrates a hypointense left paramedian pontine lesion which is hyperintense on T2-weighted images (right). This is most consistent with an infarct. Note the absence orflow void in the basilar artery. FIG. 6. Selective preoperative left vertebral angiogram, anteroposterior view, in Case 2 demonstrating no significant collateralization from the posterior cerebral or posterior communicating arteries.

CHSC 3 This 48-year-old man was employed as a paramedic. He had a history of tobacco use, hypercholesterolemia, and essential hypertension. In April, 1989, he suffered the sudden onset of objective vertigo, right hemiparests, and dysarthria. He gave a history of similar transient events related to "smoke inhalation" during workrelated activity. This event failed to clear, and a CT scan of the head demonstrated a pontine infarct which was further characterized by MR imaging (Fig. 8). He

underwent angiographic evaluation which suggested basilar thrombosis. He was placed on a course of aspirin (325 mg/day), underwent rehabilflation, and was referred for neurosurgical evaluation. At the initial evaluation, he described transient episodes of vertigo, hemiparesis alternating right-to-left, and occasional paraparesis. Repeat studies of the vertebral arteries failed to visualize the fight vertebral artery and showed distal occlusion of the left vertebral artery and a high-grade midbasilar stenosis (Fig. 9). The proximal basilar artery distribution was perfused by the anterior spinal artery and the distal basilar artery distribution was perfused by cerebellar pial vessels from the posterior inferior cerebellar arteries (Fig. 9). In addition, a 2.0 x 2.0-cm AVM was identified in the left occipital region. Cardiac evaluation was unremarkable. Coumadin therapy was not considered because the patient had a history of gastrointestinal bleeding and because of the coincidental AVM. His antihypertensive medication was adjusted to avoid relative hypotension. For several months he appeared to be stable, but he had imposed some limitations on his activity to prevent VBI symptoms. He spent the majority of his time recumbent, and was unable to work or pursue outside interests. Eight months after initial presentation, the patient began to have almost daily VBI episodes during sedentary activity. An STA-PCA bypass procedure was performed without complication, and the postoperative angiogram demonstrated patency of the bypass (Fig. 10). He has been free of VBI attacks since.

Results FiG. 7. Case 2. Left: Composite drawing summarizing the postoperative cerebral angiography. A patent superficial temporal artery (STA) collateral to a deep middle cerebral artery (MCA) has been created on the fight. Right: Postoperative right common carotid angiogram, anteroposterior view, demonstrating a patent STA collateral to a superficial MCA branch. 896

This series consisted of 12 men and seven women with a mean age of 57 years. Seven patients had left hemispheric events and five had fight-sided ischemic events. The remaining patients had VBI. Ten patients had extracranial disease in two vessels; the most common pattern involved bilateral ICA occlusion with abJ. Neurosurg. / Volume 75 / December. 1991

Disabling cerebral transient ischemic atlacks

Fl(~. 9. Case 3. L~![t: Composite drawing SLnnmarizing the preoperative cerebral angiograms. Right vertebral artery occlusion and the left vertebral artery ending in the posterior inferior eerebellar artery (PICA), which collateralizesto the basilar arte~', can [~ seen. The distal intracranial vertebral arteries and vertebrobasilar junction are supplied by the anterior spinal artery. Therc is midbasilar stenosis, and no posterior communicating vessels are visualized. The PICA supplies the pial collateral vessels. Not shown is a 2.0 x 2.0-cm left occipital arteriovenous malformation. Center and Right: Preoperative left subclavian angiograms, anteroposterior (center) and lateral (right) views, demonstrating a left vertebral vessel ending in the PICA with collateral flow to the distal inlracranial vertebral artery from the anterior spinal artery.

sent or hypoplastic PCoA's. One patient had disease in all four extracranial arteries, with three occluded vessels and one demonstrating preocclusive stenosis. The remaining patients were evenly distributed between those with disease in one or three extracranial vessels. Collateral flow from pial vessels was identified in 58% of the cases.

Eve~' patient was placed on a course of aspirin or aspirin plus Persantine, and these regimens were associated with one case of gastrointestinal bleeding. Only three patients were not given long-term anticoagulants because of a contraindication to this therapy. One of these had a basilar tip aneurysm, another an AVM, and the third an upper gastrointestinal bleed. Ten patients

FIG. 10. Case 3. Left: Composite drawing summarizing the postoperative cerebral angiograms and demonstrating a surgically created collateral vessel from the external carotid artery, to the posterior cerebral artery (PCA). Center and Right: Postoperative selective right external carotid angiograms, anteropostefior (center) and lateral (right) views, demonstrating the large superficial temporal collateral to the PCA.

