Migraine and Cerebral Infarct-like Lesions on MRI: Title and subTitle BreakAn Observation, Not a Disease

Migraine is a common, chronic-intermittent neurovascular headache disorder characterized by moderate to severe headache attacks and various combinations of gastrointestinal and autonomic nervous system disturbances.^{1 - 2} The prevalence peaks in midlife and approximately 12% of the population is affected, with women being affected 3 to 4 times more often than men.³ In approximately one-third of patients with migraine, transient neurological symptoms occur that mostly involve the visual field but can also involve the sensory or motor system (migraine aura). While the specifics of migraine pathophysiology are still being unveiled, dysfunction of brain cells and arteries has been identified as a major component.^{1 ,4}

In several neuroimaging studies and a meta-analysis,⁵ migraine has been associated with brain lesions, mostly in white matter, with the meta-analysis finding an odds ratio (OR) of 3.9 (95% confidence interval [CI], 2.3-6.7). However, the cause and mechanisms of the white matter lesions remain unclear. Most commonly, these lesions among patients with migraines have been interpreted as an indirect marker of focal cerebral hypoperfusion, a mechanism that can be induced by migraine attacks, particularly migraine attacks with aura.⁶ Many questions remain unsolved: why do some individuals with migraine have brain lesions but others do not? What is the nature of lesions among those patients and does it differ from lesions among individuals without migraine? Is there a preferential location of these lesions among patients with migraine headaches? What is the association of migraine features, such as migraine attack frequency, migraine severity, or migraine-specific treatments, with these lesions?

In this issue of JAMA, Scher and colleagues⁷ evaluate the association between migraine in middle age and late-life infarct-like lesions using data from a large, population-based cohort study from Iceland. A total of 4689 individuals (57% women) from the Age Gene/Environment Susceptibility (AGES)-Reykjavik Study, who were a mean age of 51 years at the time of headache ascertainment, were reevaluated on average 25 years later, including magnetic resonance imaging (MRI) of the brain. Information about headache was self-reported and participants with at least 1 headache per month were asked about typical migraine features, including visual disturbances during or before the headache and unilateral numbness before the headache, which was used for migraine aura classification. Performance of MRI and the definition of infarct-like lesions followed a standardized protocol. In midlife, a total of 12.2% of the participants (17.0% of women) were classified as having migraine and of those, 63% were classified as having migraine with aura. After adjustment, participants with migraine with aura had increased odds of late-life infarct-like lesions on MRI (OR, 1.4; 95% CI, 1.1-1.8), a result entirely driven by an association of migraine with aura and cerebellar lesions among women (OR, 1.9; 95% CI, 1.4-2.6). There was no association between migraine without aura or nonmigraine headache with brain lesion in any location, and the association between migraine with visual symptoms and cerebellar lesions was stronger than that for other sensory aura.

The study by Scher et al has several strengths compared with previous studies, including the large population-based sample, prospective design, and long length of follow-up. The results confirm the association between migraine and silent brain lesions and, more specifically, results from the CAMERA study,⁸ and extend these findings to a longitudinal setting. The CAMERA study, a community-based study from the Netherlands of 295 individuals with migraine and 140 controls, showed a cross-sectional association between migraine and silent infarct-like lesions in the posterior circulation territory that was apparent only among those with migraine with aura and stronger for those with at least 1 migraine episode per month.⁸ The presence of infarct-like lesions in the CAMERA study was not associated with supratentorial white matter lesions or other structural brain changes.⁹ However, the frequency of deep white matter abnormalities was higher in women with migraine compared with controls.⁹

Several points should be considered when interpreting the findings of the AGES-Reykjavik Study. First, classification of migraine and migraine aura was based on relatively broad, self-reported measures, limiting comparisons with other studies. The prevalence of migraine with aura (63%) was substantially higher than estimates from other population-based studies, which raises concerns about misclassification or biased selection processes. The prevalence of migraine with aura in the CAMERA study was also relatively high (55%). However, it would be difficult to explain why these potential misclassification biases should cause an association between migraine with aura and infarct-like lesions in prospective data.

Second, the nature of the infarct-like lesions observed on the MRI remains elusive. In the study by Scher et al and in the CAMERA study, infarct-like lesions were defined as having a size of at least 4 mm and being surrounded by an area of high signal intensity on FLAIR (fluid-attenuated inversion-recovery) images. However, these criteria were not applied to lesions in the cerebellum, ie, the region with higher frequency of infarct-like lesions among patients with migraine. These less specific criteria may increase the difficulty to distinguish, for example, between infarct-like lesions and enlarged perivascular space (ie, the Virchow-Robin space), which is particularly problematic in the intratentorial region because of reduced image quality due to close proximities of anatomic structures. Thus, some lesions might be of a different nature and perhaps of no particular pathology. Alternatively, if these lesions are infarcts, the message would be alarming, raising concerns about the potential long-term consequences of migraine on brain structure and function. However, such a conclusion does not fit current concepts of the rather benign course of common migraine on the brain, particularly in the absence of relevant clinical cognitive or other functional decline.¹⁰ In addition, with the exception of the extremely rare familiar hemiplegic form of migraine that is caused by specific gene mutations, individuals with migraine do not have cerebellar symptoms.

Third, how does this observation fit with the established association between migraine and ischemic stroke?^{11 - 14} Several differences exist when comparing the association between migraine and silent infarct-like lesions with the association between migraine and ischemic strokes. Clinically relevant strokes among patients with migraine occur most often in supratentorial regions and do not cluster in particular locations in population-based samples.¹⁴ The association between migraine and ischemic stroke is apparent in women and men, stronger in younger age, and not detectable in older age groups of most studies.^{11 ,13} In addition, specific subgroups exist in which the association is particularly strong, including those with the combination of smoking and oral contraceptive use,^{12 ,14} low cardiovascular risk profile,^{11 ,14} or a specific genetic polymorphism.¹⁵ There is currently little evidence that the mechanisms leading to infarct-like lesions are similar to the ones leading to ischemic strokes, but subgroups may exist in which mechanisms overlap. Furthermore, migraine and migraine aura are heterogeneous disease entities and primary forms should be distinguished from migraine caused by other conditions or diseases. In some patients, arterial wall dissections could play a role in migraine, but cervical artery dissections are too rare to explain increased risk of stroke or silent lesions in the posterior circulation.¹³

In summary, the clinical implications of the AGES-Reykjavik study⁷ should be interpreted with caution. In the absence of the source and the nature of infarct-like lesions and the absence of clinical symptoms or consequences, it is premature to conclude that migraine has hazardous effects on the brain. In this regard, brain scans among patients with migraine should not be initiated to detect silent brain lesions but to rule out rare secondary forms of migraine among those patients with atypical migraine forms or migraine courses. However, the study raises important questions. New studies examining the association of migraine with structural brain changes and brain function should improve understanding of the associations and perhaps further unveil migraine-specific mechanisms.