Brain (1993), 116, 187-202 Ischaemia-induced (symptomatic) migraine attacks may be more frequent than migraine-induced ischaemic insults Jes Olesen,1 Lars Friberg,4 Tom Skyh0j Olsen,1 Allan Renard Andersen,2 Niels A. Lassen,4 Per Evald Hansen3 and Agnete Karle5 'Department of Neurology, Gentofte Hospital, the ^-Department of Neurology, Rigshospitalet, University of Copenhagen, the ^Department of Neurology, Sonderborg Regional Hospital, the ^Department of Clinical Physiology and the department of Radiology, Bispebjerg Hospital, Copenhagen, Denmark SUMMARY INTRODUCTION The relationship between migraine and stroke is often unclear. In the age range 18—65 years the lifetime prevalence of migraine is 16% and 10% have had one or more attacks within the last year. One-third suffer from migraine with aura (previously classic migraine) and two-thirds from migraine without aura (previously common migraine) (Rasmussen et al., 19914*). If there were no causal relationship between migraine and ischaemic cerebrovascular disease one would expect 10% of patients under 65 years of age and Correspondence to: Jes Olesen, MD, Department of Neurology, Gentofte Hospital, University of Copenhagen, 2900-Copenhagen, Denmark. © Oxford University Press 1993 Downloaded from by guest on January 5, 2015 Fifteen consecutive patients with a diagnostic problem of ischaemia-induced migraine with aura (symptomatic migraine) or migraine-associated ischaemia (migrainous infarction) were studied in order to elucidate the mechanisms. Three had a 1 month flurry of daily attacks of migraine auras with or without headache. A severe internal carotid stenosis/occlusion and reduced regional cerebral blood flow (rCBF) was demonstrated. Borderline ischaemia may thus prime the brain for developing migrainous aura with or without migraine (symptomatic migraine). Four patients had a combination of permanent deficits after the very first migraine attack, severe atherosclerosis, risk factors for stroke, high age and no family history of migraine. In these cases the evidence indicates that thromboembolic ischaemia had triggered an attack of migraine with aura (likely symptomatic migraine). Three young females presented long-lasting typical and severe idiopathic migraine with aura. Attack-associated rCBF reduction was likely to have caused permanent, mild, visual or somatosensory deficits (migrainous infarction). In five patients the relationship between migraine and stroke remained unresolved. It seems that ischaemia-induced migraine attacks may be more frequent than migraine-induced ischaemic insults. Therefore, migraine is not as strong a risk factor for stroke as indicated by the mere coincidence of the two disorders. 188 J. OLESEN AND OTHERS M A T E R I A L AND METHODS Patients were selected consecutively from the Department of Neurology, Gentofte Hospital, University of Copenhagen by two of the authors and from the Department of Neurology, Sonderborg regional hospital by one author. All three are neurologists with a special interest in migraine. Cases which were seen by other neurologists in these departments have not been included. The selection criteria were a suspicion of migraine causing ischaemic cerebrovascular disease or vice versa. Because of the heterogeneity and complexity of the clinical picture and the necessity of documenting the migraine aura diagnosis by detailed description, it has been necessary to present the results as fairly detailed case histories. In the majority of patients rCBF was measured using the intracarotid 133-Xenon method with external stationary detectors (Olesen et al., 1971) or 133-Xenon inhalation and single photon emission tomography (SPECT) (Lauritzen and Olesen, 1984). But we also include cases whose rCBF, for various reasons, could not be studied. Patients all gave informed consent. The studies were approved by the Ethics Committee of the County of Copenhagen. Procedure far intra-arterial studies Patients were studied after fasting, in the morning with or without premedication consisting of 10 mg diazepam. After local anaesthesia, the common carotid artery was punctured in the neck and the angiography was performed through the needle. In cases with normal angiography, a thin heparinized polyethylene catheter was introduced into the internal carotid artery by means of the Seldinger technique. Regional cerebral blood flow measurements were then performed using a camera with 254 stationary detectors covering one hemisphere (Sveinsdottir et al., 