Clinical manifestations

Presentation and course

The diagnostic criteria of migrainous infarction were just revised by the International Classification of Headache Disorders, third edition (ICHD-III), as follows (coded as 1.4.3) (24):

The diagnostic criteria, in fact, do not differ from those proposed by the International Classification of Headache Disorders, second edition, code 1.5.4 (23), except for some wording.

The most common clinical scenario is of a patient who has had frequent attacks of migraine headache preceded by temporary hemianopia. Following a particularly severe attack, the deficit persists and the patient is left with a permanent loss of vision in one visual field. This visual field deficit may regress gradually over the following months, ultimately leaving a clearly defined area of permanent visual field loss. The next most common signs of migrainous infarction are hemiparesis, monoparesis, and hemisensory symptoms, characteristically having the cheiro-oral distribution (ie, involving the digits and the side of the mouth). Less frequently, neurologic deficits include ataxia or persistent dysphasia. Although persistent neurologic symptoms associated with migraine are usually visual and localized to the region of the cerebral cortex supplied by the posterior cerebral arteries, ischemic strokes attributable to migraine also commonly occur within the territory of the middle cerebral artery. A rare case with right-side occipital infarction presented as persistent visual aura for 1 week without visual defect (68). Lee and colleagues reported rare auditory symptoms as initial symptoms in two patients with migrainous infarctions (36). Brain MRI showed infarction over the upper pons and cerebellum in one patient and negative findings in the other. It is of note that the negative neuroimaging finding in the latter case did not fulfill the criteria of migrainous infarction based on the ICHD-II (23). Hoekstra-van Dalen reported that migrainous infarction is a stroke entity that causes mostly infarcts in the occipital lobe (27). Migrainous infarction involving the bilateral anterior cerebral artery territory has been reported (16). Tang and colleagues reported two patients with migrainous infarctions that involved two different cerebral arteries (67). A 29-year-old woman developed infarctions over the right posterior cerebral artery and right anterior choroidal artery simultaneously, and a 47-year-old man developed two episodes of migrainous infarction within 4 years involving the territories of the middle and posterior cerebral arteries sequentially. Even though migrainous infarction is considered more common in those younger than 45 years old, a 93-year-old woman was reported to suffer from migrainous infarction (72). In one unusual case, a 47-year-old man with visual aura was reported, who developed early recurrence of migrainous infarction— one in the cerebellum and the other in the occipital lobe (53). Chhabra and colleagues reported two rare patients with migraine with retinal aura, who developed migrainous infarction as retinal or optic nerve infarction (14). Catarci reported a 74-year-old man who developed right-side occipital ischemic stroke after visual snow phenomenon (10). His visual snow symptoms changed from bilateral and temporary to left-sided and permanent one day on awakening. The patient denied a history of migraine with aura, and his persistent visual snow disappeared completely after about 1 year.

Arboix and colleagues reported nine consecutive patients with migrainous infarction diagnosed according to the strict criteria of the International Headache Society (22; 04). Their mean age was 35.7±12 years. Six patients (67%) were women. All suffered from headache at the onset of neurologic deficit. The stroke manifested as limb weakness in five patients, sensory symptoms in five patients, hemianopia in four patients, nausea and vomiting in four patients, and aphasia in one patient. Six patients had a cerebral infarct visible on neuroimaging studies (in the territory of the middle cerebral artery in three patients, posterior cerebral artery in two patients, and superior cerebellar artery in one patient). The neuroimaging studies of the remaining three patients showed negative findings even though the patients’ neurologic deficit lasted at least 7 days.

Two large series, one from Germany (n=17) (77) and the other from Finland (n=33) (35), showed that migrainous infarction most commonly occurs in younger women with a history of migraine with aura. Most patients presented with prolonged visual aura (82%). The severity of the stroke was mild, with an NIH Stroke Scale score of 2. A total of 70.6% to 82% of patients had acute ischemic lesions in the posterior circulation. The prognosis was usually good. The differentiation between prolonged aura and migrainous infarction was difficult and was associated with delayed admission of patients to the hospital.

