J Stroke Cerebrovasc Dis
© 1994NationalStroke Association
c. Miller Fisher, M.D.
Remarks are made on various stroke topics-potential hazards in the early diagnosis of occlusion of the carotid, basilar, and middle cerebral arteries, dissection, discontinuing angicoagulants, lowering blood pressure, and delaying endarterectomy. The commentary also includes bolus versus continuous drip heparin, long-term warfarin therapy, extensive craniectomy for brain edema, the percentage narrowing of the carotid versus the residual diameter, stroke during anesthesia, the duration of a TlA, TlAs in congophilic angiopathy and the interpretation of transient spells of vertigo. Key Words: Cerebrovascular espial-Stroke.
The invitation to address this august assembly of stroke experts of the world should call for a highflown dissertation that lifts the endeavors of all. However, my working days ended several years ago, and my activities since then have been largely limited to observing the stroke scene. Therefore, I would like to present an account of this cerebrovascular espial, touching on a series of stroke topics and commenting very briefly on points in diagnosis, treatment, and nomenclature. To begin with, I would like to draw attention to some clinical situations that turn out to be more treacherous than might have seemed the case when the patient was first examined. For example, a minor neurologic deficit may seem clinically benign but unexpectedly prove disastrous. I would call them the "hidden shoals of stroke management." They are not rare, and we have all experienced them. 1. The first of these is the occurrence of a minor stroke associated with recent occlusion of the internal carotid artery when there have been no transient isch-
emic attacks (TIAs) or spells of transient monocular blindness (TMB), and no carotid bruit is present. The minor deficit may be attributed to a small embolus or lacunar infarct, not requiring prompt measures. A major deficit may be present the next day because of further embolism from the distal stump or failure of blood flow as the result of carotid obstruction. The only precaution is to keep the possibility in mind. When the dominant hemisphere is involved, special vigilance is in order. 2. A similar treacherous situation may exist in the first persistent deficit in imminent basilar artery thrombosis. A minor hemiparesis that appears disarming turns into a locked-in syndrome 12-24 h later. This has been termed the herald hemiparesis of basilar artery thrombosis (1). The initial deficit in basilar thrombosis is highly variable and defies easy definition in the absence of TIAs. Once a deficit appears in basilar artery thrombosis, a disastrous advance will occur in 24 h or less in 50% of cases, if untreated (2). Heparin is needed.
From the Neurology Service, Massachusetts General Hospital, Boston, MA, U.S.A. Address correspondence and reprint requests to Dr. C. Miller Fisher at Neurology Service, Massachusetts General Hospital, Fruit Street, Boston, MA 02114, U.S.A. This article is the Soriano Lecture presented at the Second World Congress of Stroke, Washington, DC, September 1992.
The most common cause is middle cerebral artery embolism; much less frequently, intracranial carotid dissection is the source. The basic deficit is a pure motor hemiparesis to which there is usually added some slight evidence of cortical involvement-sdysphasia, sensory loss, confusion, constructional apraxia, etc. This clinical picture has been called pure motor
