Migrainous Infarction and Migraine Triggered Epilepsy
K. Michael A. Welch
Marie-Germaine Bousser
Julien Bogouslavsky
Migraine is associated with an increased risk of strokes between and during attacks (30,115). Various forms of migraine are recognized, and they are generally classified according to the transient, though sometimes persistent, neurologic deficits that may precede, accompany, or outlast the headache phase. A number of these clinical syndromes may mimic cerebrovascular syndromes, including migraine with aura of different types but especially prolonged aura, retinal or ocular migraine, ophthalmoplegic migraine, hemiplegic migraine, and basilar-type migraine (30,67,115,156). If the deficit of a migraine attack remains, migraine-induced stroke should be suspected.
PAST CONCEPTS
In 1881, Fere (58), working with Charcot at the Salpetrière Hospital, provided one of the earliest comprehensive descriptions of the problem. He reviewed 12 patients suffering from classic migraine who also experienced language and sensorimotor symptoms. Fere later reported Charcot’s fatal case of a 53-year-old man with classic migraine since adolescence who developed permanent aphasia and right faciobrachial paralysis, offering an explanation (58) of vasospasm as the cause. Charcot again emphasized this notion of cerebral ischemia as a result of vasospasm in his discussion of a case of ophthalmoplegic migraine (32).
Among the fatal cases of migrainous cerebral infarction that have been studied at autopsy, characteristic pathologic changes have not been consistently identified (120). Buckle et al. reported the case of a 16-year-old in whom an angiogram just prior to death showed widespread arterial narrowing (22). Other clinical pathologic studies extended the concept of vasospasm (93,125,128). These reports offered the possibility that repeated attacks of severe migraine may lead to focal arterial injury, which in turn may predispose to interictal stroke of thromboembolic nature (36,57,91). These arterial changes presumably predispose the patient to thrombosis or distal embolization. In support, an autopsy of one patient with familial hemiplegic migraine (FHM) was exceptional for its demonstration of small deep infarcts, a distinctly uncommon site of migrainous cerebral infarction, indicating that lenticulostriate arteriopathy may be one mechanism, a hypothesis favored by Bruyn (19, 20, 21). The clinicopathologic features of the patients with migraine reported by Guest and Wolf (80) and Polyak (135) are more compatible with cerebral embolism. The hypothesis of local vascular alterations in the course of a migraine attack, particularly the vasodilatory phase (154), contributed to the concept of arterial dissection as a cause of migrainous stroke.
Scattered clinical studies with accompanying laboratory investigations were reported prior to the 1950s, describing serious neurologic consequences of migraine, usually permanent visual field defects (24,83,93,129, 140,142,152) prior to the 1990s. Not only was thrombosis related to vasoconstriction considered a complication of migraine, but so was intracerebral and subarachnoid hemorrhage (2,12,26,43,50,74,133). In 1962, Connor (35) reported on a series of 18 complicated migraine cases, observing a predilection for occipital cortex involvement. The largest series of cases was reported in 1965 by Pearce and Foster (131). Only two vascular malformations were detected among 29 angiograms performed, but these cases had unusual features, including seizures and loss of consciousness. This established the rarity of vascular anomalies in patients with persisting deficits.
EPIDEMIOLOGY
A review of mostly uncontrolled hospital-based studies conducted before 1989 of patients under 50 years of age with a diagnosis of stroke showed that between 1% and
17% were attributed to migraine; in two thirds of these the diagnosis was made in 1% to 8% of patients and 11% to 17% in one third (3). A compilation of studies up to the same time revealed a prevalence of 4% attributed to migraine in 448 total stroke cases, 31% of which had an unknown cause. In clinical studies, stroke was reported as more common in patients with migraine with aura (14,145) and in patients with posterior cerebral artery (PCA) strokes (14). No differences in stroke risk factors were found in migraine sufferers compared with controls without stroke, although those with migraine were more likely to have recurrent stroke, supporting migraine being an independent stroke risk factor (145). Another study of migraine with aura reported that 91% of patients who had stroke during an attack had no arterial lesions, as opposed to 9% of migraine with aura patients who suffered stroke remote from a migraine attack and 18% of patients with stroke without a migraine history (15). In some instances, however, stroke risk factors increased stroke risk in migraine with aura. This group revisited the subject over a decade later with the benefit of a large prospectively collected database in Lausanne, Switzerland (116). Of 3,502 ischemic strokes, 3.7% were judged to be migraine related and to have active migraine as defined by at least two attacks in the previous 2 months. These cases included patients who had stroke during migraine with aura according to previous International Headache Society (IHS) criteria, but also during typical migraine with aura. Dividing patient groups into young and older, with 45 years of age as the cutoff, migraine-related stroke was more prevalent (15.8%, control population 2.1%) in the younger group. Other features differed distinctly in the young group. Posterior circulation involvement and patent foramen ovale (PFO) were characteristic. Of 66 cases under 45, 24 had stroke during a migraine attack, 15 were judged to suffer from migraine without aura, and 9 with aura. Presenting clinical features of stroke were also very different in the young, including less frequent abrupt onset and motor deficit, but more frequent visual defect. The diagnosis of migraine-induced stroke in migraine without aura patients should be regarded with caution for reasons outlined below. Nevertheless, this large series of prospectively studied patients was a valuable addition to the literature, confirming the relationship of migraine with both induced and apparently comorbid ischemic stroke.