Prognosis and complications

In general, the long-term prognosis for patients with migrainous infarction is good. Milhaud and colleagues, in a prospective stroke registry, found that the outcome at 1 month was favorable in more than 70% of migraineurs with ischemic stroke (45). Arboix and colleagues reported that the mean length of hospital stay of nine consecutive patients with migrainous infarction was 9.75±6.2 days. No patients died during hospital stay, and 67% were symptom-free at discharge (04). Rothrock and colleagues followed 28 patients with migrainous stroke for a mean of 25.3 months. There were six recurrent strokes. The researchers reported that migrainous stroke itself showed comparatively poor prognosis for risk of recurrent stroke, and stroke recurrence rates of previous reports have been lower because the patient populations involved were either smaller than theirs or were followed for shorter intervals (54). It is of interest that Linetsky and colleagues reported six patients with migrainous infarction whose headache frequency and severity decreased after ischemic stroke (40). They hypothesized that the improvement in migraine may be due to reduced nociceptive transmission as the result of loss in vasoreactivity of the affected cerebral blood vessels. Two large series both showed a very favorable prognosis in patients with migrainous infarction (35; 77). A long-term study of 15 patients with migrainous infarction, mean age of 34.8 years, showed a favorable outcome in 80% of patients with a mean follow-up duration of 7.5 years.

Clinical vignette

A 32-year-old woman who worked as a laborer suffered from severe migraine for 1 week. This was followed by weakness in her limbs on the left side. She denied hypertension or any other cardiovascular risk factors and was not on oral contraceptives. Her usual pattern was one or two attacks of migraine without aura per month and, rarely, migraine with visual aura. During the week she had severe headaches and noted a visual aura (bright light) over her left visual field that lasted for 30 minutes. This lasted longer than her previous auras, and 2 hours later she noted transient left limb weakness that lasted for approximately 30 minutes. When she awoke the next day, her headache was less severe, but her left limbs were paralyzed. Neurologic examination showed left hemianopia and left hemiparesis (muscle power grade 1 to grade 2) with hyperreflexia and Babinski sign. MRI of the brain showed acute cerebral infarction over the right frontotemporal region compatible with MCA occlusion. MRA showed occlusion of the M1 segment of the right middle cerebral artery but no vasospasm. The patient had a normal electrocardiogram, cardiac echogram, carotid artery duplex sonography, ANA, prothrombin time, APTT, protein S, protein C, and anticardiolipin antibody. She was discharged on verapamil 240 mg once per day and aspirin 100 mg per day. In 6 months, she had recovered enough to walk without aid. At the 4-year follow-up, she still had occasional migraine without aura and infrequent migraine with aura. No recurrence of stroke was noted.

Biological basis

Etiology and pathogenesis

The exact cause of migrainous infarction is still not certain. Based on studies using cerebral angiography during attacks, the most important underlying mechanism for the stroke is believed to be carotid or vertebral arterial spasm resulting in a critical degree of cerebral hypoperfusion (19; 55; 56). However, this may not always be true, so other factors should be considered.

Occasionally, drug therapy is considered to be a contributing or precipitating factor. For many years, ergotamine overdoses have been thought to cause constriction of the cerebral vessels. Propranolol has also been reported to be related to permanent neurologic defects in some patients, particularly if the drugs induced postural hypotension among elderly patients (52). This relationship, however, is not well established. Mendizabal and colleagues reported a case of migrainous stroke in which treatment with propranolol was associated with stroke onset and may have played a causative role (41). Oral contraceptives, particularly in high doses, have been found to greatly increase the incidence of migrainous cerebral infarction, especially in young women (11; 45). Serotonergic medications were perhaps responsible for ischemia and the subsequent stroke (46; 65).

It is possible that some migraineurs’ strokes are a result of microemboli associated with these disorders (48). Patent foramen ovale, a potential source of paradoxical cerebral embolism, was found to be more frequent in young patients with migraine with aura (01). Milhaud and colleagues found that patent foramen ovale is an independent risk factor for ischemic stroke in young migraineurs (45). Wolf and colleagues found a high frequency in a German series (64.7%) (77). However, Laurell and colleagues did not show an increased frequency (40%) (35). Carerj and colleagues reported a higher prevalence of atrial septal aneurysm in patients with migraine with aura than in patients with migraine without aura and controls, suggesting that atrial septal aneurysm may play a part in cryptogenic stroke in migraineurs (09). Other investigators have related carotid or vertebral artery dissection to cerebral infarction in patients with migraine.

The pathologic mechanisms responsible for cerebral infarction remain unproven; arterial vasospasm (with cerebral edema), arterial wall dissection, and increased platelet aggregation are the most likely candidates. Other factors may contribute to migrainous infarction, but extracranial or intracranial vasospasm is believed to play a major role.