J. ~2,urosurg. / Volume 7.5 /December. 199l

897

P. W. McCormick, el a/. TABLE

1

Opcralive +>u~'comeo[ I8 DrO~cdur('.~ il7 17 palie~ls lrt'olt'd./i~r lra~L~i~'~l iscllenTic allackv* Procedure STA-MCA bypass (deep) STA-MCA bypass'{

No. o1" Cases

Bypass Patcnc~

Deaths

9 4

8 (g9%) 4

1 (11%) 0

Stroke Morbidity Ipsilateral Contra[ateral 0 0

I {11%) 1(25%)

Recurrence of Symptoms

Excellent/Good Outcome

I (11%) 1(25%)

8(88%) 4(100%)

(surtilce) STA-SCAbypass 3 3 0 2 (66%) 0 0 I (33%) STA-PCA bypass 1 1 0 0 0 0 I aorta-carotid & aorta-subcla1 1 0 0 0 0 1 vian artery bypass totals 18 t7 (94.4%) I (5.5%) 4 (22.2%) 2 (11.1%) 13 (72.2) * STA-MCA = superficial temporal artery-to-middlecerebral artery: STA-SCA = STA-to-superiorcerebellar artery.: STA-PCA = STA-toposterior cerebral artery. One patient underwent two surface STA-MCAbypassprocedures.

(53%) had their medication revised to allow their blood pressure to rise in an attempt to relieve their ischemic spells. Two patients had therapeutic phlebotomies and two were hospitalized for hemodynamic therapy. Postoperative angiography was obtained in all 17 patients who underwent EC-IC bypass surgery. An angiographically patent bypass was seen in eight (89%) of the nine STA-deep MCA anastomoses (Table 1) and in all four surface STA-MCA anastomoses. All bypass procedures for posterior circulation disabling TIA's were angiographically patent postoperatively. The overall patency rate for the EC-IC bypass procedures performed was 94.4%. The average long-term follow-up period for this series was 14 months (Table 1). Of the 16 surviving surgically treated patients, 13 (81%) have had a good to excellent outcome. Eleven patients (71%) with good or excellent outcomes had complete resolution of TIA's with no limiting neurological deficit. Of the two others with a good or excellent outcome, one had marked reduction of TIA's with good neurological function and the other developed a mild dysphasia. One patient had persistent TIA's in the hemisphere contralateral to the initial bypass despite a patent anastomosis; a second operation consisting of a contralateral bypass led to complete resolution of symptoms. Three patients (19%) had a poor outcome. One patient with an initially good outcome had a sudden recurrence of TIA's associated with delayed occlusion of his bypass; this was the only patient with recurrent symptoms in the series. One patient had a poor outcome, with resolution of his VBI but significant persistent cerebellar dysfunction secondary to perioperative infarction. One patient had a poor outcome with resolution of VBI, but suffered a brain-stem infarct despite a patent STA-SCA bypass. There was one perioperative death for a mortality rate of 5.5%; four (22%) perioperative strokes were recorded (two ipsilateral and two contralateral to the 898

surgically treated side). Two strokes caused minimal residual deficits associated with a good outcome, and two caused significant deficits associated with a poor outcome. Other postoperative complications included one transient fourth cranial nerve palsy, one chronic subdural hematoma, two wound infections, one postoperative angina, and one case of meningitis. On longterm follow-up review, three patients died of causes unrelated to ischemic stroke: one had a myocardial infarction, one a ruptured basilar aneurysm, and one pneumonia. The two nonsurgically treated patients continue to have frequent episodes of VBI while receiving oral anticoagulant medication.