3. Large siriaiocapsular infarct-giant lacunar infarct.
J STROKE CEREBROVASC DIS, VOL. 4, NO.1,
hemiparesis plus (PMH plus) (3). In the 24 h after presentation, there may be added a frank disabling cortical deficit. 4. Carotid and vertebral dissection. Dissection still goes unrecognized in a good many cases despite widespread publicity concerning the condition. A neurologic deficit is all the more disastrous, since patients are generally under 50 years of age. It is estimated that early recognition of carotid dissection will provide an opportunity for preventing a stroke due to embolism by using heparin, in 18% of cases (personal observation). In vertebral dissection, the early use of heparin may prevent embolism to the posterior cerebral artery (PCA), distal basilar artery (BA), or ischemia in the territory of the posterior inferior cerebellar artery (PICA) and lower BA. Headache is a prominent feature, and the most common error is to make a diagnosis of atypical migraine or occipital neuralgia. Magnetic resonance imaging (MRI), the most reliable diagnostic measure, is not widely available as an emergency procedure. In a recent instance, a man aged 51 years developed pain below the angle of the jaw and above the right eye on putting his head back to instill eyedrops. Forty-eight hours later, a severe crippling left hemiplegia ensued. The perverseness of disease has caused just the reverse. A man aged 23 years awakened with a severe right frontal headache and Horner's syndrome. Carotid dissection seemed likely and was actually confirmed by angiography. MRI, however, demonstrated pituitary apoplexy. Vertebral dissection must be kept in mind when the victim has been working with his neck extended as in painting a ceiling, making repairs under a car, or working in the confined space of an attic. In ischemic cerebrovascular disease (CVD), I pay special heed to the presence of headache as an indicator of large artery involvement. 5. Another hazard in stroke management is thediscontinuance ofanticoagulation. Stopping anticoagulants is usually tolerated, but misadventure must be a source of uneasiness. A 70-year-old woman with symptomatic stenosis of the left carotid siphon had been asymptomatic for 1 year while receiving warfarin. Preparatory to cystoscopy, the drug was discontinued 48 h before the procedure. Twenty-four hours afterwards, a severe right hemiplegia appeared, from which the patient died. Pathological examination showed recent thrombus superimposed on old thrombus related to atherosclerotic stenosis. A 76year-old man with atrial fibrillation had received warfarin for 6 years. Preparatory to abdominal surgery, it was discontinued 3 days before operation. Three days postoperatively, a severe right hemiplegia occurred, from which he died. Pathologic examination
showed a middle cerebral artery embolism. A patient who was receiving long-term warfarin for middle cerebral artery thrombosis was advised to discontinue it. Six days later, she died suddenly, and autopsy disclosed a recent myocardial infarction. A patient with frequent TIAs associated with severe carotid stenosis was receiving heparin intravenously preparatory to angiography. Heparin was discontinued 8 h before the procedure, and on arrival in the Department of Radiology the patient was found to have a severe hemiplegia. A patient with a left PCA embolus and moderate deficit had been receiving heparin intravenously for 15 days. It was discontinued preparatory to angiography, which showed no obstruction. In the next 4 days, the patient's deficit advanced to a full dominant PCA syndrome that was totally disabling. Autopsy disclosed the infarct whose circulation must have been maintained by the heparin. Even the process of switching from heparin to warfarin therapy can result in the reappearance or progression of the cerebral deficit. These are examples of events that are regularly cropping up. What can be done? Personally, a yellow flag goes up each time a decision must be made, and special consideration is given to the possibility that disaster lurks. One might draw from this series of cases the strong inference that anticoagulants can be an effective therapy. On my desk in my early years I kept a list of mishaps to remind me not to let the same mistake happen again. IfI had such a list today, items 1-5 would surely be included. Identifying these hidden shoals and singling them out for special attention cannot help but serve a good purpose in reducing our errors to the zero level. The task of making the correct decisions underscores what is obvious, namely that the field of cerebrovascular disease clearly lies in the domain of neurology. Moreover, it can be a highly expert subspecialty within neurology. There are, in addition, less portentuous pitfalls in stroke management.
6. Inadvertent lowering oftheblood pressure in ischemic stroke. Experience provides ample evidence that decreasing the blood pressure may increase the neurologic deficit, whereas elevating the pressure may decrease the deficit. I would suggest that an initial hypertensive systolic pressure of215 mm Hg or less is better left untreated.
7. Delaying carotid endarterectomy because of a stroke deficit. Some surgeons postpone endarterectomy in
this situation, but Ojemann et al. (4) have demonstrated the advisability of proceeding with operation if the deficit is only slight to moderate in severity, has occurred in the previous 48 h, and angiography shows
JSTROKE CEREBROVASC DIS, VOL. 4, NO.