Fifty years ago, Wolff and his coworkers suggested that arterial vasospasm was the primary cause of the aura preceding the migraine attack (78). This vascular theory of migraine has been supported by studies of cases of "complicated migraine" and migrainous infarction where prodromes have persisted and cerebral infarction has been documented. All investigators agree that attacks are accompanied by reduced cerebral blood flow during the aura. Meyer and colleagues reported significant reductions in regional cerebral blood flow during the prodromal interval with 133Xe inhalation (44). The evidence that vasospasm is the primary cause of the symptoms of migrainous aura is supported by regional cerebral blood flow measurements, arteriographic data, and limited autopsy studies demonstrating the absence of intrinsic vascular or cardiac disease despite extensive brain infarction.

It is possible that patients with migrainous infarction have rheological abnormalities predisposing to platelet aggregation, coagulation disorders, and intravascular thrombosis. Welch contends that "spreading oligemia" can lead to platelet accumulation in hypoperfused cerebral areas (75). Vasoactive substances released by platelets could further reduce cerebral blood flow to ischemic levels. The proportion of platelet microaggregates in the circulation has been shown to be greater than in control individuals during the prodromal period of an attack. Augmentation of this intrinsic platelet abnormality by extrinsic factors (eg, oral contraceptives, alcohol abuse, cigarette smoking, and diabetes mellitus) combined with intrinsic conditions (eg, mitral valve prolapse or migrainous vasospasm) are probable factors in the genesis of migrainous strokes. The frequent presence of antiphospholipid antibodies and lupus anticoagulant in young patients with stroke has suggested that these factors play an important part in migrainous stroke (64). However, this finding remains controversial (70). Santos and colleagues reported a 43-year-old woman who developed migraine aura-like symptoms before developing right-side, middle cerebral artery ischemic infarction (57). During operation, several episodes of cortical spreading depression were recorded with increasing levels of glutamate and lactate/pyruvate ratio. This case provided a link between migraine aura and stroke by cortical spreading depolarization.

The von Willebrand factor, a large, multimeric glycoprotein acting as a procoagulant via stimulation of platelet adherence, activation, and aggregation, has been purported to be an independent risk factor for ischemic stroke. The role of von Willebrand factor in migrainous infarction was explained by Tietjen and colleagues, who found that migraineurs with or without prior stroke had significantly higher von Willebrand factor antigen and activity than controls (69). Proinflammatory platelet adhesion to leukocytes occurring during the headache-free interval in patients with migraine is similar to that seen in those with acute coronary or cerebrovascular syndromes (79). This may support the hypothesis that migraine attacks predispose to stroke by inducing platelet-related hypercoagulability.

According to the ICHD-3, migrainous infarction has to occur in patients with migraine with aura. It is possible that the mechanisms underlying migrainous infarction and migraine with prolonged aura might be the same. However, a study adopting 31P-MR spectroscopy found the cortical energy reserve in patients with migrainous infarction was similar to that in controls but higher than in patients with migraine with prolonged aura (60).

Wolf and colleagues found that acute ischemic lesions in patients with migrainous infarction were often multiple and located in distinct arterial territories and that there were no overlapping ischemic lesions; therefore, they suggest that hemodynamic compromise during the development of migraine is unlikely to be the cause of infarction (77). Chen and Eikermann-Haerter, in a comprehensive review, suggest that the mechanisms linking migraine and stroke connection are multifactorial, with genetic predisposition, aura-related electrophysiological mechanisms (cortical spreading depolarization), and cerebral microembolism being the most convincing ones (13).

Epidemiology

The incidence of migrainous infarction is rare, according to the strict criteria proposed by the International Headache Society. Henrich and colleagues reported that the incidence rate of first migrainous infarction was 3.36 per 100,000 person-years (26). However, in the absence of other stroke risk factors, this estimate was reduced to 1.44 per 100,000 person-years. In a study of the Barcelona Stroke Registry, Arboix and colleagues reported that the group of patients with migrainous infarction accounted for only 0.6% of all first-ever acute strokes, 0.8% of ischemic strokes, 12.8% of ischemic strokes of unusual etiology, and 13.7% of ischemic strokes in young adults 45 years of age or younger (04). Therefore, stroke can occur during migraine attacks, but the association between migraine and stroke is infrequent (06; 71). It was also reported that ischemia-induced migraine attacks may be more frequent than migraine-induced ischemic insults (49). A large series in Germany found 17 patients among 8137 patients with stroke (0.2%) over an 11-year period (77).