17% were attributed to migraine; in two thirds of these the diagnosis was made in 1% to 8% of patients and 11% to 17% in one third (3). A compilation of studies up to the same time revealed a prevalence of 4% attributed to migraine in 448 total stroke cases, 31% of which had an unknown cause. In clinical studies, stroke was reported as more common in patients with migraine with aura (14,145) and in patients with posterior cerebral artery (PCA) strokes (14). No differences in stroke risk factors were found in migraine sufferers compared with controls without stroke, although those with migraine were more likely to have recurrent stroke, supporting migraine being an independent stroke risk factor (145). Another study of migraine with aura reported that 91% of patients who had stroke during an attack had no arterial lesions, as opposed to 9% of migraine with aura patients who suffered stroke remote from a migraine attack and 18% of patients with stroke without a migraine history (15). In some instances, however, stroke risk factors increased stroke risk in migraine with aura. This group revisited the subject over a decade later with the benefit of a large prospectively collected database in Lausanne, Switzerland (116). Of 3,502 ischemic strokes, 3.7% were judged to be migraine related and to have active migraine as defined by at least two attacks in the previous 2 months. These cases included patients who had stroke during migraine with aura according to previous International Headache Society (IHS) criteria, but also during typical migraine with aura. Dividing patient groups into young and older, with 45 years of age as the cutoff, migraine-related stroke was more prevalent (15.8%, control population 2.1%) in the younger group. Other features differed distinctly in the young group. Posterior circulation involvement and patent foramen ovale (PFO) were characteristic. Of 66 cases under 45, 24 had stroke during a migraine attack, 15 were judged to suffer from migraine without aura, and 9 with aura. Presenting clinical features of stroke were also very different in the young, including less frequent abrupt onset and motor deficit, but more frequent visual defect. The diagnosis of migraine-induced stroke in migraine without aura patients should be regarded with caution for reasons outlined below. Nevertheless, this large series of prospectively studied patients was a valuable addition to the literature, confirming the relationship of migraine with both induced and apparently comorbid ischemic stroke.
The overall incidence of “migrainous infarction” has been estimated at 3.36 per 100,000 population per year (95% confidence interval [CI] 0.87 to 4.8) but in the absence of other stroke risk factors becomes 1.44 per 100,000 population per year (95% CI 0.00 to 3.07) (89). This rate is similar to that reported later in subjects under 50 (14); migrainous infarction accounted for 25% of cerebral infarcts. To place these data in context, the overall incidence of ischemic stroke under age 50 ranges from 6.5 per 100,000 to 22.8 per 100,000 (99,109).
As opposed to hospital data basequestioning, rigorously designed true epidemiologic studies systematically addressing an association between migraine and stroke have been few. In a retrospective study of parents of migraine sufferers conducted early on without benefit of the IHS classification and first published in 1988, no increased risk of stroke was found, but the frequency of hypertension was 1.7 times greater in persons with migraine than in those without (106). In an inconclusive study conducted in 1975, the Collaborative Group for the Study of Stroke in Young Women found that the relative risk (RR) of thrombotic stroke was twofold higher, greater for women with migraine compared with a neighbor but not with hospital controls (79). A hospital-based controlled study of 89 cases found that ischemic stroke was increased more than twofold in patients with migraine with aura (88), but when stroke risk factors were excluded in this group, there was no longer a statistically significant association.
A systematic examination in a large-scale prospective epidemiologic study of men and women showed that, after controlling for established risk factors for stroke, both migraine and severe nonspecific headache were associated with a significantly increased risk for stroke (115). The risk for stroke associated with migraine decreased as the age at stroke increased. Also in a large population-based study involving only men and migraine identified only by associated neurologic deficit—in other words largely migraine with aura—the overall risk of stroke was 2.2% (79).
