WMH classification based on a modified version of the Scheltens visual rating scale indicated that a majority of migraine patients (63.9%) had WMH on T2 and FLAIR imaging which were most prominent in the frontal lobes. In fact, fewer than 2% of patients who did not have WMH in the frontal lobe had WMH in the basal ganglia or infratentorial regions and only 5.6% had WMH in the periventricular WM. WMH were mostly distributed bilaterally (61%) and right hemisphere and left hemisphere lesions occurred with equal frequency.
Our results are in accordance with those from Xiu and colleagues who assessed WMH in 69 migraine patients, 24 of whom had WMH. Similar to our results, Xiu reported WMH to be most prevalent in the frontal lobe (74.9%) followed by the parietal lobes. In our study, 14% of individuals (n = 21) with frontal lobe hyperintensities had WMH located in the insula which is an intriguing finding as the insula is a known ‘cortical hub’ and a key region of the salience network (17) involved in cognitive and interoceptive components of the pain experience (18) and an area known in migraine to demonstrate alterations brain functional connectivity using resting-state imaging (19–21). Furthermore, using positron emission tomography [11 C]HOwe(9) PBR28 imaging, Hadjikhani and collegues found increased binding of the 18 kDa translocator protein, a marker of glial activation, in individuals with migraine with aura in the bilateral insula (22). Furthermore, authors also noted a positive correlation between insula [11 C] PBR28 standard uptake ratio and migraine attack frequency further suggesting insula involvement in neuroinflammation and nociception.
In the present study, migraine patients who had WMH were significantly older compared to migraine patients who did not have WMH. Furthermore, those who were older tended to have larger WMH for lobar (frontal, temporal, parietal and occipital) and periventricular regions compared to younger patients. Frazekas reported that 11% of symptom-free subjects (n = 87, ages 31–83) had WMH in the 4th decade of life which increased to 83% in subjects over the age of 70. This percentage was even higher in those with cardiovascular risk factors (23). It is possible that the larger WMH found in our study could be at least partially attributable to factors other than migraine.
There is substantial evidence that migraine is associated with an increased risk for WMH. However, more work is needed that defines the characteristics of WMH attributed to migraine (i.e. size, shape, number, distribution) and differentiates them from those associated with other diseases, such as small vessel ischemic disease, lacunar infarcts, or multiple sclerosis. The majority of lobar WMH were under 3mm in size. In the opinion of the neuroradiologist who reviewed the imaging WMH were most commonly punctate in shape. None of them were visible on T1-weighted imaging as hypointense, as can be seen with lacunar infarcts, small vessel ischemic white matter changes, and multiple sclerosis. Furthermore, none of the lobar lesions in our migraine cohort were confluent (i.e. none had a score of 6 on the Scheltens scale), as can be seen in these other three entities. An additional observation in our migraine cohort was that few patients had periventricular white matter lesions contiguous with the lateral ventricles or capping of the lateral ventricles. Similarly, few patients had involvement of the basal ganglia, brainstem, or other posterior fossa structures. These features might prove unique to migraine and be useful for differentiating WMH attributable to migraine from those due to other diseases. Interestingly, 30 % of those with WMH identified using the Scheltens scale did not have WMH identified in the clinical imaging report. Possible explanations for a negative clinical report in 30% of the cases could be that either the lesions were too small to be appreciated, or alternatively, WMH were not felt to be clinically relevant and therefore not described. As there are a paucity of studies that have used systematic methods to assess and compare WMH between disorders, at times it can be difficult in clinical practice to determine if WMH are due to migraine or if they are better explained by another etiology. (24). Classification of WMH using the modified Scheltens scale may contribute to a better identification and understanding of the distribution of migraine WMH which could help to clarify the relationship between WMH and headache features.
The distribution and size of WMH in our migraine cohort seems to differ from that of focal WMH associated with lacunar infarcts. Ryu et al. (25) used a modified version of the Scheltens scale to assess lacunar infarcts and found a left hemisphere dominance of WMH in the corona radiata, basal ganglia, thalamus and internal capsule. Lesions seen in lacunar infarcts are usually small lesions that are hyperintense on T2 and FLAIR and have low signal on T1, correlating with pathological descriptions of gliosis and encephalomalacia respectively (26). The absence of low signal on T1 in this migraine cohort might imply that these lesions are not due to encephalomalacia. Similar to lacunar infarcts, multiple sclerosis lesions are bright on T2 and FLAIR images and can demonstrate low T1 signal when the lesions are chronic (27). The absence of low T1 signal in lesions seen in our migraine cohort distinguishes the WMH lesions from lacunar infarcts and multiple sclerosis. It is interesting that even in older migraine patients WMH are not low signal and therefore may not represent a final common pathway of encephalomalacia that is seen with lacunar infarcts and demyelination.
There are several limitations of our study. All subjects included in this study had brain MRIs ordered for clinical reasons. Thus, they might have specific characteristics that led the clinician to order an MRI (e.g. atypical symptoms, abnormal neurological examinations). Thus, the findings from this study might not be generalizable to the general migraine population. All subjects in this study were enrolled from a headache specialty clinic and the results might not be generalizable to the general population of people with migraine. It is not known whether some of the enrolled individuals had, vascular risk factors, or other conditions which might increase their risk of having WMH attributable to conditions other than migraine. Lastly, as the age range of this study cohort was broad, it cannot be ruled out that some of the WMH found in older subjects were age-related and not necessarily related to migraine.