Unlike the rare epidemiological survey of migrainous infarction, the association between migraine and stroke has been well documented (08; 42; 73). A case-control study in young women also revealed that oral contraceptives, high blood pressure, and smoking were high-risk factors for infarction in migraineurs (11). Schwaag and colleagues repeated this association that migraine was a significant risk factor for juvenile stroke in Germany (62). The risk was even higher in people younger than 35 years of age and in females. In a cross-sectional study, migraineurs, particularly those who had migraine with aura, had a higher cerebrovascular risk than individuals without a history of migraine (59). By following a cohort of the Women’s Health Study for a mean of 9.9 years, Kurth and colleagues demonstrated that migraine with aura but not without aura was associated with about a two-fold increased risk of ischemic stroke as well as myocardial infarction and ischemic cardiovascular death and a 1.7-fold increased risk of coronary revascularization (31). Compared with women without migraine, Kurth and colleagues observed a J-shaped relationship between migraine frequency and cardiovascular events, with a higher risk for very infrequent (< monthly) and frequent migraine attacks (> or = weekly) and a lower risk for monthly migraine (33). The risk was only increased for women with migraine with aura. Stroke was associated with migraine with aura and a migraine frequency of at least one per week. The prospective data from the Physicians’ Health Study also indicate an increased risk of stroke in men with migraine (RR: 1.84; 95% CI: 1.10–3.08) who were aged 40 to 45 years at study entry. This association was not found in older age groups (32). The same group further demonstrated that the association between migraine with aura and cardiovascular disease varies by vascular risk status. The association was strongest in the lowest cardiovascular risk score group stratified by the Framingham risk score (34). A meta-analysis confirmed the association between migraine aura and stroke but not myocardial infarction (61).

MRI has been successfully used to evaluate clinical and subclinical lesions in the brain. Subclinical cerebral lesions, especially in the posterior circulation or white matter, were reported to be more frequent in patients with migraine (especially migraine with aura) in a case-control MRI study, ie, CAMERA (Cerebral Abnormalities in Migraine, an Epidemiological Risk Analysis) study (30). The same group, in another study, demonstrated that most (88%) infratentorial infarct-like lesions had a vascular border zone location in the cerebellum and, further, that a combination of (possibly migraine attack-related) hypoperfusion and embolism is the likeliest mechanism for posterior circulation infarction in migraine (29). Scher and colleagues studied the association between migraine in middle age and late-life infarct-like lesions in a prospective, large, population-based cohort, ie, Age Gene/Environment Susceptibility (AGES)-Reykjavik Study from Iceland (58). After on average 25 years, subjects with migraine with aura had increased odds of late-life infarct-like lesions on MRI (OR=1.4), a result entirely driven by an association of migraine with aura and cerebellar lesions among women (OR=1.9). There was no association between migraine without aura or nonmigraine headache with the infarct-like brain lesions.

Prevention

Featherstone reported that patients with migrainous stroke were usually young adults with a past history of migraine with aura whose headache attacks had worsened in severity and duration just before the stroke (19). Therefore, it is wise to initiate prophylactic measures for stroke during these prodromal intervals. Comorbid cardiovascular risk factors (including angina pectoris, diabetes mellitus, hyperlipidemia, hypertension, collagen disease, blood dyscrasias, and cigarette smoking) should also be controlled.

It has been reported that a history of migraine is a risk factor for cerebral ischemia in patients younger than 45 years of age (15; 08; 62). Migraine-related stroke is more common in individuals with aura than those without aura, but it does occur in patients with migraine without aura (37). Patients with a history of migraine-associated stroke are at significantly increased risk for recurrent strokes (54). Avoiding migraine drugs with marked vasoconstrictive action (ergotamine or possibly sumatriptan) and using them in accordance with labeling and removing other vascular risk factors (smoking and oral contraceptives) are additional measures for the prevention of migraine-related stroke, especially in patients with migraine with aura (37; 06; 50). On the subject of oral contraceptives and hormone replacement therapy in women with migraine, the International Headache Society Task Force recommended that combined oral contraceptives should be stopped when the following conditions exist: (1) new persisting headache, (2) new onset of migraine aura, (3) increased headache frequency or intensity, and (4) development of unusual aura symptoms, particularly prolonged aura (07).

Differential diagnosis

Because migraine is so common in the general population, it is not unusual to observe patients with cerebral infarction who also have a history of migraine. The fact that a cerebral infarction develops around the time of a migraine headache does not necessarily mean that they are related.