Probably the most convincing evidence of stroke risk and comorbidity of stroke and migraine come from a succession of three important case-controlled studies. In the first of these studies, no overall association between migraine and ischemic stroke was found, but among women aged less than 45, migraine and stroke were significantly associated (157) (Table 64-1). Risk of stroke was 3 times control for migraine without aura and 6 times the risk of controls for migraine with aura. Further, young women with migraine who smoked increased their stroke risk to approximately 10 times control, more than 3 times greater than young women without migraine who smoked. For young women with migraine on oral contraceptives the
risk of stroke was 14 times control, and 4 times the risk of stroke and the dose of estrogen: the odds ratio was 4.8 for pills containing 50 μg of estrogen, 2.7 for 30 to 40 μg, and 1.7 for 20 μg and 1 μg progesterone. In none of these cases was the stroke induced by the migraine attack.
risk of stroke was 14 times control, and 4 times the risk of stroke and the dose of estrogen: the odds ratio was 4.8 for pills containing 50 μg of estrogen, 2.7 for 30 to 40 μg, and 1.7 for 20 μg and 1 μg progesterone. In none of these cases was the stroke induced by the migraine attack.
TABLE 64-1 Migraine and Stroke Risk in Women Under 45 Years: Case Control Study n = 72 | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||
TABLE 64-2 Migraine Stroke Risk in Women Under 44: Case Control Study n = 308 | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||
In a separate case-controlled study of 308 patients with either transient ischemic attacks or stroke, a history of migraine was more frequently than controls (14.9% versus 9.1%) (29) (Table 64-2). Migraine was the only significant risk factor (odds ratio 3.7) in women below 35 years of age. Although these risk figures appear startlingly high in both studies, it must be remembered that the absolute risk of stroke for this patient population translates to around 17 to 19 per 100,000 woman-years, which is very low.
Most recently a hospital based case-control study involving five European centers studied 291 women aged 20 to 44 years with ischemic, hemorrhagic, or unclassified arterial stroke compared with 736 age- and hospital-matched controls (31). Adjusted odds ratios associated with a personal history of migraine were 1.78 (95% CI 1.14 to 2.77), 3.54 (95% CI 1.30 to 9.61), and 1.10 (95% CI 0.63 to 1.94) for all stroke, ischemic stroke, and hemorrhagic stroke, respectively. Odds ratios for ischemic stroke were similar for migraine with aura (3.81, 95% CI 1.26 to 11.5) and migraine without aura (2.97, 95% CI 0.66 to 13.5). A family history of migraine, irrespective of personal history, was also associated with increased odds ratios, not only for ischemic stroke but also hemorrhagic stroke. Use of oral contraceptives or a history of high blood pressure or smoking had greater than multiplicative effects on the odds ratios for ischemic stroke associated with migraine alone, although only smoking was statistically significant. Change in the frequency or type of migraine on using oral contraceptives did not predict subsequent stroke. Between 20 and 40% of strokes possibly were induced during a migraine attack.
In a prospective cohort study conducted among 39,754 U.S. health professionals of age 45 and older who participated in the Women’s Health Study, patients with migraine with aura had an increased risk of any stroke (1.53, 95% CI 1.02 to 2.31) and higher risk of ischemic stroke (1.71, 95% CI 1.11 to 2.66). In women younger than 55 the risk of total and ischemic stroke was even higher. The absolute risk increase was low, however, at 3.8 additional cases per year for 10,000 women. No increased risk could be documented for migraine without aura (103).
Not surprisingly the migraine and stroke literature has been subjected to meta-analysis in recent years. In 14 studies, 11 case-controlled and 3 cohort, conducted between 1966 and 2004, the meta-analysis concluded the RR of stroke in migraine overall was 2.16 (95% CI 1.89 to 2.48). The risk was consistent in both migraine with and without aura and markedly higher when oral contraceptives were used (RR 8.72, 95% CI 5.05 to 15.05) (55). Of additional interest, data from the Genetic Epidemiology of Migraine population-based study in the Netherlands, migraineurs were more likely to smoke and have parental history of early myocardial infarction. People with migraine with aura were more likely to have an unfavorable cholesterol profile, elevated blood pressure, and a history of early-onset coronary and cerebrovascular disease. Women with aura were more likely to be using oral contraceptives. The odds of having an elevated Framingham risk score for coronary heart disease were approximately doubled for the aura patients (146). In summation, an association between migraine and stroke appears confirmed, especially in women younger than 45 and men with migraine plus aura. Common risk factors for stroke in general increase this risk.
