Several recent studies have shown that in migraine patients, the prevalence of an incomplete Circle of Willis is higher than in controls.1-3 This might suggest that such an anatomic anomaly is a risk factor for developing migraine. As an explanation, it has been commonly proposed that an incomplete Circle of Willis could prevent regional cerebral blood flow adaptation in the face of increased metabolic demand. A resulting local ischemia could then, in hyperexcitable persons, lead to a cortical spreading depression that is by many investigators believed to be the cause of migraine attacks.
Unfortunately, these studies did not measure regional cerebral blood flow during increased brain activity or, eg, a one-sided carotid artery occlusion. We doubt whether in subjects with an incomplete Circle of Willis such conditions will importantly diminish cerebral perfusion. In this paper, we argue that an incomplete Circle of Willis might increase the risk for migraine by elevated wall shear stress in small-diameter anastomotic vessels.
Aberrations in the Circle of Willis do not automatically cause a dramatic reduction of anastomotic flow to the brain. Only a simultaneous absence of both the anterior communicating artery (or an A1 segment) and a posterior communicating artery (or P1 segment) would definitely prohibit flow compensation via the Circle of Willis (see Figure). The presence of this combination seems, however, rare because it was not observed in a study on 360 normal fixed brains.
Furthermore, even a moderate decrease of cerebral blood flow (from 47 to 37 mL/100 g/min) is sufficient to attenuate attention and motor reactions. It is hard to believe that in half of the population, having similar morphologic variants, such signs of diminished brain function, will be elicited by increased brain activity or by head rotation. Indeed, occlusion of a carotid artery in patients with incomplete Circle of Willis, anesthetized for open arch surgery or kept conscious for carotid endarterectomy, did not elicit signs of transient cerebral ischemia, even not in the presence of severe bilateral carotid artery stenosis.[6, 7]
In addition, secondary collateral pathways are formed by the external carotid artery via the ophthalmic artery and via leptomeningeal anastomoses at the brain surface.
An imminent shortage of regional cerebral blood supply will obviously be corrected by a rise of blood flow velocity in the patent portions of the circuit, as a result of an increased pressure gradient. An example of this is observed in a study with carotid occlusion in healthy volunteers, showing increased flow velocity, as measured by transcranial color-coded duplex sonography, in a posterior communicating artery.
An increased flow velocity may enhance wall shear stress. Presence of abnormally high wall shear stresses in the Circle of Willis is well-known and thought to be one of the main pathogenic factors in the development of saccular cerebral aneurysms in that region. The increased prevalence of migraine in patients with an unruptured cerebral aneurysm (about 40% compared with 8.8% of controls) suggests that high shear stress is also a factor in the genesis of migraine.
In a model-based study of a complete Circle of Willis shear stress in a 2-mm wide, anterior communicating artery was calculated to be 72 Pa during partial obstruction of one carotid artery. It is to be expected that this value will be even higher in an incomplete Circle of Willis wherein one or more vessels have a much smaller diameter. These values by far exceed those known to elicit agonist-independent platelet aggregation in-vitro (8 Pa in platelet-rich plasma and 31.5 Pa in non-anticoagulated blood flowing from an anticubital vein). In vivo, high shear stress is believed to cause formation of unstable platelet aggregates especially in the presence of endothelial dysfunction (as in stenotic arteries) and increased platelet aggregability. These conditions are, next to other genetic vasculopathies, often present in migraine patients. Endothelial dysfunction is specifically found in the territory of the posterior cerebral artery, exactly that region that is supplied by the portion of Circle of Willis that is most often incomplete.
Literature on increased platelet aggregability in migraine patients has been surveyed in a recent paper. In that paper, we have documented the theory that shear-induced platelet aggregation in stenotic or otherwise narrowed vessels to the brain might elicit a, mostly unilateral, migraine attack by local release of platelet serotonin. This neurotransmittor is a known pain mediator and may in high concentration (resulting from strong platelet aggregation) induce constriction of extracerebral arteries. A subsequent progressing regional hypoxia might be responsible for aura signs and diminished pain sensation. In a lower concentration (resulting from diminishing or initially less strong aggregation), serotonin will cause a long-lasting vasodilation and pain. The importance of platelet aggregation in the pathology of migraine is supported by the fact that many migraine triggers enhance platelet aggregability, while a number of well-known anti-migraine medicines have a platelet-inhibiting effect.
Noteworthy are recent anecdotical messages from migraine patients who temporarily used clopidogrel for platelet inhibition after a cardiac intervention: a relieve of migraine during, and reappearance after ending the medication. The efficacy of this medicine is presently tested in 2 randomized, placebo-controlled trials (see 17 for references). Of note, clopidogrel and other thienopyridines might be superior to aspirin in relieving migraine, as the latter only weakly inhibits shear-induced platelet aggregation.
Existence of elevated shear stress in the Circle of Willis may, in addition, clarify why the risk for small cerebral and cerebellar infarcts is increased both in subjects with incomplete Circle of Willis and in migraine patients. Aggregation of platelets might, except for inducing migraine by release of serotonin, cause small sites of ischemia by the shedding of micro-thrombi that travel to distal portions of the cerebral vasculature. The limited size of micro-infarcts makes it less likely that they would result from diminished blood flow in one of the large cerebral arteries.
In conclusion, in the presence of increased platelet aggregability and endothelial dysfunction, an incomplete Circle of Willis might predispose to migraine by elevated wall shear stress in small-diameter anastomotic vessels.