Abstract
Migraineurs with frequent aura (≥1 per month) are at a 12-fold increased risk of posterior circulation infarctions. Although the mechanism is unclear, these lesions are located in the arterial borderzone of the cerebellum. This case illustrates, for the first time, crossed cerebellar diaschisis in a patient with migraine with prolonged aura, and raises the possibility that hypoperfusion may be responsible in part for the development of these lesions.
Introduction
A recent Dutch population-based magnetic resonance imaging (MRI) study has demonstrated that migraineurs with more than one aura episode per month have a 12-fold increased risk of cerebellar infarctions compared with controls matched for age, sex and vascular risk factors (1). The CAMERA study also found that migraine sufferers had an increased risk of hyperintense T-2-weighted lesions in the pontine tegmentum, and that female migraineurs had more supratentorial deep white matter lesions (1, 2). The risk of lesions increased with attack frequency and was independent of cardiovascular risk factors.
Although the mechanism of cerebellar infarctions in migraineurs is not known, the overwhelming majority of infarctions are located in the vascular border zone of the cerebellum (3). This has led to speculation that these infarctions may be the result of migraine attack-related hypoperfusion and/or cerebral embolism, rather than atherosclerosis-related thrombosis.
We report the case of a migraneur with prolonged aura in whom subtraction ictal single photon emission computed tomography (SPECT) coregistered with MRI (SISCOM) demonstrated crossed cerebellar diaschisis (CCD) during an episode of aura. We comment on the clinical and pathophysiological relevance of this finding in the context of the recent findings of the CAMERA study.
Case report
A 61-year-old woman presented with 20 years' history of stereotyped neurological episodes. Each episode was preceded by several hours of profound fatigue. The episode began with paraesthesias that involved the right side of her face and gradually over the course of 30–60 min would descend to involve the right hand, arm and leg. While the symptoms disappeared within 1–2 h, they would often recur multiple times over a 1-week period. The sensory symptoms were followed within 30–60 min by a mild or moderate bilateral frontal throbbing headache, nausea, mental confusion, and expressive language dysfunction and word-finding difficulty. There was no associated photophobia or phonophobia. These symptoms could last up to 1 week, before completely resolving. Neurological examination was remarkable only for mild tandem gait ataxia. There was no evidence of ocular motility deficits, gaze-evoked nystagmus, orappendicular dysmetria. Speech and language were normal and formal neuropsychological testing revealed no deficits in any cognitive domain. Laboratory testing revealed a normal complete blood count, sedimentation rate, C-reactive protein and chemistry panel, and negative antinuclear, double-stranded DNA, extractable nuclear and anticardiolipin antibodies, antineutrophilic cytoplasmic antibody, cryoglobulins, lupus anticoagulant, protein C and S, protein C resistance, and mutation analyses for notch 3 (CADASIL), CACNA1A [familial hemiplegic migraine (FHM)] and factor V Leiden. Paraneoplastic autoantibody panel, which was ordered prior to evaluation by the author for reasons that are not clear, was normal. Electroencephalography during wakefulness and light sleep during the course of an attack revealed a dysrhythmia Grade 1, with intermittent rhythmic theta activity, which was maximal in the left hemisphere. Transthoracic echocardiography, without agitated saline administration, was normal. The atrial septum appeared intact. Brain MRI revealed mild cerebral atrophy, mild right cerebellar hemisphere atrophy and deep white matter T-2 hyperintensities, more prominent in the left hemisphere (Fig. 1a). To exlude the possibility of an epileptic focus, brain SPECT was performed within 30 min of the onset of a typical attack during the march of sensory symptoms. Baseline brain SPECT was performed 3 weeks after complete resolution of this attack. Subtraction-ictal SPECT coregistered with MRI (SISCOM), a computer-based image fusion technique originally developed for identification of epileptogenic foci in seizure patients (4), revealed a reversible reduction in cerebral blood flow in the left hemisphere during an attack (Fig. 1b, blue coloured areas, arrow), with associated CCD, in the arterial borderzone of the right cerebellar hemisphere (Fig. 1c, blue coloured areas, arrow). Whereas the literature reports to date have documented the application and usefulness of SISCOM in epilepsy patients, there are few if any reports of the use of this technique in non-epileptic patients with reversible cerebrovascular deficits. The SISCOM software used was ANALYSE 6.0 (Biomedical Imaging Resource, Mayo Foundation, Rochester, MN USA). The gamma camera is a dual-head Seimens E-CAM gamma camera with low-energy, high-resolution, parallel-hole collimators. The acquisition parameters for this study were 64 stops per head, 35 s per stop, 128 × 128 × 16 matrix. The MRI used was a GE Signa 1.5-T, head coil, FSPGR 10 TE, 23TR. The radiopharmaceutical employed was ∼20 mCi 99mTc-ECD (Neurolite) (Cardinal Health Radiopharmacy, Phoenix, AZ, USA) for each separate SPECT study. Both the ictal and interictal dosages were administered to the patient in a quiet, dimly lit room. The patient was treated with acetazolamide, lamotrigene, verapamil, topiramate, and pregabalin, levitiracetam and gabapentin. She responded to lamotrigene and topiramate, but was unable to tolerate either drug long-term due to skin reaction (lamotrigene) and cognitive side-effects (topiramate). She is presently on clonazepam 1 mg qhs and her attacks occur with a frequency of approximately one per month.