1977; Olesen et al., 1981a). The catheter was kept open with intermittent injections of heparinized saline (4 IU/ml). For each measurement —10 mCi (400 MBq) of 133-Xenon dissolved in Downloaded from by guest on January 5, 2015 somewhat fewer in the higher age range with ischaemic cerebrovascular disease to also suffer from migraine. The proportion having an attack at the exact time of onset of a stroke would of course be minimal since most migraineurs have less than one attack a month (Rasmussen et al., 1991a). Recently, evidence has been increasing for a more than chance association between the two diseases. Thus, Henrich and Horwitz (1989) found significantly more migraineurs with aura among stroke victims than among controls, despite the fact that none of her stroke probands had a migraine attack in association with the stroke. Bogousslavsky et al. (1988) found that 10% of strokes in young adults occurred during an actual attack of migraine with aura (classic migraine). To analyse the temporal association and causal relationship between migraine and thromboembolic cerebrovascular disease poses, however, several difficulties have been recently discussed in papers by Iglesia and Bousser (1990) and by Welch and Levine (1990): migraine in general and migraine with aura in particular have been poorly defined (Ad Hoc Committee, 1962) until the advent of the international headache classification (Headache Classification Committee of the International Headache Society, 1988). Migraine aura may occur without headache and mimic transient ischaemic attack. Thromboembolic transient ischaemic attack and stroke are frequently followed by headache and, more surprisingly, headache often occurs before a definite thromboembolic event (sentinel headache) (Gorelick et al., 1986). The differentiation of ischaemic cerebrovascular disease and migraine aura is based on the slow march of aura symptoms and the sudden maximal deficits caused by embolic episodes and most other ischaemic cerebrovascular diseases. The march, however, is often not elicited in routine neurological interviews according to our experience. We present a 5-year prospective study of the borderland between migraine and stroke from a defined population. Regional cerebral blood flow (rCBF) elucidated the complicated and, as it turned out, bidirectional relationships. MIGRAINE AND ISCHAEMIC CEREBROVASCULAR DISEASE 189 2 - 3 ml of saline was injected as a bolus into the internal carotid artery, and the wash out was followed for 2 min. Measurements were repeated at 10—20 min intervals. Patients usually developed their aura symptoms after one or two rCBF measurements and measurements were repeated until the late aura phase or into the headache phase (two to six measurements over 1 — 1.5 h). No permanent deficits or other complications were encountered. Regional cerebral blood flow was calculated using the initial slope method and values were translated into colour-coded images (Sveinsdottir et al., 1977; Olesen a al., 1981a). Areas with rCBF deviating > 14% from hemispheric average were considered abnormal according to our normal material (Olsen a al., 1987). Procedure for tomographic studies If at home, patients telephoned at the very onset of an attack and came as soon as possible by taxi. Hospitaladmitted patients were instructed to inform staff immediately if they had an attack. Most patients were studied during an attack as well as when they were free of symptoms. A Tomomatic 64 (1981 —1986) and a Tomomatic 232 (1986-1989) (Medimatic, Copenhagen) were used. The apparatus, calculation paradigms and other technical details have previously been described in detail (Lauritzen and Olesen, 1984; Andersen et al., 1988). Areas of focally abnormal flow were determined on the basis of a normal material using the asymmetry index as previously described (Andersen et al., 1988). RESULTS CASE HISTORIES Group la. Borderline perfusion behind stenosis causing a flurry of symptomatic migraine auras with or without headache CASE 1. A 40-year-old female admitted because of a flurry of migraine attacks. There was no family history of migraine or atherosclerosis and the patient had never previously suffered from migraine. She had never been hospitalized before and had generally been of excellent health except for tension-type headaches - 1 5 days a month, usually worst on the right side. Two weeks before admission she had an unusually severe headache during which her vision was poor in the left eye. This could