C. MILLER FISHER
some patency of the lumen. The reversal of progression can be most dramatic. The notion that surgery may cause a pale infarct to become hemorrhagic is unfounded. The proper management of a stroke in progress can be a formidable challenge. The greater challenge is prevention of the stroke in the first place. The pathogenesis of most strokes is understood and preventive measures, short-term and long-term, are highly feasible-control of blood pressure, antiatheroma measures, anticoagulants for atrial fibrillators, smoking cessation, listening for carotid bruits, heeding TIAs, prompt treatment of early stroke symptoms, etc. At medical grand rounds, I recently made the pronouncement to internists that, nowadays if a patient has a stroke, his physician must be regarded as at fault. Although the statement was an exaggeration, it placed the situation in the correct perspective, and it did not fall on deaf ears. The same applies in other situations as well. For example, if a patient with carotid disease later dies suddenly from myocardial infarction, was his care suboptimal? 8. TI,e prevention of femoral uein thrombosis and pulmOllary embolism. Treatment of the acute stroke patient should include at least subcutaneous heparin and regular surveillance. 9. Misdiagnosis by MRI of stroke ill tumor cases. Despite increasing experience, there continues to be a trickle of cases in which a superficial glioma closely resembles and is misdiagnosed as infarction on MRI. Practically, the error makes little or no difference, but the repercussions may be disproportionately great.
10. Unanticipated cerebral hemorrhage illatrial fibrillation. It is a basic rule in atrial fibrillation that a stroke is embolic until disproved. Occasionally another process, hemorrhage, seizure, etc. is responsible. A 75year-old woman with atrial fibrillation awakened with a severe left hemiplegia with sensory loss, anosognosia, neglect, impersistence, unknowing eye closure, clipped speech, etc. Alterness and memory were intact. Cerebral embolism seemed a certainty, but computed tomography demonstrated a 2.5-cm right thalamic hemorrhage. Now I would like to alter course and comment on
some aspects of therapy. 11. Bolus heparin versus continuous drip heparin. For 20 years or more, it was routine on our neurologic service to use intermittent bolus injections of heparin (3,500-6,000 units every 4 h) in the treatment of incipient arterial thrombosis-basilar, carotid, middle cerebral, PCA, etc. This would be maintained for 1-3 weeks, at which time a changeover to warfarin was made . When the continuous drip method was introduced for venous thrombosis, it was soon applied to 48 ] STROKE CEREBROVASC DIS, VOL. 4, NO. 1, 1994
cerebral arterial thrombosis as well. It is my clinical experience that the drip method is less effective and the complications no fewer. Maintenance of a steady, satisfactory level of anticoagulation is highly uncertain using the drip method. Occasionally, for example in basilar thrombosis, when intermittent deficits are continuing or the deficit is progressing while heparin is being given by drip, changing over to bolus heparin arrests the progression. Whenever heparin drip fails, I feel bolus heparin should have been instituted. In ischemic cerebrovascular disease, heparin is one of our most effective therapies, and unequivocal guidelines for its use need to be established. Only clinical results can determine the optimal dose of heparin. 12. Holding thepatient vertically upside doum. Two patients who were acutely comatose as a result of basilar artery thrombosis and given up for lost were revived when given a bolus of heparin intravenously and held upside down with their heads on a low stool and their feet in the air. Within 10 min, they began to respond and eventually in the following weeks recovered and returned to their former work, one as vicepresident of a bank, the other as a house painter. This approach might be tried more often by house officers less sensitive to the gibes of their colleagues. In disturbances of the retinal circulation when blood flowin the arteries and veins is observed ophthalmoscopically to be sluggish, tilting the patient head down on a tilt table may result in a remarkable speeding up of the flow.
