Abstract
Migraine has been identified in a recent meta-analysis to be an independent risk factor for stroke with a relative risk of 2.16 (95% confidence interval 1.48, 2.48), the risk being higher in migraine with aura (MA), in women taking oral contraceptives (1) and in smokers (2). A higher prevalence of white matter abnormalities is also seen in migraine (3–6). Migraine has been associated with supratentorial posterior circulation infarcts but not an excess of cerebellar infarcts compared with non-migraineurs (5, 7). However, in the recent Cerebral Abnormalities in Migraine, an Epidemiological Risk Analysis [CAMERA (6)] study, small (<2 cm) cerebellar infarcts were demonstrated on magnetic resonance imaging (MRI) in migraineurs, particularly those with MA who had never had a clinical stroke. These lesions were located predominantly (>90%) in cerebellar arterial borderzone territories. Here we describe a patient with MA who presented with a prolonged headache associated with mild gait ataxia and evidence of acute infarcts in arterial borderzones in both cerebellar hemispheres on diffusion weighted MRI (DW-MRI).
Case report
A 38-year-old man with a history of migraine with visual aura awoke with a throbbing right frontal headache with mild photophobia but no aura or phonophobia. His symptoms improved moderately with aspirin and sleep. Six hours later he developed sudden-onset unsteadiness and generalized weakness with intermittent vomiting for 10 h. He was admitted to a regional hospital and a computed tomography (CT) brain scan the following morning showed a small (11 × 8 mm) left cerebellar infarct (Fig. 1a) and a left subtentorial calcific lesion thought to be a meningioma. His headache persisted over the next 6 days for which he took acetaminophen only. On day 3 after presentation he had a stereotypical migraine starting with a scintillation scotoma in the right visual field which enlarged and then receded over 30 min and was followed by a right-sided headache with photophobia which lasted 60 min. Two days later on awakening he felt immediately flushed, sweaty and generally weak with nausea. He experienced vertigo, gait unsteadiness, vomiting and developed a left frontal headache with photophobia. The majority of these symptoms abated over 6 h. A repeat CT brain scan (5 h from the onset of symptoms) on the 6th day after presentation showed a new right medial cerebellar infarct (12 × 7mm, Fig. 1b). The patient was given 325 mg aspirin, started on 81 mg aspirin/day and transferred to our hospital.
Axial computed tomography (CT) scans at presentation (a) and 6 days later (b) at a local hospital (CT slice thickness 2.5 mm in posterior fossa) showing a left then a new right cerebellar hypodensity, respectively (arrows). T2-weighted (c) and diffusion-weighted (d) magnetic resonance (1-T General Electronic Signa magnet, 5 mm slice thickness) axial images of the cerebellum showing three cerebellar infarcts (two deep and one cortical; arrows) and apparent diffusion coefficient map (e) on day 7 post initial presentation showing an area of hypodensity in the right cerebellum (arrow).
His first migraine occurred at age 14 years and he had had one or more migraine attack annually until the prior 3 months, during which he had one MA per month. There were no clear triggers for this increased migraine frequency. The only other medical history was of previous episodes of presyncope which had been investigated 1 year ago with a 24-h Holter monitor and a cardiac stress test, which were both normal. He was an ex-smoker and drank 20 units of alcohol per week. There was no history of ergot, triptan or illicit drug use. There was no family history of stroke or ischaemic heart disease, but his brother had typical visual auras without headache and his son had migraine without aura.
On examination on the day of his first MRI scan his pulse was 60 and regular and blood pressure 140/80. His cardiorespiratory, abdominal and neurological examinations were unremarkable except for some mild unsteadiness of gait with minor difficulty tandem walking.
MRI of the brain on the day 7 after presentation showed bilateral cerebellar lesions on T1 and T2-weighted images with restricted diffusion on DW-MRI consistent with infarction (see Fig. 1). The infarcts involved cortical and subcortical areas of the posteromedial poles of both cerebellar hemispheres, sparing the inferior and superior regions of the cerebellum. Three infarcts were noted in the left cerebellar hemisphere, four on the right and one on the left side of the vermis. The more inferior infarcts were in the territory of the medial posterior inferior cerebellar artery (mPICA) bilaterally, whereas the more superiorly situated infarcts were in the borderzone between the mPICA and medial superior cerebellar artery [mSCA, see Fig. 1 (8, 9)]. A single deep white-matter lesion in the right frontal lobe was noted. MR angiography of the extra and intracranial arteries showed no occlusive lesions or evidence of arterial dissection. The left vertebral artery was dominant. The right PICA and both anterior inferior cerebellar arteries (AICA) and SCAs were present, but no left PICA was seen.