not be further described. Five days before admission, during a walk, she suddenly got paraesthesiae in the right side of the face slowly spreading in the course of minutes to involve all of the right half of the head and the right half of the tongue. Subsequently the paraesthesiae spread to the first, then the fourth, then the third right finger. The right aim felt stiff but there was no weakness. The time from the first sensation to maximal spread was 15 min and the total duration was 30 min and was not followed by headache. Thereafter she had one or two similar attacks a day all developing over 10—15 min and lasting almost exactly 30 min. Some of the attacks began in the fingers and then spread to the face, others were associated with right homonymous hemianopsia. One attack was associated with throbbing, left-sided occipitoparietal headache and two attacks were associated with predominantly right-sided stabbing headache but the rest were not associated with headache. During the first 3 days of admission the patient had fluctuating symptoms of paraesthesiae and numbness in the right hand and the right side of the face, coming and going and not associated with headache. Then these symptoms cleared and the patient did not have any further migraine auras or other neurological symptoms apart from the usual headaches. Regional cerebral blood flow was measured with 133-Xenon SPECT on the day of admission and repeated Downloaded from by guest on January 5, 2015 For the purpose of this paper, patients fell into one of four groups. Group la comprised three patients with occlusion or severe, haemodynamically significant stenosis of the internal carotid artery causing a flurry of symptomatic attacks of migraine aura with or without headache. Group lb comprised patients where the initial event was also likely to be primary ischaemic cerebrovascular disease with secondarily elicited (symptomatic) migraine. In Group 2, migrainous aura and associated blood flow reduction was presumably the direct cause of neurological deficits. Group 3 comprised five cases where the relationship between migraine and stroke could not be clearly determined. These cases are not further discussed. Symptoms 63 60 4.0 53 51 58 68 4.4 5.0 52 62 5.2 -4 -5 15* 16* 7 0 59 56 71 71 -5 0 62 79 22* 55 74 26* 62 62 55 74 26* 50 71 30* 54 69 21* 19* 4 45 47 53 52 4 -2 45 55 18* 41 51 20* 47 49 45 53 15* 42 52 43 48 10* Superior Middle/inferior temporal temporal left/right left/right % asymmetry % asymmetry 17* 5 18* 5 51 5 0 - 2 59 62 52 65 20* 47 61 23* 53 56 49 60 45 61 26* 48 58 Parietal left/right % asymmetry 5 15* 56 50 2 66 58 4 62 62 59 61 55 55 -1 -1 0 3 0 49 63 22* 58 61 52 61 8 52 48 69 67 72 76 66 72 62 66 -8 -3 5 8 6 62 73 15* 55 66 17* 58 63 Occipital Basal ganglia left/right left/right % asymmetry % asymmetry Case 1 had a flurry of migraine attacks and severe carotid narrowing due probably to spontaneous intimal dissection. Regional cerebral blood flow (ml/100 g/min) calculated from the middle slice (OM+5). 'Significant side asymmetry (/ > <0.05). 3 p.m. 4 June 1987 None 30 March 1987 None 7 10 9 March 1987 Slight tingling 3 p.m. right cheek: 6 11 a.m. 8 March 1987 Slight bilateral 8 p.m. headache 5 53 57 51 65 22* 46 59 22* 48 58 17* Frontal left/right % asymmetry 4.4 4.0 7 March 1987 Pins and* needles Noon in right hand and lip 8 March 1987 Bilateral 2 p.m. headache worse on left side 4.3 3 March 1987 None 9 p.m. 6 March 1987 None 4 p.m. 4 Study no. Date and time of study Arterial pCO2 (% 4.0 TABLE 1. REGIONAL CEREBRAL BLOOD FLOW MEASUREMENTS IN CASE 1 Downloaded from by guest on January 5, 2015 X rn H m z > z a o O r ra MIGRAINE AND ISCHAEMIC CEREBROVASCULAR DISEASE 191 the following 3 days and again after 3 weeks and 3 months. The results are presented in Table 1 and Fig. 1. There was severely reduced rCBF in most of the left hemisphere during the first days after admission demonstrating haernodynamic significance of the stenosis. On the third day rCBF fluctuated without changes in neurological status indicating a functional disturbance perhaps spreading depression. After 3 weeks and 3 months rCBF was normal. CMC HO ,,Ji*tt u0 lot c*roU4 ittnoftis. rtttWUW SfflPTOrtt Tf— STUOVl »MiutortoMM»«(al/100f/*in> STUDY 2 SUnl. Site* 2 . 