Daily Migraine With Visual Aura Associated With an Occipital Arteriovenous Malformation


Address all correspondence to Dr. Egilius L.H. Spierings, 25 Walnut Street, Suite 102, Wellesley Hills, MA 02481-2106.


A 51-year-old woman with daily attacks of migraine with visual aura is described. The aura always occurred on the right and the headache always on the left side of the head, suggesting a structural lesion in the left occipital lobe. The lesion appeared to be an arteriovenous malformation of which almost full obliteration resulted in a decrease in frequency of the aura and in intensity of the headache. Subsequent treatment of borderline hypothyroidism with levothyroxine brought about a dramatic improvement in frequency of both the aura and the headache. The case is discussed in the light of our present understanding of the pathogenesis of the migraine attack.

The daily occurrence of attacks of migraine with aura, as recently described in this journal by Merims and Kuritzky,1 is rarely seen in medical practice and is probably also rare in the general population. The only case that I have encountered in over 15 years of full-time headache practice was associated with an occipital arteriovenous malformation. Stereotactic gamma-knife radiosurgery of the malformation decreased the frequency of the aura, and the intensity but not the frequency of the headache. Subsequent treatment of borderline hypothyroidism with levothyroxine brought about a dramatic improvement in frequency of both the aura and the headache.


A 51-year-old, right-handed woman experienced vomiting spells as a young child, which were so intense that they caused dehydration. However, she did not get headaches until after her second pregnancy when she was 26 years old. The headaches have been severe since their onset and preceded by a visual disturbance. The visual disturbance consists of horizontal, black, squiggly lines in the right visual field, which move downward as new ones are added from the top. It lasts for 10 to 15 minutes and is followed by headache within 5 to 10 minutes. The headache is located on the left, behind the eye and in the frontotemporal area. It is steady in nature but becomes throbbing when more intense. It builds to its maximum intensity in 20 minutes and lasts for 6 hours. It is associated with photophobia in both eyes and in one fourth, it is also associated with nausea; it is only occasionally accompanied by vomiting. The headache is made worse by light, physical activity, and coughing; lying down, sleep, and applying pressure or cold or both to the left eye make it somewhat better.

Initially, after their onset at aged 26, the attacks occurred infrequently, ie, once every 3 or 4 months, usually in the week following menstruation. When the patient was 39 years old, they occurred once every 2 weeks and when she was 44 years old, several times per week. During the year that followed, the attacks rapidly increased in frequency and when she was 45 years old, they became daily (Figure 1). When the attacks occurred daily, the headache was present on awakening in the morning 1 or 2 days per week. Otherwise, the attacks came about during the day but not at any particular time. The headache was always preceded by the visual disturbance and the visual disturbance, in turn, almost invariably by an urge to defecate. The headache woke her up at night infrequently, ie, only once or twice per month. She did not experience headaches other than those described above. For treatment of the headaches, she was taking ibuprofen 800 mg, one or two tablets per day, or Midrin (a combination of isometheptene, dichloralphenazone, and acetaminophen), two capsules per day; tramadol 50 mg, two to four tablets per day; and acetaminophen 500 mg with codeine 30 mg, six to eight tablets per day.

Figure 1.—.

Estimated attack frequency per year (with exponential trend line) from aged 36 to 45 years in a patient with progressive migraine with aura.

The patient's past medical history was remarkable for an appendectomy at aged 7 years. She stopped menstruating at aged 50 years and was not on estrogen therapy because of worsening of the headaches. She was taking medroxyprogesterone, 2.5 mg once daily, for hot flashes with benefit and without impact on the headaches. Her family history was significant for migraine aura without headache in her mother, who had had the visual disturbance for 25 years with the distinct difference that it alternated sides. Her father and sister were free of migraine but one of her two children experienced migraine attacks infrequently. The patient's physical examination was remarkable for obesity and nonpitting pedal edema; the neurological examination revealed a very soft bruit over the left carotid artery and orbit. The latter finding, in addition to the fixed and crossed lateralization of the aura and headache, led to neurodiagnostic imaging, which confirmed the presence of an arteriovenous malformation in the left occipitotemporal region (Figure 2). The malformation was treated twice with stereotactic gamma-knife radiosurgery, which caused its almost full obliteration. However, the headaches continued to occur daily although they were overall less intense. Rather than almost daily, they were now preceded by the visual disturbance in 40%, ie, about half of the times that the headache came about during the day.

Figure 2.—.

Magnetic resonance image of the brain, T1-weighted coronal view, showing a relatively large arteriovenous malformation in the left occipitotemporal region.