13. Extensive craniectomy for uncontrollable brain edema. Occasionally, after middle cerebral artery infarction, particularly in patients under 50 years of age, there occurs massive brain edema uncontrolled by steroids, hyperventilation, mannitol, and phenobarbital. In this situation, when the patient is young and the nondominant hemisphere is involved, early consideration might be given to craniectomy and incision of the dura mater. Frontotemporal craniectomy combined with temporal lobectomy may suffice, but extension of the craniectomy to the occipital region is more effective. The immediate threat, to the patient with brain edema is midline displacement at the tentorium. Frontotemporal craniectomy barely extends to the plane in which the horizontal force on the brainstem is greatest. There are also younger patients with aneurysmal subarachnoid hemorrhage in whom there occurs acute brain swelling and coma. Extensive bilateral craniectomy may be miraculously efficacious. A physician's wife, aged 36 years, gradually became comatose in the 48 h after an acute subarachnoid hemorrhage. Dr. Robert Ojernann, as a desperate measure, performed a bilateral subtotal craniectomy with incision of the
dura and the insertion of a gusset. Four hours after surgery, the patient was fully alert and literally intact neurologically. The brain is being strangled for want of 35 ml more of space. Analogous therapy may be applicable in acute meningitis with posterior fossa compression. Knowledge of the pathologic aspects of these processes may provide ideas for new therapeutic approaches. 14. Heparin and mannitol in basilar bifurcation embolism. I have been greatly impressed that some patients who are comatose or stuporous as a result of embolism to the upper bifurcation of the basilar artery respond to an immediate intravenous bolus of heparin combined with an infusion of mannitol. Were this observation a valid datum, it could be explained as the interplay of three factors. First, the persisting clinical picture resulting from embolism to this site usually does not develop monophasically, but rather the deficit progresses or evolves in stuttering fashion over a period of 24-36 h. During this time, heparin could prevent superimposed clotting in regions of sluggish flow. Another possibility is that heparin actually affected some lysis of the embolus. Jaques (5) who reported the anticoagulant effect of heparin in experimental animals had used glass cannulas surgically inserted into the divided carotid arteries of dogs. He noted that not only was clotting prevented but sediment deposited in the cannulas disappeared, indicating that thrombus that had already formed underwent lysis. Jaques later suggested that heparin, by virtue of its anticoagulant effect, may have allowed natural fibrinolysis to occur, thus accounting for dissolution of the deposit in the cannulas (personal communication, 1992). A third possibility would be that mannitol produced shrinkage of the embolic mass, thereby promoting greater flow. 15. Long-term warfarin therapy. In the Boston area in a randomized study of warfarin in atrial fibrillation directed by Kistler, strokes were reduced to almost zero in the treated group (6). Another striking result was that there were 2.4 times as many deaths in the control group (26:11). Only one of these deaths was the result of a stroke. There is an indication here that warfarin had a significant salutary effect, and continuation of long-term warfarin therapy should be considered in arteriopaths with involvement of multiple regions. There remain for commentary a few further points in cerebrovascular disease. 16. Using percentage narrowing in measuring the residual arterial lumen. In recent studies of endarterectomy for carotid stenosis, the residual lumen has been expressed as a percentage of the diameter of the normal artery rather than simply using the diameter of the
residual lumen. It is unlikely that the original diameter of the carotid artery determines cerebral blood flow, since in most persons the vessel can be ligated with little effect on the cerebral blood flow (7). Experience indicates that it is the residual lumen that is the crucial factor. Determining a percentage requires the introduction of a denominator whose accuracy is less certain than that of the numerator. The denominator varies from patient to patient, is influenced by such factors as poststenotic dilatation and the carotid slim phenomenon, and does not have a pathologic correlate. It is not certain that this method of measurement can be widely applied with reliability. We,at the Massachusetts General Hospital, have found that measurement of the residual lumen is simpler and more accurate than the percentage method. 