Discussion Disability from TIA's is an emerging clinical entity. Part of its recognition has been fostered by the general acceptance of the international EC-IC Bypass Study Group's conclusions. 8 Although the conclusions of this study are internally consistent and appropriate for the population sample studied, they are easily misinterpreted as demonstrating no indication for surgical augmentation of collateral CBF in the entire universe of patients with symptomatic cerebral ischemia and major extracranial vessel occlusion. The failure of some specific angiographically identified subgroups and positron emission tomography subgroups to benefit from surgery has generally been interpreted to mean that the surgery is of no benefit to any subgroup of patients) '8"~sA more appropriate conclusion would be that no diagnostically identified subgroup of patients has shown benefit from surgery over an established medical management protocol. This conclusion begs the question of whether any subgroup exists that does benefit from surgery, and suggests that patients who fail to benefit from an aggressive medical management protocol, as indicated in this report, might be identified as such a subgroup. 6 This work attempts to address this issue by retrospecZ Neurosurg. / l~#ume 75 /December, 1991

Disabling cerebral transient ischemic attacks tively evaluating a group of 19 nonconsecutive, highly selected patients with well-defined TIA and VBI syndromes. Each patient was carefully evaluated with diagnostic studies and failed to respond to an aggressive medical management protocol. In addition, each patient reached a point where his/her ischemic attacks became severely disabling. This group of medically refractory TIA patients were almost all treated by surgery, and their response to surgery offers insight into the potential of surgical treatment. Study Desig~ Critical analysis of this work begins with the fact that the data are not prospective, and the natural histow of this selected group of patients for stroke and death is unknown. The lack of prospective data is appropriate for the preliminary nature of these results but weakens the conclusion concerning surgeu, favorable or unfavorable, based on these data alone. The lack of natural history data on stroke and death makes the impact of surgery difficult to judge. These patients had progressed to complete disability prior to surgery, and we suggest that this is as valid an endpoint as stroke. Patient Evahtation Evaluation of the diagnostic studies used raises several important issues. The hemodynamic consequences of the cerebral vascular lesions identified angiographically in each patient are not predictable. An individual with a single collateral vessel supplying a hemisphere may either be asymptomatic, develop TIA's, develop disabling TIA's, or have a cerebral infarction. The impact of an adequate collateral blood supply on the natural history of ICA occlusion is firmly established. Surgical experience with blood flow confirms the importance of collateral flow. 5"14"17"22 An asymptomatic hemisphere fed only by collateral supply is not an uncommon clinical finding. Therefore, the angiographic data are only meaningful when taken together with additional diagnostic and clinical information. However, it is interesting to note the frequency with which disabling TIA's were associated with a large vascular territory, collaterally supplied by pial vessels. This occurred in 58% of our patients, and may represent an important sign of hemodynamic inadequacy in the supporting collateral system. The use of noninvasive CBF studies such as those involving the xenon technique is common. There is, however, limited information supporting the clinical utility of such measurements. In fact, it has been observed with inhalation xenon CBF measurement that a predictable pattern of reduced flow is seen with ICA occlusion irrespective of the patient's symptoms.~~ Significant asymmetry of blood flow measured between the two hemispheres has not been shown to add additional sensitivity to cerebral angiography as a measure of stroke risk. Nevertheless, the asymmetry of flow seen in our patients with anterior circulation disease suggests J. Neurosttrg. / Volume 75/December, 1991

that the angiographic disease identified is associated with oligemia. Of great importance in our approach to persistent T[A's or VBI are the issues of disability and incapacitation. The patient must be evaluated as an individual, not as a statistic, and every reasonable effort should be made to relieve the suffering of an individual unable to work and incapacitated in routine activity. Accordingly, the determination of work disability made by the attending physicians and the determination of incapacitation made by the patient figured heavily into the decision to operate. Medical Treatment The issue of medical treatment is controversial, and there are numerous reports describing medical management for cerebrovascular disease. ~2.~s Grotta'2 described the reduction of risk factors and the administration of aspirin (325 to 1300 rag/day) and anticoagulant drugs as therapy for prevention of stroke after TIA's or stroke. The concept of "best medical therapy" is poorly defined, but in the international EC/IC Bypass Study, ~ it included risk factor reduction and aspirin. The "aggressive" medical management our patients received is consistent with the published approaches to treating post-TIA patients and those with worsening TIA's. '2"~5 It is fair to say that our patients failed to benefit from a consistently aggressive medical approach to persistent TIA's, and "best medical" management might include less familiar treatment regimens. Embolic vs. Hemodynamic Transient lsehemic Attacks Recurrent TIA's generally have one of two pathophysiological mechanisms: hemodynamic or embolic. The use of surgical therapy to create additional collateral circulation is a treatment for a hemodynamic cause rather than an embolic source of cerebral ischemia. It is important to separate patients with cerebral ischemia into hemodynamically or embolically caused groups based on the pathophysiology of their TIA syndromes. This differentiation was achieved in our study by using multiple diagnostic tests with strict criteria for patient entry into the disabling TIA group. First, the echocardiogram and 24-hour cardiac monitoring had to be within the normal range, thus screening for those with a common source of embolic events. Second, close correlation of the angiographically compromised circulation, the xenon/CT-delqned area of oligemia, and the patient's symptoms had to be achieved. This correlation is most consistent with hemodynamic compromise as the etiology of TIA's, and accounts for their recurrence in a single vascular distribution. Embolic disease would be more likely to cause ischemia in variable vascular distributions. Finally, the failure of the described patients to respond to aggressive medical therapy for embolic disease with antiplatelet and anti899