The patient was prescribed metoclopramide, 10 mg orally, to be taken at the onset of the visual disturbance, followed in 10 to 15 minutes by Midrin, two capsules at the onset, followed by one every half hour, with a maximum of six capsules. She was prescribed indomethacin 50-mg suppository, one every half hour, with a maximum of four, for rescue or when the headache was present on awakening in the morning or woke her up at night. For particularly severe headaches, she also required Cafergot, a combination of ergotamine and caffeine, one third of a suppository, preceded by prochlorperazine, 25 mg rectally, because of nausea.

As she was feeling very tired despite adequate sleep, felt cold in general, and had gained 50 pounds over a 10-year period with nonpitting pedal edema on the examination, a thyroid function test was performed. It revealed a free thyroxine level of 0.9 ng/dL (reference range, 0.8 to 1.8), thyroid stimulating hormone (TSH) level of 2.4 μIU/mL (reference range, 0.3 to 5.0), and negative thyroid antibodies. The free thyroxine and TSH levels were interpreted as suggesting borderline hypothyroidism and the patient was treated tentatively with levothyroxine, .05 mg per day. Unexpectedly, a week later, the attacks began to improve, with a gradual decrease in frequency of the aura and a decrease in intensity of the headaches (Figure 3). After a month, the attacks no longer occurred daily but 6 to 10 days per month, with two thirds of the headaches preceded by the aura. She also felt more energetic, was able to do more during the day, and felt less inclined to take a nap, which she used to do almost daily. A repeat thyroid function test revealed a free thyroxine level of 1.1 ng/dL and a TSH level of 0.9 μIU/mL. The dose of the levothyroxine was increased to .075 mg per day, which resulted in a further decrease in frequency of the attacks to 2 or 3 days per month. The free thyroxine level was subsequently found to be 1.2 ng/dL and the TSH level 0.5 μIU/mL, at which time the dose of the levothyroxine was left unchanged. The pattern of occurrence of the attacks has remained stable over the 2 years since, with 80% of the headaches preceded by the visual disturbance.

Figure 3.—.

Headache frequency in days per week. Treatment of borderline hypothyroidism was initiated at week 6 with levothyroxine .05 mg per day; the dose was increased to .075 mg per day at week 31.


In their paper on daily migraine with aura,1 Merims and Kuritzky described five patients between the ages of 50 and 80 years, who at one point in time developed daily attacks. At least three had a history of nondaily attacks, ranging from once every 2 months to once or twice per week and in at least two, the transition to daily attacks was acute. The aura was visual in all five patients, lasted for 5 to 60 minutes, and consisted of zigzag lines (two patients), circles (one patient), negative scotomas (one patient), and blurring (one patient). However, blurring of vision is not a neurological but an ophthalmological symptom and is a common associated symptom of headache in general.2 In my opinion, when it occurs without other visual symptoms, it does not qualify as a migraine aura sensu stricti, although even inability to focus has been considered as such.3 It may be more a manifestation of eyestrain leading into the headache, as I described previously under “mimics of migraine with aura.”4 All patients had computed tomography of the brain, but no information is provided as to whether the scans were performed with contrast to exclude a vascular lesion, such as an arteriovenous malformation. Treatment was not particularly successful in any of the patients, although a temporary benefit with phenytoin was reported in two.

In the patient described here, the fixed and crossed lateralization of the aura and headache suggested a structural lesion in the left hemisphere. The visual nature of the aura localized the lesion to the occipital lobe and the bruit over the left carotid artery and orbit pointed to it being vascular. The association of migraine with aura with a cerebral arteriovenous malformation is intriguing and raises more questions than it answers.5 My mentor in headache management, John R. Graham, MD, MACP, believed that in such cases, the arteriovenous malformation is not the cause of the migraine but determines its presentation in terms of lateralization of the aura and headache. Migraine is determined by genetic factors and this also seems to be the case in the above patient: her mother and one of her two children suffered from migraine, although the mother only experienced the aura. In the mother's case, the aura alternated sides as is more typical of migraine and suggestive of a functional rather than a structural disorder. Migraine is par excellence a functional disorder but in the above patient, it might have been given the appearance of a structural one by the presence of the arteriovenous malformation.

The mechanism underlying the migraine aura is currently thought to be Leão's spreading depression,6 which is a wave of excitation followed by prolonged depression of cortical neuronal activity. However, it is not known what triggers the spreading excitation and how it causes the headache when it follows the aura, assuming that there is a causal relationship between the two. The aura and headache typically occur sequentially, with one following the other within a tight time frame, and this sequential occurrence has been considered indicative of a causal relationship between the two. However, there is no evidence that the migraine aura and the headache are causally related to each other and that the aura is, indeed, the cause of the headache. The traditional view is that the migraine aura is caused by localized cerebral ischemia resulting from vasospasm, which is followed by reactive vasodilation. The cerebral vasodilation is supposedly accompanied by extracranial vasodilation, which, in turn, causes neurogenic inflammation in the extracranial tissues through stretching of the perivascular nerve fibers. It is ultimately the neurogenic inflammation that turns the extracranial vasodilation into the painful headache of migraine.

