Abstract
There is no question that significant vascular changes may occur during a migraine attack. The question is whether these vascular phenomena are a cause or a consequence of migraine pathophysiology and migraine symptoms. While the simplistic appeal of the ‘vascular hypothesis’ of migraine has had remarkable staying power, studies of the clinical features of migraine, and its physiological and pharmacological mechanisms, have provided strong evidence that effectively refutes a primary role for the vasculature in migraine pathogenesis.
It is important to start with the concept of migraine as not merely a headache, but rather a multi-faceted clinical phenomenon that can occur over hours to days. Diverse prodromal symptoms including yawning, mood change, neck pain and polyuria, among others, are common, and electronic diary studies have shown that these symptoms may occur up to several hours before the onset of headache (1). Aura symptoms include both positive and negative visual phenomena, and characteristic sensory or language dysfunction. Nausea, light and sound sensitivity, mechanical allodynia, vertigo and fatigue commonly occur before, during, or after headache. These symptoms indicate elaborate changes in brain chemistry and physiology – there is no reasonable explanation of these symptoms based on a primary vascular mechanism.
Regarding migraine headache, the ‘vascular hypothesis’ is centered on the premise that migraine pain is caused by vasodilation. A variety of imaging and pharmacological studies, however, have directly challenged this concept. Pioneering blood flow studies performed by Olesen and colleagues (2) found that a migraine attack may include both hypo- and hyper-perfusion phases. These studies found that the headache component of the attack begins during the hypoperfusion phase and continues into the hyperperfusion phase, but the hyperperfusion may continue long after the headache has ended. Put simply, there is no correlation between blood flow changes and headache. Recent magnetic resonance angiography (MRA) studies have similarly shown no correlation between changes in vascular caliber and headache in either evoked or spontaneous migraine. Schoonman et al. (3) found that while nitroglycerin did evoke some cerebral and meningeal vasodilation, migraine headache did not occur until after the caliber of vessels had returned to baseline. Nagata et al. (4) reported that spontaneous migraine was not associated with any dilation of the middle meningeal artery as measured by MRA. Rahmann and colleagues (5) showed that vasoactive intestinal peptide is a potent dilator of cerebral vessels, yet does not cause headache. Conversely, sildenafil is known to evoke headache and migraine, but does not cause cerebral vasodilation or alteration in cerebrovascular function (6). Taken together, these studies clearly indicate that vasodilation is neither necessary nor sufficient to cause migraine headache.
It has also long been assumed that the pulsating quality of migraine pain is a reflection of vascular pulsation. Recent studies by Ahn (7), however, call this assumption into question, suggesting that in most patients the rate of migraine pain pulsation is in fact slower than the arterial pulse rate. In some patients, the rates of pain pulsation were similar to the arterial pulse, but the two rates were observed to come in and out of phase over time. These studies raise the possibility that pain pulsation could be driven by a mechanism involving neural oscillations rather than by perception of the arterial pulse.
Multiple acute and preventive medications such as caffeine, ergotamines, triptans, beta-blockers and calcium channel blockers have in the past been presumed to work via vascular mechanisms, and vasoconstriction had been assumed to be a primary mechanism of acute migraine therapy. But each of these medications has alternative mechanisms involving either nociception or brain excitability that can explain their therapeutic efficacy. Triptans, for example, inhibit nociceptive neuronal activity in the trigeminal nucleus caudalis, periaqueductal gray, and thalamus (8). In addition, multiple medications without consistent vascular effects are effective migraine therapies. Topiramate and divalproate sodium have no direct vascular effects on the vasculature. Aspirin and other anti-inflammatories are effective acute therapies with no effects on vascular caliber. CGRP receptor antagonists also have clearly demonstrated efficacy as acute treatments for migraine, but do not constrict blood vessels. Vascular mechanisms are therefore not required for medications to be effective in the acute or preventive treatment of migraine.
Another tenet of the vascular hypothesis has been that vascular changes trigger nociceptive responses via peripheral inputs through the trigeminal nerve. There has not, however, been any direct evidence for a role for the peripheral trigeminal system in what has long been assumed to be a ‘trigeminovascular’ phenomenon. An important case report by Matharu et al. (9) supported by personal unpublished observations of other cases, definitively demonstrates that cluster headache, the quintessential ‘trigeminovascular headache’ can occur in the setting of complete peripheral trigeminal denervation. Although there are obvious limitations associated with case reports, and cluster headache is not migraine, the persistence of triptan-responsive headache in the absence of peripheral trigeminal input nonetheless represents a significant challenge to the traditional idea of migraine as a ‘trigeminovascular headache’.
The denouement of the longstanding debate regarding the role of the vasculature in migraine may be the understanding that no simple hypothesis is capable of explaining such a complex biological and clinical phenomenon. While extensive evidence indicates that the traditional view of a primary role for the vasculature in migraine may be overly simplistic, it would also be misguided to conclude that the vasculature plays no role at all. For example, some laboratory studies indicate that a disruption of normal neurovascular coupling may be an important migraine mechanism (10). Even if vascular changes are not the primary mediators of migraine, it is nonetheless important to consider the possibility that blood vessels are a significant piece of this intricate pathophysiological puzzle.