First, we could replicate the finding that patients with migraine show a higher incidence of WMH than age- and sex-matched healthy control subjects. This finding is however unspecific since both migraine patients and healthy subjects did not show any infarctions or bleedings.
The study was primarily aimed to clarify the role of WMH in differentiating migraine subtypes and its relation to clinical characteristics of migraine and MRI signs of cerebrovascular lesions, i.e. asymptomatic infarctions in young patients without vascular risk factors and other causes of cerebrovascular diseases. Eligibility criteria were designed to eliminate patients with any possible causes of WMH, subclinical small brain infarctions and microbleedings except migraine. The rationale for the study was the ambiguity of evidence supporting diagnostic and prognostic significance of WMH in migraine11, 13, 15 that may play a role in differential diagnosis and risk assessment of vascular complications. WMH of vascular origin are the most widespread and non-specific MRI sign 17, 49–52. It is difficult to differentiate WMH as a sign of migraine or as of vascular origin, and this may partially explain the inconsistency between the results of previous studies.
In this study, the incidence of WMH in different migraine subtypes was 38.7–44.4% that complies with the middle position within the previously reported values between 29% and 73%4–8. The incidence of WMH was non-significantly higher in migraine with aura patients (44.4%), similarly to the results of other studies5, 6, 9, 10. No significant differences in the incidence of WMH, its distribution across cerebral lobes and size of lesion between groups were found. Universal differentiating features for all types of migraine were: predominance of WMH within the frontal lobe and a lower incidence of lesions within the parietal and temporal lobes and their absence within other lobes; equal incidence of juxtacortical and deep location along with less frequent periventricular localization; small (2.5 [1.5; 3] mm) lesion size; the absence of a tendency to merge. Previous studies have reported not only the localization of WMH within frontal, parietal and temporal lobes9, 13, 12, 18, 26, 30–32, but also their infratentorial and occipital location11, 18, 30, 34, the involvement of periventricular white matter6, 8, 10, 13, 26, various sizes of WMH, including confluent lesions31.
It should be noted that our approach to evaluate the white matter allowed us to conclude that WMH in migraine are characterized by juxtacortical localization which was previously mentioned only in few studies20, 31, 32. We also demonstrated a direct association of the increase of juxtacortical WMH lesions within the frontal lobe with age of the patients and duration of the disorder which most probably indicates a role of mechanisms which determine the frontal localization of pain during migraine attack, and the extravasation liquid part of the blood during an attack, in the development of these lesions.
Thus, the pattern of WMH in our population of young migraine patients without vascular risk factors and other causes of cerebrovascular lesions differs from the results of other studies that did not implement such restrictions in patient selection. These differences include the absence of the following signs of WMH: disseminated distribution; periventricular localization; absence of tendency to confluence (i.e., MRI signs described in older migraine patients and signs of age-dependent SVD)43,53. The influence of age and vascular factors on the disseminated distribution of WMH in migraine was also confirmed in a recent study by Meilan A. et al. (2020). Among the investigated factors, only the age > 45 years was associated with the severity of WMH, while the type of migraine and the frequency of attacks did not26.
Despite different pathophysiological mechanisms underlying different migraine subtypes54, they are characterized by similar pattern of WMH. This may be explained by a presence of a common link of WMH pathogenesis, maybe associated with vasodilation and plasma extravasation. In our study, we did not find any relationship between severity of WMH and the severity of migraine as it was demonstrated in previous studies4, 11, 13, 14, 18, 19, 24–26. The exclusion of vascular risk factors allowed us to suggest that there were no other causes leading to the increased permeability of the blood-brain barrier other than migraine; therefore, it can be assumed that the underlying mechanisms of WMH pathogenesis include individual properties of the endothelium associated with high vascular wall permeability. This predisposition may be related to a disturbance in the production of the Vascular Endothelial Growth Factor (VEGF). It affects the vascular permeability and leads to an increase in the level of nitrogen oxide by stimulating the nitrogen oxide synthase with consequent vasodilatation 55. Earlier, elevated VEGF levels during a migraine attack 56 and decreased VEGF levels during the interictal period 57 were demonstrated, as well as the association between the genetic profile of VEGF and different sensitivity to migraine 58. The revealed similarity of WMH pattern in different migraine types confirms that specific features of activation and sensitization of trigeminovascular pathways, brain stem and diencephalic nuclei play an important role in the development of various migraine subtypes 59, 60.
In our study groups with different types of migraine, no subclinical brain infarctions were seen, this is consistent with the role of vascular risk factors in their development. This may be also somewhat supported by the fact that in a Cerebral Abnormalities in Migraine, an Epidemiological Risk Analysis study (CAMERA), which established an association between migraine and asymptomatic infarctions in the posterior circulation, the age of the patients was higher 5, 13, 18, 33, 34. We also did not detect any microbleedings, previously described in older patients with migraine 35. Although in the aforementioned study microbleedings were more common in migraine patients than in the control group35, it should be noted that the proportion of patients with hypertension, a leading cause of small vessel damage, was also significantly higher migraine patients. Therefore, it could not be confirmed that migraine was the sole cause for the higher incidence of microbleedings in this group. Since the leading mechanism of microbleedings is the increased permeability of blood-brain barrier, which is typical for both the headache phase of migraine attacks54 and SVD (including hypertensive and other age-dependent types of this disorders) 50–53, the combined influence of these conditions cannot be excluded either.