8 0 M sue* 2. X«rwn MtfUtten r i m stutfu Mwwn Inhittw* tUm atuty FIG. 1. Regional cerebral blood flow studied with 133-Xenon and SPECT in Case 1. Values of rCBF in ml/100 g/min are translated into colours on the scale in the middle. On the day of admission at 4 p.m. and 9 p.m. rCBF was markedly reduced in most of the left hemisphere. The patient, however, was asymptomatic (study 1 and study 2). A vascular occlusion causing such rCBF changes would normally be associated with marked deficits. It is therefore presumed that, in addition to the haemodynamically significant stenosis of the internal carotid artery {see Fig. 2) rCBF was functionally depressed by the repeated migrainous auras. Further studies showed fluctuating rCBF which normalized gradually {see Table 1). Computerized tomography (CT) on the day after admission and 4 months later were normal. Results of carotid angiography are shown in Fig. 2. The patient was followed up for 2.5 years during which she never had another migraine aura. Conclusion. Severe narrowing of the internal carotid artery probably due to spontaneous dissection, caused borderline cerebral perfusion but without infarction. This induced typical migraine auras with and without headache despite the lack of a family or personal history of migraine. As the stenosis decreased in severity, rCBF normalized and migraine auras disappeared. CASE 2. This 42-year-old female was referred because of sudden aggravation of her migraine attacks 1 month prior to admission. There was no family history of early atherosclerosis or migraine. She had never been admitted to a hospital but had suffered from two or three migraine attacks a year for 20 years. They were Downloaded from by guest on January 5, 2015 Sllctl. J. O L E S E N A N D O T H E R S Downloaded from by guest on January 5, 2015 FIG. 2. Angiography results in Case 1. A, severe narrowing of the intraosseous part of the internal carotid artery of the fourth day of admission, B, fibromuscular dysplasia of the middle free portion of the internal carotid artery not compromising the lumen, c, less pronounced narrowing of the intraosseous part of the internal carotid artery 3 weeks Inter. stereotyped, always beginning with left visual disturbances lasting 15—30 min. Vision was hazy with flickering spots and occasional zigzag lines but no definite fortification spectrum and no topographical spread could be elicited. In 20% of attacks visual disturbances were followed by tingling in the left fingertips marching slowly up the arm, then spreading to the face and, occasionally, to half of the tongue in which case there was associated speech difficulty. Time from onset to maximal extension of sensory symptoms was 15 min. Tingling sometimes spread down the left side of the body into the left leg, then clearing in the order in which it appeared, lasting a total of 45—60 min. Occasionally, left-sided weakness was experienced. After the aura an invariably right-sided, pulsating, severe headache associated with nausea, photo- and phonophobia followed. It required bed rest and lasted 4 - 5 h. One month before admission the patient had a particularly severe attack which, for the first time, was associated with dizziness. Severe headache continued for 1 week during which the usual aura symptoms recurred every day, sometimes several times a day. There was a persisting slight weakness of the left arm on admission to the local hospital 1 week after onset. A CT was unremarkable. Daily attacks continued for a total of 3 weeks, then diminished to one attack during the fourth week at which time the patient was studied by us. At the end of the fourth week a direct right carotid arteriography revealed a threadlike internal carotid MIGRAINE AND ISCHAEMIC CEREBROVASCULAR DISEASE 193 stenosis (Fig. 3). An intravenous digital subtraction angiography 3 days later was unchanged. A 133-Xenon SPECT study showed a right hemispheric rCBF decrease (Fig. 4). The entire right hemisphere was unresponsive to diamox, whereas the left hemisphere exhibited a 40% increase (Fig. 4). Conclusion. A haemodynamically significant thread-like irregular narrowing of the internal carotid artery probably due to spontaneous dissection was associated with daily attacks of migraine auras and headache. Conclusion. The 1-month-lasting flurry of migraine with aura occurred before the onset of the middle cerebral artery infarct, at a time when the patient's internal carotid artery probably became severely stenosed or occluded. The sudden onset of severe neurological deficits was probably due to embolization to the middle cerebral artery (Ringelstein et al., 1983). Group lb. Other cases of ischaemia-induced migraine (symptomatic migraine) CASE 4. A 54-year-old