P. W. McCormick, el al. coagulant drugs supports the hemodynamic basis of their TIA's. Surgical Trealmenl

The surgical therapy used in these cases was tailored to the individual patient with the goal of supplying adequate collateral circulation at minimum risk to the patient. Generally, a "deep" STA-MCA bypass is performed to a large MCA branch in the proximal sylvian fissure] This procedure supplies a high volume of collateral blood flow with excellent long-term patency. A "superficial" bypass to a cortical MCA branch was performed when a ~aeep bypass was technically not feasible. The VBI patients were treated with STA-SCA or STA-PCA bypass procedures, Each type of bypass resulted in a high volume of collateral blood flow. One patient received an aorta-common carotid and subclavian artery Y bypass because of very complex proximal occlusive disease. The single perioperative death was related to panhemispheric infarction of the surgically untreated asymptomatic side 8 hours postoperatively. Occlusion of a pre-existing 98% petrous carotid artery stenosis on the untreated side was implicated by angiography. Stroke morbidity involved contralateral cerebral infarction as well, These events reflect the complex flow response in the circle of Willis and cerebral conducting vessels to a bypass. Infarcts following STA-SCA bypass surgery occurred in two patients. This complication rate is disproportionately high compared to our larger series. 4 A nonstrokerelated morbidity rate of 35% was limited to transient treatable problems, all of which resolved without sequelae. The combined mortality and stroke morbidity rate for surgical therapy of disabling TIA's in this series was 27.5%, and the combined mortality and ipsilateral stroke morbidity rate was 16.5%. These figures reflect a significant risk associated with surgical intervention, but should be interpreted together with a long-term outcome of resolution of symptoms and return to premorbid activity levels (good or excellent outcomes) in 81% of cases. The long-term follow-up data presented here demonstrate that the efficacy of surgery, is lasting in 94% of patients with a mean follow-up period of 14 months. It is important to emphasize that, when surgical therapy is delayed in patients who are neurologically unstable, the patient becomes a poorer surgical risk and has a higher complication rate. This fact has been reported in our series of patients with posterior circulation ischemia undergoing bypass surgery. 4 and by Sundt, et al., 2~ in their experience with carotid endarterectomy. Conclusions These data suggest that a subgroup exists of patients who fail to respond to aggressive medical therapy. A series of 11 such patients was treated surgically by Meyer, et al. (FB Meyer, personal communication, 900