ECG showed sinus bradycardia, with normal PR and QT intervals. A transoesophageal echocardiogram was normal with no valvular abnormality, patent foramen ovale or atrial septal aneurysm. The following blood tests were normal or negative: complete blood count, international normalized ratio, activated partial thromboplastin time, erythrocyte sedimentation rate, renal, liver and thyroid functions tests, C-reactive protein, glucose, hypercoaguable and vasculitic blood screen (including lupus anticoagulant, anticardiolipin, antithrombin III, protein C and S, factor V Leiden mutation, C3, C4, rheumatoid factor, extractable nuclear antibodies, antinuclear and antidouble-stranded DNA antibodies). Total cholesterol was 5 (total cholesterol:HDL ratio of 4) and urinalysis normal. The patient was continued on aspirin 81 mg/day and discharged home. His symptoms of mild gait unsteadiness improved to normal by 2 weeks after discharge. On review at 3 months he had no further headaches and a normal neurological examination. A repeat MRI brain scan showed some diminution of the cerebellar infarcts (not shown).
Discussion
This is the first report of multiple small acute cerebellar infarcts involving borderzone territories demonstrated by DW-MRI in a patient with MA. The cerebellar infarcts described here are similar to the ‘infarct-like lesions’ observed in the CAMERA study, lending support to the hypothesis of Kruit et al. (6) that these lesions occur as acute infarcts in migraineurs. This case fulfills two out of the three criteria for a migrainous stroke (10); first, neuroimaging demonstrated an infarct in an area appropriate for his symptoms and no other cause of stroke was found. However, the third criterion was not fulfilled; the primary symptoms did not begin with a stereotypical MA or manifest stroke symptoms similar to the aura. It has recently been highlighted that migrainous infarction has probably been overdiagnosed in the past and that normal aetiological investigations in a migraineur who has a stroke do not prove that migraine is the cause (2). Further, stroke can give rise to migrainous symptoms (2). As far as possible, we have excluded the main causes of stroke in a younger man, e.g. arterial dissection, cardiac thromboembolism (including patent foramen ovale), vasculitis, hypercoaguable states and premature atherosclerosis.
Our patient’s neurological examination was unremarkable except for very minor gait unsteadiness. One study has demonstrated subclinical hypermetria in migraineurs (particularly MA) using computerized optoelectronic tracking of reaching movements compared with healthy volunteers (11). In one series (12) cerebellar infarcts <2 cm diameter were subclinical in 37% of patients compared with only 4% of territorial cerebellar infarcts. Our patient’s course is in keeping with previous reports that borderzone cerebellar infarcts tend to produce transient benign symptoms, mild or absent cerebellar signs and minimal long-term disability (13, 14).
The fact that we could not identify a left PICA was probably due to the presence of the common anatomical variant of a single AICA-PICA trunk on the left (8) rather than a left PICA occlusion, particularly as the cerebellar infarcts were bilateral. The infarcts identified in our patient arose in the mPICA territory or in the borderzone between the mSCA and mPICA territories. This latter category accounted for 37% of cerebellar infarcts in the CAMERA study (6). In one case series, small cerebellar infarcts <2 cm diameter (mean age 60 years) tended to arise in borderzone territories (12). Two recent case series using MRI and DW-MRI have described cerebellar infarcts in deep or cortical borderzone territories in 26% and 23% of all patients with PICA or SCA territory infarcts, respectively (13, 14). In these studies the likeliest causes of infarction were cardioembolic, extracranial large artery disease or in situ branch thrombosis; 20% were of unknown cause. The latter study (SCA group) recorded migraine as a stroke risk factor, although no data on migraine prevalence in the borderzone infarct group were given. However, migraine was suggested as a possible mechanism in one case (13). Borderzone cerebellar infarcts may therefore be a previously under-recognized stroke subtype which is more readily identified using MRI.
The reason for the predilection for cerebellar involvement in migraine-related stroke is unclear. In familial hemiplegic migraine (FHM) proton MR spectroscopy has demonstrated a correlation between metabolic changes (reduced N-acetyl aspartate and glutamate levels, and elevated myo-inositol levels) in the superior cerebellar vermis and gait ataxia scores (15). Further, crossed cerebellar diaschisis (i.e. contralateral cerebellar hypoperfusion) has also been documented in FHM (16). Migrainous aura is associated with occipital lobe hypoperfusion using perfusion MRI, but no change in DW-MRI (17), although this study did not assess cerebellar perfusion MRI. Cerebellar borderzone infarcts have been postulated to be secondary to hypoperfusion of watershed areas (6) and it is possible that this occurs at the time of contralateral occipital hypoperfusion during the visual aura of a migraine attack. Arterial thrombosis is also proposed as a mechanism for cerebellar infarcts in migraine (7) and a recent population-based study has described a higher prevalence of cardiovascular disease risk factors in migraineurs compared with non-migraineurs (18). This may explain the higher stroke prevalence in migraine (1), although our patient had no clear cardiovascular risk factors. Another putative mechanism for cerebellar infarction in migraine is vasospasm (2).
In summary, we describe an otherwise healthy patient with MA presenting with acute bilateral small borderzone cerebellar infarcts demonstrated by DW-MRI. This suggests that patients with migraine are susceptible to small borderzone cerebellar infarcts as indicated in the CAMERA study (1) that may produce mild symptoms or be clinically silent.
Conflict of interest
None declared.
Footnotes
Acknowledgements
J.R. is supported by the Dalhousie University Internal Medicine Research Foundation and an unrestricted educational grant from Hoffman-LaRoche.