By Dara Jamieson, MD

Associate Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Jamieson reports she is a consultant for Bayer and Boehringer-Ingelheim.

SYNOPSIS: Hypoxia-induced migraine attacks with and without aura, in an experimental paradigm, were accompanied by dilation of cranial arteries in individuals who have migraine with aura, as well as in healthy control subjects.

SOURCE: Arngrim N, et al. Migraine induced by hypoxia: An MRI spectroscopy and angiography study. Brain 2016:139;723-737.

The interplay between migraine, alterations in cerebral blood flow, and changes in cerebral oxygen delivery continues to be explored in an effort to explain migraine both with and without aura. Danish investigators tested whether experimental hypoxia could trigger attacks of migraine, as well as metabolic and arterial circumference changes, in patients suffering from migraine with aura. Patients with migraine with aura, who had a typical aura associated with every migraine attack and who also had a minimum of one attack per month, were recruited from the outpatient clinic at the Danish Headache Centre. In a randomized, double-blind, crossover study design, 15 migraine with aura patients were exposed to 180 minutes of normobaric hypoxia (resulting in a capillary oxygen saturation of 70-75%) or to the sham inhalation of atmospheric air on two separate days. Also, 14 healthy controls, recruited via a Danish website for recruitment of volunteers for health research, were exposed to equivalent levels of hypoxia and underwent the same testing. Glutamate and lactate concentrations in the visual cortex were measured by magnetic resonance spectroscopy (MRS). The circumference of standardized extracranial and intracranial cerebral arteries was measured by 3T high-resolution magnetic resonance angiography (MRA). A T1 sequence on MRI ruled out underlying brain pathology.

Eight of 15 migraine patients (53%) experienced migraine-like attacks during hypoxia in comparison to one (7%) during the sham procedure (P = 0.039). With hypoxia, three patients reported migraine with aura, and five patients reported migraine without aura. The median time to onset of the migraine-like attacks was 105 minutes, with a median attack duration of 6 hours. One person in the control group of 14 had a migraine-like attack associated with hypoxia. Four out of 15 migraine patients experienced a short period of uncharacteristic visual disturbances, considered to be an atypical aura. For three migraine patients, the visual disturbance occurred during hypoxia, and in one patient it occurred after hypoxia.

Hypoxia did not change glutamate concentration in the visual cortex compared to sham. Hypoxia increased lactate concentration by 61%, as compared to by 7% during sham inhalation. The circumference of the cranial arteries increased with hypoxia. These changes were noted across all experimental groups, as there was no difference in the metabolic or vascular responses to hypoxia between migraine patients and controls. The authors extrapolated from their results that hypoxia may provoke migraine headache and aura symptoms in some patients.


These investigators found that hypoxia induced migraine-like attacks with and without aura in patients with migraine with aura, without detecting unique cerebral metabolic or vascular changes. Increase in arterial circumference was found due to hypoxia in both experimental groups, not correlating with migraine triggering. In some of the patients, atypical visual auras were also triggered by hypoxia, perhaps mirroring the migraine auras without headache that some migraineurs experienced. Their work adds to our knowledge of what can cause a migraine attack, and specifically a migraine with aura; but, the mechanisms behind the migraine-inducing effect of hypoxia is still not apparent as no unique correlates were found in those migraineurs with a hypoxia-induced attack. Hypoxia was examined because of the observation that migraine is especially prevalent in areas of higher elevation. However, hypoxia is not a common trigger for most migraine sufferers, and the induced levels of hypoxia were more decreased than in most habitable elevations.

The patients tested all had only migraine with aura, which is unusual as most patients with migraine with aura have auras associated with only a minority of their headaches. The individuals tested seemed to have a particular vulnerability to triggering of attacks that may not be representative of the general migraine population. The probability of inducing a migraine attack appears to be a function of the severity and the duration of the hypoxia, which in this experimental design was profound and prolonged. Another correlate of these hypoxia-induced attacks could be the abruptness of the homeostatic perturbation. Change in weather, sleep, eating habits, and stress can trigger migraine headaches. A revealing experiment would be to gradually decrease oxygen saturation, allowing for adjustment at each decremental level, to mitigate the effect of the abruptness of the change in oxygenation. Acclimatization to high altitude areas of relative hypoxia decreases triggering of headaches.

Despite some knowledge of its pathophysiology, the cause of migraine still remains mysterious, with investigators piecing together information like the three blind men examining different parts of an elephant. One method of investigation is to identify a trigger and then to see if the circumstances of triggering in a vulnerable migraine sufferer can lead to an understanding of why migraine attacks occur under other circumstances. This investigation looked at a relatively unusual trigger for most migraine sufferers. Other similar experimental protocols should be designed around more common triggers, such as sleep deprivation and red wine.