17. Stroke related to general anesthesia and surgery. This type of stroke remains little known. Caplan et al. (8) presented a series of cases in which infarction in the territory of the basilar artery occurred intraoperatively or postoperatively in patients undergoing noncardiac surgery under general anesthesia. It was postulated that mechanical deformation of the vertebral arteries with resultant dissection or intimal tear occurred during manipulation of the neck. A pathologic study was not available. In the following case, a healthy 46-year-old woman underwent an abdominal operation for a small ovarian carcinoma. One hour after recovering from the general anesthetic, she became comatose. Autopsy disclosed embolic material blocking the upper bifurcation of the basilar artery. There was no embolic source in the heart, lungs, or aorta. The vertebral arteries were not examined in the neck. The internal carotid artery may be similarly affected . In a personal case, a 34-year-old woman developed bilateral carotid dissections during an appendectomy under general anesthesia. A persistent postoperative headache preceded a severe right hemiplegia and aphasia by 4 days. 18. Theduration ofa TLA. Currently, the upper limit for the duration of a TIA is commonly placed at 24 h. At the meeting of the second ad hoc committee on nomenclature in CVD, the duration was placed at 230 min (9). Spells ofTMB last less than 10 min in 90% of instances. Prodromal TIAs heralding a thrombotic stroke last less than 10 II1in in 70% of cases (10). Extending the duration of a TIA to 24 h results in the inclusion of embolic and thrombotic strokes of brief duration, precluding a reliable correlation of clinical, pathologic, vascular, and imaging data. In the Boston atrial fibrillation study (6), three patients each had a single spell that lasted 2-3 h, 12 h, and 18 h, respectively. None was as brief as 20-30 min. 19. 'TLAs" in amyloid angiopathy. Several instances
J STROKE CEREBROVASC DIS, VOL . 4, NO.1,
C. MILLER FISHER
of spells resembling TIAs have been reported in congophilic amyloid angiopathy (CAA) (11). Loeb (personal communication, 1992) has pointed out the resemblances of these spells to the creeping paresthesias of migraine accompaniments and to Leao's spreading depression. In his two cases, the spells ceased on anticonvulsant medication, raising the possibility of sensory seizures. In this regard, Tan (personal communication, 1992) reported that migrainous accompaniments in all of 10 instances were controlled with anti-se izure medication. Furthermore, I have been informed of a patient whose daily noncarotid TMBsof undetermined nature ceased on carbamazepine therapy. These interesting observations relating to migrainous phenomena warrant further exploration. It might be mentioned here that it is widely held that CAA is not diagnosable by angiography. Armed with the details of the microscopic pathology, I am impressed that it is possible using a hand lens to observe diagnostic irregularities of the small cortical branches of the middle cerebral artery when normal angiograms are used for comparison. My colleagues are not persuaded. As part of the pathologic picture in CAA, small areas of cortical infarction have been described. Serial sections of CAA cortex show that such areas are actually the outlying borders of small hemorrhages. 20. TIAs in marantic endocarditis. It is well-known that TlAs may occur in the antiphospholipid antigen (APLA) pro coagulant syndrome, in which LibmanSacks endocardial verrucae are a feature. The same may hold for marantic endocarditis, for in two pathologically studied personal cases with that diagnosis spells of TMB and transient numbness of the limbs had occurred. The literature contains two further cases (12). Although the nature of the procoagulant state associated with marantic endocarditis remains a puzzle, apparently it can be associated with transient neurologic phenomena just as occurred in the APLA syndrome. 21. Transient episodes of vertigo. Episodes of unaccompanied vertigo after the age of 50 years regularly raise the spectre of a brainstem stroke. In the clinical interpretation, it is well to have a procedural format. When the sudden onset of vertigo is associated with a change in acuity of hearing or the onset of tinnitus as the only accompaniments, the diagnosis of an endorgan labyrinthopathy is almost certain. When the onset of vertigo is accompanied by diplopia, paresthesias, cerebellar ataxia, direction-changing horizontal nystagmus, vertical nystagmus, internuclear ophthalmoplegia, or other brainstem signs, the diagnosis of brainstem vascular disease can safely be made. When neither of these clinical pictures pertains and the ver50