Headache may occur acutely with any stroke; approximately 30% of patients with stroke have headaches (05). At times, pulsatile headaches follow both atherothrombotic and embolic cerebral infarction. Thus, in a migraineur, headaches at the onset of a cerebral infarction may not necessarily be of the migraine type. Some conditions—in particular, arterial dissection, arteriovenous malformation, cardioembolic stroke, and intracranial neoplasm—can mimic migrainous infarction.

Idiopathic thunderclap headache with vasospasm presenting as severe intractable headache may be followed by a delayed stroke. It mimics migrainous infarction; however, the onset of headache is much more abrupt (within a split second) in patients with thunderclap headache (17).

Shuaib described five patients with migraine headaches who developed cerebral infarctions (63). The initial diagnosis in all cases was migrainous infarction, but the diagnosis was later revised. Three patients had arterial dissection (one proven at autopsy), one had marantic endocarditis (proven at autopsy), and one had generalized atherosclerosis associated with diabetes.

Diagnostic workup

Because distinct diagnostic criteria are lacking, migrainous infarction should be considered a diagnosis of exclusion. The diagnosis demands a well-established history of migraine and exclusion of other conditions that cause stroke. Computed tomography, MRI (including diffusion-weighted image, perfusion-weighted image, MRS, and MRA), cerebral arteriography, transcranial Doppler evaluations, and transesophageal echocardiography evaluations should be performed when possible. Lumbar puncture and blood studies to help exclude vasculitis should also be carried out.

MRI techniques provide more information for acute cerebral infarction and may be the diagnostic tool of choice for migrainous infarction (47), particularly if the clinical examination demonstrates brainstem or cerebellar signs. It might also be useful in patients with prolonged aura (66). It has been reported that elevation of cerebral lactate can be detected among patients with migrainous infarction by localized 1H MRS (74). Liang and Scott (38) reported a patient with migrainous infarction whose MRI showed cortical laminar necrosis. The authors hypothesized that sluggish perfusion combined with a variety of metabolic derangements (spreading depression in the case of migraine) probably accounted for this unusual finding. Two patients with migrainous infarction showed cortical laminar necrosis in their MRI studies (03; 51). Arai and associates reported on the chronological changes in neuroimaging findings in a 64-year-old man with migrainous infarction (02). Of note, the MRI abnormalities suggested the presence of vasogenic edema, prolonged hyperperfusion, and alteration of the blood-brain barrier followed by irreversible brain damage.

When carried out at the time of the acute process, cerebral arteriography often reveals abnormalities during migrainous stroke. Dukes and Vieth first reported angiographic observations in migraine with aura (18). Sequential angiography made during the prodromal phase showed progressive decreases in the caliber of the internal carotid arterial system. Lieberman and colleagues reported a case of complicated "diplegic migraine" with angiographic demonstration of vasospasm of both the cervical carotid and intracranial middle cerebral arteries (39). However, cerebral arteriography carries a significant complication rate, and the risk-to-benefit ratio should be carefully weighed in the acute stages of presumed migrainous infarction.

Gomez and colleagues reported an interesting case of a young woman with a longstanding history of migraine with aura who developed acute cerebellar infarction in association with reversible vasospasm of the vertebral arteries (21). The vasospasm was demonstrated by transcranial Doppler sonography. However, the differential diagnosis of thunderclap headache with vasospasm should also be considered in this case (17).

Management

Prophylactic treatment of migrainous infarction in patients with prolonged aura includes platelet antiaggregants or calcium channel blockers. Aspirin administration should be considered as prophylactic treatment against migrainous infarction. Patients who have anticardiolipin syndrome may need anticoagulants. Calcium channel blockers are recommended for patients at risk for migrainous infarction. Nimodipine reduces cerebral vasoconstriction, and several studies have demonstrated the potential effectiveness of calcium channel blockers in the prophylactic treatment of migraine headaches (43).

Ergotamine, triptans, and serotonergic medications may initiate or worsen intracranial vasospasm and dysautoregulation in patients with migrainous infarction. These drugs should not be given to patients with prolonged aura. Some experts believe that peripheral beta-blockers such as propranolol should also be withheld because they may worsen intracranial vasoconstriction. For prophylactic purposes, verapamil should be considered for migraine attacks and for migrainous infarction.

Special considerations

Pregnancy

Pregnancy is clearly a risk factor for stroke, but several mechanisms are involved (76). However, the risk of migrainous infarction was not studied in pregnant women.