Graham believed that in the presence of an arteriovenous malformation, the hemodynamic changes within it cause the localized cerebral ischemia that brings about the migraine aura. As there is no evidence that cerebral ischemia is a trigger of Leão's spreading depression, I am inclined to believe that the presence of an arteriovenous malformation forms an irritative cerebrocortical focus, which in migraineurs further lowers the local threshold for the aura to occur. With regard to the aura-headache relationship, the consistent crossed occurrence of the two in the above-described patient makes it hard to believe that they are not causally related. This is different in migraine with aura in general, where not only alternation of the aura and headache is seen but also “neurological non-sense” in terms of the aura and headache occurring on the same side. Also, typically in a patient with migraine with aura, the headaches are not always preceded by aura and the aura may not always be followed by headache. The former is explained by the aura phenomenon occurring in a so-called clinically silent area of the cerebral cortex. Evidence for this may be provided by the four case studies7-9 in which attacks of migraine without aura were observed to be accompanied by “spreading oligemia,” the supposed equivalent in humans of spreading depression in animals. This assumes that difficulty coordinating the movements of one eye and inability to focus do not qualify as aura symptoms. Aura not followed by headache as in migraine aura without headache has even defied any attempt to explain it.

It is noteworthy that in the above patient, the almost full obliteration of the arteriovenous malformation by gamma-knife radiosurgery was followed by a partial disconnecting of the aura and the headache. A similar situation is seen in many patients with migraine with aura, ie, the headache is not always preceded by the aura. It is difficult to accept that this was due to a partial displacement of the aura phenomenon to cortical areas from which no symptoms emerge. The only other explanation could be that the aura and headache are not causally related, that both are initiated by the same mechanism, and that obliteration of the arteriovenous malformation decreased the intensity of the irritative focus and the likelihood of the aura to occur. This still leaves the question as to how the headache is related to the arteriovenous malformation, which it consistently occurred ipsilateral to. In the above patient, the headache responded to Cafergot, which contains the potent vasoconstrictor, ergotamine, and does not have analgesic properties per se. The headache must, therefore, be vascular in nature while in its presentation, it is indistinguishable from migraine. It is possible that the arteriovenous malformation only determined its lateralization and because the headache is vascular rather than neuronal, it occurred ipsilateral rather than contralateral. In this regard, it is interesting that in this patient, part of the blood supply of the arteriovenous malformation, although small, went through the external carotid artery. This may have affected the hemodynamics within this arterial system in such a way that the vascular changes of migraine were more likely to take hold here.

In our patient, we had relatively good records to track the increase in attack frequency over time, which showed it to be exponential rather than linear. This is the type of progression that is generally seen when intermittent migraine or tension-type headaches develop into (almost) daily headaches over time. The increase in headache frequency over time is often attributed particularly to the regular use of analgesics or vasoconstrictors for headache, in the latter category, especially that of caffeine. However, in migraine with aura, it is difficult to see that this is the case because while the headache may be analgesic or vasoconstrictor sensitive or both, the aura surely is not. In addition, Mathew et al10 observed that almost half of the patients with migraine with aura who develop “chronic daily headache” first transform into migraine without aura. The progression of migraine with aura as took place in the above patient, more likely involves a mechanism akin to kindling as it is known from animal-experimental studies of epilepsy.

Finally, the response of the migraine condition to treatment of borderline hypothyroidism with levothyroxine deserves attention because it was unexpected and dramatic. The possibility of it being due to “placebo effect” or, more correctly, due to nonspecific aspects of treatment cannot be excluded but is, in my opinion, unlikely because of its dramatic and sustained nature. An explanation of the benefit of the treatment, at least with regard to the migraine aura, could lie in the decrease in cerebral excitability associated with correction of hypothyroidism.11 With regard to the benefit from treatment, attention should also be paid to the time of onset of the attacks following pregnancy. Pregnancy is associated with a temporary boost in thyroid function, due to the stimulating effect of the placental hormone, human chorionic gonadotropin, on the thyroid gland. This effect ceases after pregnancy, resulting in a drop in thyroid function, which may have triggered the occurrence of the attacks in the patient. The attacks subsequently did not abate until thyroid hormone was added to the system and a condition of borderline hypothyroidism was alleviated.

Acknowledgment:  Appreciation is expressed to my wife, Malina, for editorial assistance in the preparation of the manuscript.