(a) Axial FLAIR magnetic resonance images reveal T2 signal hyperintensity in the cerebral white matter. The signal changes are most prominent in the left posterior periventricular white matter. (b) Ictal single photon emission computed tomography coregistered with magnetic resonance imaging reveals decreased cerebral perfusion involving the left hemisphere and contralateral cerebellum, predominantly in the arterial borderzone of the right cerebellar hemisphere.
Discussion
While unilateral hemispheric reductions in cerebral blood flow have been demonstrated during typical migraine aura (5), this report of migraine with prolonged aura is the first to demonstrate CCD. This finding suggests that the increased risk of borderzone cerebellar infarctions in migraineurs with aura may be related to attack-related reductions in cerebellar blood flow.
CCD has been reported in patients with ischaemic and haemorrhage hemispheric stroke, during carotid amytal procedures, and in a single patient with FHM during a prolonged (1 week) attack (6). CCD is thought to be due to interruption of crossing corticopontocerebellar fibres resulting in transneuronal metabolic and blood flow alterations that are distant to and on the opposite side of the primary area of hypoperfusion or hypometabolism. Since CCD appears to correlate with the severity of hypoperfusion, it may not be seen frequently in migraineurs with aura, since the hemispheric cerebral blood flow deficit is usually short-lasting and invariably below ischaemic thresholds (5). However, it is possible that CCD may occur in susceptible patients during aura, especially in those with prolonged aura.
The paucity of arterial anastomoses in the cerebellar borderzone, combined with the impairment in adaptive posterior circulation haemodynamic mechanisms in migraineurs with aura (7), may render susceptible individuals vulnerable to ischaemic cerebellar infarctions as a result of CCD during attacks of aura. It is possible that subischaemic threshold reductions in cerebellar blood flow may also contribute to subclinical cerebellar dysfunction and progressive cerebellar atrophy seen, respectively, in patients with typical aura and FHM (8, 9). The P/Q type voltage-gated calcium channel, which is highly concentrated in the cerebellum, may render FHM patients with CACNA1A mutations selectively vulnerable to progressive cerebellar degeneration as a result of glutamate-mediated excitotoxicity during repeated episodes of hemiplegic migraine.
If CCD is responsible in part for the borderzone cerebellar infarctions in migraineurs with aura, the lesions would be expected to predominate in the cerebellar hemisphere ipsilateral to the clinical symptoms of the aura, or contralateral to the hemisphere with reduced cerebral blood flow. It is unclear whether this is indeed the case, but future studies should investigate this possibility. It is also possible that cardiac right-to-left shunts (RLS) may contribute to cerebellar infarctions in migraineurs with aura. The prevalence of RLS in patients with migraine with aura is about 2.5 times higher compared with migraineurs without aura or a non-migraine control group (10). Future studies should examine the prevalence of RLS in migraineurs with cerebellar infarctions compared with those without such lesions.
This case illustrates, for the first time, CCD in a patient with migraine with prolonged non-hemiplegic aura. Although this patient had no evidence of a right cerebellar infarction, in light of the recent finding of an increased prevalence of borderzone cerebellar infarctions in migraineurs with aura, CCD may contribute to the development of these lesions. Further cerebral blood flow investigations in migraine patients during aura are warranted to reproduce this finding and to explore the prevalence of CCD and the migraine subtypes in which it occurs. In addition, correlation between infarct location and aura laterality may help to support a role of CCD in the pathogenesis of cerebellar infarctions. Finally, the association between cardiac RLS and cerebellar infarctions in migraineurs with aura may help in determining whether, and to what extent, cerebral embolism may play an aetiological role in the cerebellar infarctions seen in this group of patients.