Persistent Migraine Aura: New Cases, a Literature Review, and Ideas About Pathophysiology

Sam Thissen, MD; Iris G. Vos, MD; Tobien H. Schreuder, MD; Wendy M.J. Schreurs, MD; Linda A. Postma, PhD; Peter J. Koehler, PhD


Headache. 2014;54(8):1290-1309. 

In This Article


We reviewed data of 47 patients with PMA and with this number attempted to discern any patterns within this remarkable phenomenon. There is a broad variation in age of onset, with an average age of 30. The majority of the patients were women. Both age and female preponderance may be explained by the prevalence of MA in the general population and not be a characteristic of PMA itself.[31] The wide variation in duration of symptoms makes it difficult to predict the prognosis and outcome of these symptoms.

Accurate history taking is the most important diagnostic tool. A normal neurologic, as well as ophthalmologic examination, is required to diagnose PMA, although visual field exam could be abnormal because of negative symptoms (ie, scotoma) as part of the aura. Obtaining an MR of the brain is necessary to exclude infarction and other pathology.[6] This was the case in 44 (94%) patients. Therefore, we cannot be completely certain of the diagnosis PMA in three cases. In one case there was a CT scan without signs of infarction. In one case no information about imaging studies was provided. In our first case there was some doubt about a cortical infarction on the left. However, this possible lesion could not explain the patient's symptoms and was therefore considered an incidental finding. The fifth patient that we presented above, and who was excluded from the series, is illustrative in a way that typical PMA-TA may be present with a small infarction and therefore demonstrating that MR is necessary and should be studied carefully. This kind of lesions could easily escape detection on CT and we wondered whether more powerful MR scan would enable detection of lesions in the other patients.

Lance and Goadsby made a distinction between PMA-TA and PMA-PPVD.[5] There were more PPVD (27) cases than TA (19) in our review. The precise significance of this distinction is unknown, but there may be a difference in underlying pathophysiology and treatment efficacy. Besides a significant difference in duration of symptoms, we have been unable to find any differences between these groups in the present review (Table 3).

Previously, Lashley's experience of his own visual aura led to his proposal that the aura was due to a spreading abnormality that migrated over the visual cortex at a rate of 3–5 mm per minute.[32] Presently, visual aura is thought to be associated with the phenomenon of cortical spreading depression (CSD) usually starting in the occipital cortex as described by Leao. CSD is a neuronal depolarization wave with subsequent suppression of electrical activity that moves across contiguous cortical areas at a rate of about 2–5 mm per minute after mechanical or chemical perturbation of the cortex in experimental animals.[33] CSD is followed by changes of the regional cerebral blood flow (rCBF). During the aura phase, patients develop rCBF reduction (oligemia), which does not reach critical values (ischemia). Oligemia gradually spreads anteriorly in the course of 15–45 minutes.[34] Studies during attacks showed alterations in relative cerebral blood flow, cerebral blood volume, as well as tissue MTT in the grey matter of the occipital cortex contralateral to the affected visual hemifield and, more rarely, hypoperfusion of the whole hemisphere.[35] Hadjikhani et al reported a case of spontaneous migraine studied with functional MRI techniques, revealing a slow neuronal change in the occipital cortex, moving forward at a rate of 3–6 mm per minute, during the aura phase.[36] These data might create expectations with respect to functional studies in PMA.

Although the specific pathophysiology of PMA without infarction is still unknown,[17] several theories on the underlying pathophysiological mechanism have been proposed. These include abnormal energy metabolism in the brain, brain magnesium levels being significantly lower than that of non-migraineurs, greater reactivity of N-methyl-D-aspartate receptors to glutamate, a lowered threshold to CSD, and loss of inhibitory GABA-ergic interneurons; they may combine to cause sustained visual aura.[14] Other authors oppose a sustained hyperexcitability of the visual cortex without significant dynamic modulation. This study also suggests a pathophysiologic link to sustained excitatory links possibly related to reverberating CSD.[37]

Several studies have been undertaken to investigate these underlying pathophysiological hypotheses in PMA by using Tc99m-HMPAO-SPECT, FDG-PET and MR–perfusion-weighted imaging (PWI).[4,13–15,18,21,38] As reviewed above, these investigations show conflicting data.

