Vascular changes have a primary role in migraine

Introduction

In the ongoing debate about migraine pathogenesis it is crucial whether vascular factors are of primary importance or the blood vessels are just ‘bystanders’.

In Edward Liveing's the first monograph on migraine entitled ‘On Megrim, Sick-headache, and Some Allied Disorders: A Contribution to the Pathology of Nerve-Storms’ (1) he proposed the neural theory of migraine, ascribing the problem to ‘disturbances of the autonomic nervous system’, and that ‘vasomotor disturbance is secondary in origin’. The ‘neural’ theory has been revived in recent decades and some researchers have even proposed that migraine pain is caused by abnormal central interpretation of normal sensory input in the trigeminal sensory system. However, no scientific evidence supporting the theory of ‘abnormal central interpretation’ has yet emerged. Following a debate format I would like to present the scientific evidence supporting a primary role of vascular changes in migraine.

History

In the second century, Galen suggested that throbbing pain during headache originated from blood vessels. In 1672 Thomas Willis proposed the first vascular hypothesis of migraine. Graham and Wolff (2) provided the first observations in humans demonstrating that focal head pain may be elicited from both extra- and intracranial vessels. In 1940 Ray and Wolff (3) reported that stimulation or distension of the large cranial arteries evoked head pain associated with the feeling of nausea or sickness. More recently it was reported that focal headache may be induced by balloon dilatation of cerebral arteries (4). The similarity between referred pain locations following stimulation of large cerebral arteries and headache patterns in migraine provide the strongest support yet for the involvement of vascular nociceptors in migraine pain.

Migraine with aura

The results of studies of regional cerebral blood flow strongly suggest that aura is caused by cortical spreading depression (CSD) (5,6). Although we agree that the aura is a cortical phenomenon we do not know the pathophysiological mechanisms triggering the CSD or aura. The questions for discussion here is whether vascular changes may trigger aura and whether aura may trigger nociception and head pain.

To address the first question we have to go back to the first in vivo CBF study in migraine aura patients (5). In this study patients underwent a conventional angiography and aura symptoms or CSD was triggered immediately after the angiographic procedure. The most plausible explanation for aura triggering mechanism could be haemodynamic changes induced by puncture or by injecting a contrast agent into the carotid artery. In support, patients with internal carotid artery (ICA) dissection may present with transient symptoms resembling migraine with aura (7). Furthermore, migraine attacks with auras are associated with underlying hereditary cerebrovascular disorders, such as cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL). CADASIL is caused by an abnormal accumulation of mutated NOTCH3 in vascular smooth-muscle cells both in cerebral and extracerebral vessels (8). In recent years another interesting association has been suggested between migraine with aura and a patent foramen ovale (PFO) (9,10). Interestingly, the frequency of medium to large PFOs was shown to be greater among patients with cryptogenic infarcts than among infarcts of determined cause (11) and a recent population-based study reported that a moderate-to-large right-to-left cardiac shunt was an independent risk factor for posterior circulation territory brain infarcts and infratentorial hyperintense lesions in migraineurs with aura (12). Comorbidity of migraine aura with acquired or hereditary cerebrovascular disorders and PFO suggests a strong link between blood vessels and aura, and suggests that vascular changes may evoke aura attacks or CSD (13). In support, recent experimental data in mice demonstrated that cerebral microembolism triggers CSD (14).

The relationship between aura and head pain is complex and difficult to study in humans. It could be speculated that primary vascular changes causing aura may also activate sensory trigeminal afferents. To date, there is no human evidence to support this hypothesis. The study by Lauritzen et al. (15) reported significantly reduced vascular reactivity to CO₂ in the hypoperfused parts of the brain after aura, but to what extent this contributes to headache after aura is unknown. Recent animal studies investigated possible mechanisms linking CSD and headache. Using a novel method, Bolay et al. (16) provided the first evidence that CSD might cause unilateral plasma protein leakage within the overlying dura and brainstem activation, both blocked by trigeminal denervation. Another model developed by Burstein and colleagues took Bolay et al.’s work a step further, by recording single-unit activity of meningeal nociceptors (17) and central trigeminovascular neurons (18) before and after CSD. These studies demonstrated that CSD could lead to long-lasting activation of nociceptors and activation of central trigeminal neurons. Taken together these data provide crucial evidence on the relationship between CSD and head pain.

Migraine without aura

For more than 100 years and after the first report by Graham and Wolff of extracranial vasodilatation during migraine attacks (2), the vascular hypothesis of migraine has received support. In the following I will review the scientific data supporting the involvement of vascular changes in migraine pain.

Provocation experiments have shown that vasoactive substances, such as glyceryl trinitrate (GTN) (19), calcitonin gene-related peptide (CGRP) (20) and pituitary adenylate cyclase activating polypeptide-38 (PACAP38) (21), induce initial dilation and delayed migraine-like attacks in migraine suffers indistinguishable from their spontaneous attacks. Interestingly, unlike GTN, CGRP and PACAP38 do not cross the blood brain barrier but both still provoke migraine attacks.

In 1991 Friberg et al. (22) reported that migraine headache was associated with middle cerebral artery (MCA) dilatation (20%) on the headache side and that the MCA velocity on the headache side returned to normal after successful treatment with sumatriptan. Thomsen et al. (23) showed decreased velocity in the MCA on the pain side of half-sided migraine attacks and Iversen et al. (24) found the diameter of the superficial temporal artery to be larger on the pain side than on the pain-free side during attack. Recently, we used high-resolution 3T magnetic resonance angiography to examine vascular changes during CGRP-provoked migraine attacks without aura (25). Diameter changes of middle meningeal artery (MMA) and MCA were examined before and during delayed migraine attacks. In addition, diameter changes of arteries were recorded 20 minutes after subcutaneous sumatriptan administration. The study reported that migraine attacks were associated with dilatation (∼12%) of both MMA and MCA. We found that in half-sided headache there was dilatation on the headache side and in double-sided headache side there was bilateral dilatation. Sumatriptan administration resulted in bilateral MMA contraction and amelioration of the headache while MCA remained unchanged (25). These data clearly demonstrate that extra and intracranial arteries are only dilated on the pain side during half-sided migraine attacks, and that bilateral dilatation is observed in double-sided headache during CGRP induced migraine attacks. Furthermore, these results supported by previous animal studies suggest that sumatriptan exerts its anti-migraine effect by a combined constriction of extracerebral arteries and inhibition of nociceptive input from these arteries.

Is mechanical dilatation sufficient to activate nociceptors and cause migraine pain?

Arteries may dilate markedly, such as during blood pressure decreases, and heart rate may increase during physical exercise without accompanying head pain. Therefore, dilatation alone cannot explain migraine pain. Experiments in animals suggest that pain-sensitive nerve fibres can be activated by leakage of sensitizing vasoactive substances from trigeminal nerve terminals or by efferent activity in parasympathetic nerves. Given these data, it is plausible to suggest that the release of vasoactive substances around brain vessels would cause sensitization of trigeminal afferents, vasodilatation and migraine pain.

Conclusion

Both my distinguished opponent Professor Andrew Charles and I agree that migraine pathophysiology is complex and there is no simple and definitive support pro or contra vascular changes. It is necessary to keep searching for new scientific data to explore the pathophysiological mechanisms underlying migraine. At present we can conclude that ‘…migraine cannot be understood without a clear understanding of the dynamic role of the blood vessel in its pathogenesis’ (26).