CLINICAL INVESTIGATION
Clinical reports in the past 50 years have provided more convincing evidence for an association between migraine and cerebral infarction, especially those cases documented using angiography, computed tomography (CT) scanning, positron emission tomography (PET), and, most recently, magnetic resonance imaging (MRI). There is little to support any of the possible mechanisms raised by older studies.
Angiography
Kaul et al. (98) reported angiographic findings of four patients with migraine. One had an occlusion of the main trunk of the PCA, presenting with right hemianopia, hemianesthesia, spontaneous pain, transient amnesia, and dysphasia. All four of these patients had branch occlusions of the calcarine artery. Each patient suffered from transient visual disturbances, and the deficits occurred during a severe migraine. Vasospasm attributed to migraine has been documented by arteriography, the first in 1964 by Dukes and Vieth (49). A 44-year-old man suffered from left hemianopia with numbness of the left side of the body, followed by a throbbing right hemicranial headache. The patient underwent arteriography. A first and second set of films demonstrated increasingly poor filling of the intracranial
internal carotid system at a time when focal deficit was maximal. During the headache phase, good intracranial filling of the internal carotid was observed. Other cases reported in subsequent years confirm arterial occlusion (72,110).
internal carotid system at a time when focal deficit was maximal. During the headache phase, good intracranial filling of the internal carotid was observed. Other cases reported in subsequent years confirm arterial occlusion (72,110).
Computed Tomography
The use of CT scanning beginning in the mid-1970s permitted documentation of lesions compatible with cerebral infarction in many cases (6,23,25,40,92,111). In each instance, CT showed a low-density lesion in the cortex, most commonly in the occipital lobe. Review of CT series of selected migrainous patients with or without local focal neurologic deficits showed that the prevalence of abnormalities ranged from 34% to 71%. Cala and Mastaglia (25) reported on 94 patients with a history of “recurrent migrainous headaches,” of whom 6 showed evidence of cerebral infraction. Four had fixed visual field defects with mesial occipital low densities. Cerebral edema, particularly in the periventricular white matter, was evident in another six patients. Baker (6) described diffuse low-density zones during a migraine attack, which disappeared on subsequent CT examination. Hungerford et al. (92) studied 53 patients who had “exceptionally severe” migraine or serious clinical complications including hemiplegia. The most frequently encountered abnormalities were cerebral atrophy (30%) and infarction (10%). Similar findings were reported by Mathew (111). Rascol et al. (137) reported CT-confirmed cerebral infarction occurring in the course of a migrainous attack. Ten patients had syndromes referable to the middle cerebral artery, whereas the remaining four had hemianopic defects owing to PCA territory infarctions. Arteriography, performed in each patient 2 days to 6 months after stroke onset, gave an abnormal result in nine, showing internal carotid artery occlusion in one, middle or PCA stem occlusion in four, and branch occlusions in the remaining four patients.
Magnetic Resonance Imaging
The diagnosis of migraine-induced stroke, both infarction and neuronal necrosis, has been greatly enhanced by the use of MRI. The greatest value of MRI, however, is in the differential diagnosis of stroke owing to migraine from other causes in cases that present as migraine mimics. From a research viewpoint, great interest stemmed from observations of increased white matter lesions in approximately 30% of routinely studied migraine patients compared with healthy controls (12%) (94). Lesions were found in the centrum semiovale and frontal white matter, in some cases extending to deeper structures in the region of the basal ganglia. In some series such findings were more prevalent in migraine subtypes associated with neurologic aura (56). Not all case series found a greater incidence than controls, however (37). Another series found a higher incidence of white matter lesions in patients with migraine, as well as in patients with tension-type headache (45). The mechanisms of these changes remain to be determined. If relevant, they may represent small foci of ischemic infarction of obscure origin, or gliosis. Meta-analysis has again solidified the evidence of an association between migraine and white matter lesions. From seven case-controlled studies, it was determined that patients with migraine have a 3.9 times greater risk compared to controls (95% CI 2.26 to 6.72). The risk was present even in young migraine sufferers with no known stroke risk factors (149).
The most recent and rigorous analysis of MRI findings in migraine are strengthened by being population based. Kruitt et al. performed a cross-sectional prevalence study of Dutch adults aged 30 to 60 years (102). Although there was no overall increase in infarct prevalence between migraine and controls, prevalence was high in the cerebellar region of the posterior circulation (5.4 versus 0.7%). The risk was higher in migraine with aura and in migraineurs with more than one attack a month. In women, deep white matter lesions were seen more often, and again increased with the frequency of migraine attacks but in migraine with or without aura. Why this is only in women remains to be established. Also, why only deep as opposed to peripheral white matter lesions are increased in migraine remains uncertain, although this might implicate different arterial territories.