J STROKE CEREBROVASC DIS, VOL. 4, NO.1,
tigo is unaccompanied or isolated, the following observations may serve as guideposts. (a) Although basing clinical diagnosis on statistics is anathema, it is important to realize that isolated dizziness is accounted for by cerebrovascular disease in not more than one case in a hundred. (b) TlAs of dizziness occur in 80% of basilar ischemia cases, but, in 95% of such cases, there are other neurologic accompaniments as described above. Therefore, unaccompanied dizziness is rare even in vertebrobasilar ischemia. (c) Basilar TIAs last 25 min or less in 85% of cases. Therefore, if unaccompanied vertigo persists for hours or days, the diagnosis of vascular disease is excluded. This includes its subsidence on lying quietly and its recrudescence on movement. (d) Examination of the patient during a period of vertigo is desirable in order that the absence of nystagmus can be established as well as other brainstem signs. (e) Dizziness beginning in the morning on awakening and turning over in bed (matutinal dizziness) is benign until disproved (13). (f) An eighth-nerve tumor does not cause on-off spells of vertigo. (g) End-organ rotatory dizziness is more often systematized in one plane or another than stroke dizziness, which is often described as "every which way." (h) In acute end-organ dizziness, horizontal nystagmus is not direction-changing. Over the years, I have had the opportunity to confirm this in some 10 instances. (i) If dizziness persists for days or weeks, it is almost certainly of endorgan origin. (j) There often is a history of a similar episode a few years before. (k) When the spell of vertigo is unaccompanied, I find it is a safe policy to wait and watch, confident that should brainstem ischemia be responsible, it will be announced by the appearance of brainstem accompaniments in time to initiate heparin therapy. 22. Notions to bediscontinued. (a) Already referred to is the idea that endarterectomy may convert a pale infarct into a hemorrhagic one. (b) The concept of a shower of emboli is not based on clinical experience or pathological study. (c) Attributing any small stroke to a lacune is wrong, since an infarct or hemorrhage at any site may be small. (d) The principle of ischemiaat-a-distance, i.e., that stenosis in one carotid system can cause ischemia on the other side or in the posterior circulation, should be invoked with the greatest reluctance. (e) Drop attack as a stroke phenomenon is probably a nonentity. In conclusion, may I say that I shall be pleased ifyou have found something of interest and of practical value. The topics have been a potpourri, but that is a consequence of having to sit by the side of the road casting an eye on the passing scene. I have even offered some advice, which is part of being a kibitzer.
Acknowledgment: The support of the Manny and Ruthy Cohen Foundation, Inc., Washington, DC, is acknowledged.
References 1. Fisher CM. The "herald hemiparesis" of basilar artery occlusion. Arch NeuroI1988;45:1301-3. 2. Fisher CM. Clinical syndromes in cerebrallhrombosis, hyp ertensive hemorrhage and ruptured saccular aneurysm. CUll Neurosurg 1975;22:117-47. 3. Fisher CM. Lacunar infarcts-a review. Cerebrovasc Dis 1991;1:311-20. 4. Ojemann RG, Heros RC, Crowell RM. Surgical mallagemellt of cerebrovascular disease, 2nd ed, Baltimore: Williams and Wilkins, 1988. 5. Jaques LB. Effects of intravenous injections of heparin in the dog. Am] PhysioI1939;125 :98-107. 6. The Boston Area Anticoagulation Trialfor Atrial fibrillation Investigators. The effect of low-dose warfarin on
7. 8. 9. 10. 11. 12.
the risk of stroke in patients with non-rheumatic atrial fibrillation. N Ellgl ] Med 1990;323:1505-11. Miller JD, [awal K, Jennet B. Safety of carotid ligation and its role in the management of intracranial aneurysms. ] Neurol NeurosurgPsychiatry 1977;40:64-72. Caplan LR, Tettenborn B, DeWitt LD. Brainstem and cerebellar infarcts after non-cardiac surgery. Neurology 1991;41(Suppll):367. NINCDS Ad Hoc Committee on Cerebrovascular Diseases . A classification and outline of cerebrovascular diseases. II. Stroke 1975;6:564-612. Fisher CM. Concerning transient ischemic attacks. Cleve Clill ] Med 1987;54:3-11. Smith DB, Hitchcock M, Philpott PJ. Cerebral amyloid angiopathy presenting as transient ischemic attacks. Neurosurgery 1985;63:963-4. Ashenhurst EM,Chertkow G. Cerebral embolism due to non-bacterial thrombotic endocarditis. CallMed Assoc] 1962;86:313-7. Fisher CM. Vertigo in cerebrovascular disease. Arch OtolaryllgoI1967;17:614-7.
] STROKECEREBROVASC DIS, VOL. 4, NO.1, 1994