Eleven PMA cases have been studied with Tc99m-HMPAO-SPECT; they show inconsistent data.[4,7,12,14,25] In five cases there was bilateral hypoactivity, and in four cases hemi-hypoperfusion. Only in two cases the SPECT showed normal activity of the bilateral occipital cortex. Occipital hypoperfusion on SPECT might be due to reduced neuronal activity and metabolism during the supposed reverberating waves of spreading depression. Focal hypoperfusion might not be the true pathogenetic origin.[17] Nevertheless, this indicates that the occipital cortex is involved in the pathogenesis of persistent visual auras. The inconsistency of the data presented might be explained by the difference in clinical symptoms. One could think that a typical aura in one hemifield would show hypoperfusion on the contralateral side, whereas symptoms in both visual fields may be correlated with bilateral hypoperfusion. However, we could not identify such an association. Despite these findings, this still does not exclude such an association. Findings with SPECT seem subjective and difficult to interpret with a lack of standardization. This makes it difficult to interpret the presented results.

Four cases have been studied with FDG-PET.[18,23] Three showed normal results. In one case there was bilateral hyperactivity. Mathew et al found hyperactivity in the occipital region, indicating hypermetabolism of the contralateral cortex in patients with migraine aura.[38]

Only one study, using MR-PWI, could identify hypoperfusion during PMA.[14] However, no conclusions can be drawn from these conflicting data. Revision and standardization of the imaging protocols could possibly lead to other results. Therefore, further studies are needed to establish the role of Tc99m-HMPAO-SPECT and FDG-PET in migraine aura and PMA. This could lead to a better understanding of its underlying pathophysiology (Table 4). Thinking of the one patient, described above, with a subtle occipital cortical infarction and hypoperfusion on the PET scan, one may wonder whether CT scans done in some of the patients were sufficient to detect small lesions. Moreover, it may be questioned whether MR scans presently used for routine clinical purposes (usually 1.5 Tesla with variable slice thickness) are sensitive enough to pick up small cortical lesions. Lesions found in migraineurs (mainly those with aura) in the Cerebral Abnormalities in Migraine, an Epidemiological Risk Analysis (CAMERA) study were largely found in the cerebellum, but patients with PMA were not included. Could small (without causing hemianopia) cortical infarctions in the occipital cortex be responsible for PMA by causing loss of inhibitory GABA-ergic interneurons resulting in an imbalance in local networks leading to reverberating CSDs?[39,40]

The treatment of PMA is purely based on case reports and expert opinion. As could be expected in such a rare phenomenon, there are no double-blind, placebo-controlled studies. Only a few drugs have been described to be effective: Acetylsalicylic acid,[10] baclofen,[1] divalproex sodium,[10] furosemide,[11,20,25] lamotrigine,[12,17,25] nifedipine,[15] nimodipine,[15] and sertraline.[7]

Basically, treatment is based upon the theory of hyperexcitability.[1] In our cases only lamotrigine was effective, with resolution of symptoms in one case. In literature, lamotrigine also seems the most described effective treatment (Table 5).

Wang et al tried to find a prediction model for the prognosis of PMA. The analysis involved 29 patients. They concluded that PPVD had a poorer prognosis, and higher Visual Aura Rating Scale scores, which are more typical of migraine visual aura, predicted a better outcome. For those with a potential for complete resolution, improvement would occur early in the course.[17] We have been unable to reproduce these figures mainly because of incomplete data in these 47 cases.

Recently Schankin et al described the phenomenon "visual snow" or "positive persistent visual disturbance" as a distinct disease entity from MA. They describe 120 patients with this unique clinical syndrome. The criteria they formulated include visual snow during daytime or at night plus at least one of the following: floaters, persistent after-images, "hard time seeing at night," "little cells that travel on a wiggly path," photophobia, "moving objects leave trails," flashes and "swirls with eyes closed." The etiology is currently unknown.[41,42]