Finally, diffusion tensor MRI performed in migraine subjects with white matter lesions revealed lower diffusivity histogram peak heights suggesting subcortical structural damage. These measures correlated with the extent of cortical reorganization determined on blood oxygen level-dependent functional MRI activation of the supplementary motor cortex. These findings suggest that functional cortical changes occur in migraine patients who have white matter changes whose cause remains to be determined (143).
CLINICAL SUBTYPES OF MIGRAINE-RELATED STROKE
One drawback in understanding the dilemma that faces the diagnostician has been a lack of consistency in the definition of migraine-related stroke. Three major issues must be considered. Stroke occurs in the course of the migraine attack, causing true migraine-induced cerebral infarction. Migraine may cause stroke because other risk factors for stroke are present to interact with the migraine-induced pathogenesis. Stroke may present as a migraine syndrome, either symptomatic migraine or as a migraine mimic. The IHS classification (86) and new techniques of brain
imaging have served to clarify the association between migraine and stroke.
imaging have served to clarify the association between migraine and stroke.
Coexisting Stroke and Migraine
Definition
A clearly defined clinical stroke syndrome must occur remotely in time from a typical attack of migraine.
Stroke in the young is rare and migraine is common. Clearly, the two conditions can coexist without migraine being a contributive factor to stroke. When the two conditions coexist in the young, the true cause of stroke may be difficult to elucidate. Comorbidity of stroke risk in migraine sufferers seems apparent from the case-controlled series reviewed, wherein women younger 45 years of age with migraine have increased risk of stroke, particularly between attacks. The cause of such strokes remains to be defined, but is likely to be multifactorial. This increases the clinical significance of coincident stroke and should serve to raise clinical consciousness to the need for stroke risk factor awareness in all migraine sufferers. Although comorbid factors may be present (such as increased platelet aggregation or mitral valve prolapse), many are uncertain risk factors for stroke. Recent reports of an association between migraine and PFO warrant discussion, however. A higher frequency of PFO was reported in migraine with aura patients (5) and an association of migraine and PFO in patients with ischemic stroke (116). In a prospective study of 587 cryptogenic stroke patients, 46% had a PFO (104). Migraine was present in 27.3% of patients with PFO, whereas migraine was present in 14% without PFO (OR 1.75; 95% CI 1.08 to 2.82). Migraine was present in 36% of patients with PFO and stroke whereas only 16% of non-PFO patients with stroke had migraine. In small uncontrolled series of cases, migraine was reduced after PFO closure (151,168). With the caution that an association with PFO does not necessarily imply a cause of migraine, speculation about mechanisms has invoked serotonin release from activated platelets uncleared from the lungs, or microemboli-induced microvascular ischemia triggering spreading depression. No doubt this will continue to be a controversial topic, the fear being that a flurry of irresolute PFO closure procedures for migraine prevention will follow these reports. As this chapter goes to press, several clinical trials of device closure of the PFO in migraine prevention are underway.
Stroke with Clinical Features of Migraine
Definition
A structural lesion unrelated to migraine pathogenesis that presents with clinical features typical of migraine.
Symptomatic Migraine
Certain central nervous system or cerebrovascular disorders episodically causes symptoms typical of migraine with neurologic aura. Indeed, ischemia itself may precipitate an attack of migraine (126). In such cases migraine can be considered symptomatic of the underlying disorder. For example, cerebral arteriovenous malformations frequently masquerade as migraine with aura. The side of the aura and the headache are strongly correlated in such cases. Stroke and migraine are major features of the MELAS syndrome associated with the 3,243-point mitochondrial DNA mutation in the tRNA Leu (UUR) gene (51). Migraine attacks also may be associated with cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) (8,155) possibly being symptomatic of the membrane dysfunction associated with this disorder. In fact, a breakthrough in establishing the genetic basis of FHM was achieved because of the clinical investigation of CADASIL (127). Ten different missense mutations were identified in the Notch 3 genes of 14 unrelated families with CADASIL. The Notch genes are intimately involved in intercellular signaling during development. Proteins belonging to the Notch family are transmembrane receptors. Nine of the 10 mutations either added or mutated a cysteine residue in one of the epidermal growth factor (EGF)-like repeats; EGF-like motifs are to be found in the extracellular domain. It is likely that this mutation strongly affects protein conformation, although how this leads to CADASIL remains to be established. Considering the mechanisms whereby MELAS and CADASIL cause migraine attacks, possibly membrane instability and abnormality of cell signaling causing hyperexcitability of neuronal membranes could be the underlying basis.
Related posts:
Full access? Get Clinical Tree