1989). The data presented here suggest that the risk of surgical therapy may be justified by the effectiveness of the operation, but this conclusion is preliminary, and natural history data are necessary to evaluate it further. On the other hand, the disabled condition of the patients described represents an ethical barrier to a natural history study. We conclude that surgery is an effective therapeutic alternative for patients who meet the criteria of disabling TIA. The morbidity associated with surgery is high in this neurologically unstable group of patients, but their mortality rate is low. Long-term follow-up results suggest that the benefits of surgery are lasting and that patient lifestyle is improved by it. References 1. Ausman JI, Diaz FG: Critique of the extracranial-intracranial bypass study. Surg Neurol 26:218-221, 1986 2. Ausman JI, Diaz FG, de los Reyes RA, et al: Posterior circulation revascularization. Superficial temporal artery to superior cerebellar artery anastomosis. J Neurosurg 56: 766-776, 1982 3. Ausman JI, Diaz FG, Vacca DF, et al: Posterior fossa revasculafization, in Schmidek HH, Sweet WH (eds): Operative Neurosurgical Techniques. Indications, Methods and Results, ed 2. New York: Grune & Stratton, Vo[ 1, 1988, pp 807-818 4. Ausman JI, Diaz FG, Vacca DF, el al: Superficial temporal and occipital artery bypass pedicles to superior, anterior inferior, and posterior inferior cerebellar arteries for vertebrobasilar insufficiency. J Neurosurg 72: 554-558, 1990 5. Castaigne P, Lhermitte F, Gautier JC, et al: Internal carotid artery occlusion. A study of 61 instances in 50 patients with post-mortem data. Brain 93:231-258, 1970 6. Day AL, Rhoton AL Jr, Little JR: The extracranialintracranial bypass study. Surg Neurol 26:222-226, 1986 7. Diaz FG, Umansky F, Mehla B, et al: Cerebral revascularization to a main limb of the middle cerebral arter3, in the Sylvian fissure. An alternative approach to conventional anastomosis. J Neurosurg 63:21-29, 1985 8. EC/IC Bypass Study Group: Failure of extracranial-intracranial arterial bypass to reduce the risk of ischemic stroke. Results of an international randomized trial. N Engl J Med 313:1191-1200, 1985 9. EC/IC Bypass Study Group: The international cooperative study ofextracranial/intracranial arterial anastomosis (EC/IC bypass study): methodology and entry, characteristics. Stroke 16:397-406, 1985 10. Ewing JR, Robertson WM, Brown GG, et al: mXenon inhalation: accuracy in detection of ischemic cerebral regions and angiographic lesions, in Wood JH (ed): Cerebral Blood Flow. Physiologic and Clinical Aspects. New York: McGraw-Hill, 1987, pp 202-218 11. Ewing JR, Welch KMA, Robertson WM, et al: A probe by probe identification of focal cerebral ischemia using the ~33~Xenon inhalation technique. J Cereb Blood Flow Metab (Suppl) 3:586-587, 1983 12. Grotta JC: Current medical and surgical therapy for cerebrovaseular disease. N Engl J Med 317:1505-1516, 1987 13. Hopkins LN, Budny JL: A rational surgical approach to vertebrobasilar ischemia, in Spetzler RF, Selman WR, J. Neurosurg. / Volume 75 / December, 1991

Disabling cerebral uaiasicnt ischcmic attacks

14, i5, 16. 17, 18.

19.

Ca:ler LP. ct al icdsi: Cerebral Roascularization for Stroke, New York: Thieme-Strau.m. 1c185. pp 490-5()4 Landoll AM. Millikan CH: Pa,hogencsis of cerebral infarciion secondary to mcchanic:~l c:~rol!d amcr) occlusion Stroke 1:52-62. ]L)7(i ke'~ine SR: Acute cerebral ischcnli:~ in a criticcd care unit, A review of diagnosis and management..~.rch Intern Med 149:90-98. 1989 Millikan CH. Bauer RB. Goldschmidt J. et al: X classd'ication and outline of cerebro,ascular disease lI. Stroke 6:555-61(~, /975 Nicholls SC, Bergelin R, Strandness DE: Neurologic ',equelae o!" unilalcrat carotid arter~ occlusion: immediate and late, J Yase Surf{ 10:542-548, 198") Powers WJ, Grubb RL Jr. Raichle \IE: Clinical results of extracranial-imracrania', bypass surge~ in patienvs with hemodynamic cerebro\aseular disease, d Neurosurg 71~: 61-67, ]989 Sadasivan B. Diaz FG, Ausman JI: Reconstructive ~as-

J. Neurosure. / I'o/ume 75 /Dc'ce~Tber, /991

,.'JlaI ~,U:'g~.'l~,.Curr 1'her \eurol Surg 2:147-150. 1989 21,. Stmdt TM Jr. S'.uldok BA, Whisnant JP: Carotid endartcrectom,:. Connplications and preoperative assessment of risk. 31a)o (fin Proc 50:301-306. 1975 21 VChisnanl JP. gaslbrd JR. Bernstein EF, ctal: Classifica"ion ef c'ercbro~ascular disease lII. Special report from zhe National Instilute of Neurological Disorders and Szrokc. Stroke 21:637-676, 1990 22. ~,Vhisnam JP. Sandok BA. Sundt TM Jr: Carotid endarterectom) t\)r unilateral carotid system transient cerebral ischemiu. 3,1a.vo Clin Proe 58:I71-175. ~983

Manuscript received December I, I990. Accepted in final form April 29, t991. Addre.~.r reprinl rectucs~,~ lo. James 1. Ausman, M.D.. Ph.D., Depa~ament of NeurosurgeD', University of Illinois College of Medicine. M/C 799 912 South Wood Street, Chicago, Illinois 60612

901