female with no family history of migraine but a maternal history of stroke was admitted for attacks of uncertain origin. She had suffered non-insulin-dependent diabetes mellitus for 5 years, mild arterial hypertension for 6 months and smoked 15 cigarettes a day. Routine blood chemistry, chest X-ray and electrocardiogram (ECG) were normal. She had not previously suffered from migraine when, 7 months before the present admission, she suddenly had the onset of dizziness, nausea, general malaise, headache and she fell to the ground with a left-sided hemiparesis and left-sided sensory disturbances. During the next 5 h she improved markedly, but a very mild left-sided hemiparesis and left-sided mild hemisensory disturbances persisted. A scan was unremarkable. She subsequently had four attacks of tingling in the left leg slowly spreading to the arm and later to the face. It was followed by left-sided hemiparesis and speech difficulty which could not be characterized. She also had a star-shaped figure in front of the eyes which could not be characterized further. The duration of attacks could only be indicated with considerable uncertainty but probably was — 10 min. The neurological disturbances were followed by violent usually bilaterally headache and nausea. The left arm was slightly spastic and she had a slight left-sided hemiparesis of supranuclear distribution. Reflexes were normal. Pain and touch sensation was decreased on the left half of the body and face. There was a slight instability of the left arm on straight arm testing. An aortocervicography followed by selective right carotid angiography (7 months after the onset of permanent symptoms) were normal. Angiography induced a typical migraine attack with green and blue zigzag lines and a left-sided homonymous visual field defect. Regional cerebral bloodflow was decreased occipitally in a pattern typical of migraine with aura. Conclusion. Absence of a family or personal history of migraine, relatively high age, multiple risk factors for atherosclerosis and sudden onset of symptoms indicate that the first episode was thromboembolic. Subsequently she developed attacks of migraine with typical aura, originating from the same area probably because the initial thrornboembolic event had primed the brain. Downloaded from by guest on January 5, 2015 CASE 3. This 50-year-old male had no family history of atherosclerosis or migraine. He suffered a noninsulin-dependent diabetes mellitus and arterial hypertension, smoked 20 cigarettes and drank 10 beers a day. Four years before admission he suffered an ischaemic stroke with left-sided hemiparesis. It was not initiated by a migraine attack but was followed by right-sided headache. Four months later he had only very minor paresis in the left arm and a normal left leg. Since the age of 15 years he had had approximately one attack a month of flickering scotoma followed by paraesthesiae involving the left side. They usually spread from the fingers gradually up the arm and then involved the cheek and finally the left half of the tongue. The symptoms developed over 10—20 min and lasted 30 min. They were usually followed by a severe right-sided pulsating headache, nausea and vomiting. One month before the admission his migraine suddenly worsened, attacks occurring almost daily. On the evening before admission he had one of the usual attacks and went to bed. As he awoke in the middle of the night he had a left-sided hemianopsia and left-sided hemiparalysis and was taken to the hospital. He was hemiparalytic, had neglect, was somnolent and complained of severe right-sided headache for several days. Remission of symptoms was slow. On the day after admission rCBF measured by SPECT was markedly reduced in the right hemisphere. Computerized tomography showed a right hemisphere infarct corresponding precisely to the middle cerebral artery territory. The angiography showed occlusion of the right internal carotid artery. Blood flow measurements repeated after 1 and 3 days showed unchanged hypoperfusion. The patient was marginally improved after 3 months. J. OLESEN AND OTHERS 194 7 I FIG. 3. Direct carotid angiography in Case 2. A, severe narrowing of the internal carotid artery with only thread-like filling proximal to the stenosis, B, a later picture shows passage through the artery and collateral filling of the siphon and intracranial arteries. Downloaded from by guest on January 5, 2015 Hfl/'RRfl IX MIGRAINE AND ISCHAEMIC CEREBROVASCULAR DISEASE Mffct tot c«rotJ