Abstract
Background
According to ICHD-II, and as proposed for ICHD-III, non-hemiplegic migraine aura (NHMA) symptoms last between five and 60 minutes whereas hemiplegic migraine aura can be longer. In ICHD-III it is proposed to label aura longer than an hour and less than a week as probable migraine with aura. We tested whether this was appropriate based on the available literature.
Methods
We performed a systematic literature search identifying articles pertaining to a typical or prolonged duration of NHMA. We also performed a comprehensive literature search in order to identify all population-based studies or case series in which clinical features of NHMA, including but not restricted to aura duration, were reported, in order to gain a complete coverage of the available scientific data on aura duration.
Results
We did not find any article exclusively focusing on the prevalence of a prolonged aura or more generally on typical NHMA duration. We found 10 articles that investigated NHMA features, including the aura duration. Five articles recorded the proportion of patients in whom whole NHMA lasted for more than one hour, which was the case in 12%–37% of patients. Six articles reported some information on the duration of single NHMA symptoms: visual aura disturbances lasting for more than one hour occurred in 6%–10% of patients, sensory aura in 14%–27% of patients and aphasic aura in 17%–60% of patients.
Conclusions
The data indicate the duration of NHMA may be longer than one hour in a significant proportion of migraineurs. This seems to be especially true for non-visual aura symptoms. The term probable seems inappropriate in ICHD-III so we propose reinstating the category of prolonged aura for patients with symptoms longer than an hour and less than one week.
Introduction
Migraine with aura (MA) is characterized by recurrent attacks of visual, sensory, motor, aphasic or “basilar-type” (brainstem) symptoms. The presence of headache and its relation to the aura is variable. A detailed description of aura symptoms is crucial when diagnosing MA as there are no biological markers. The differential diagnosis includes cerebrovascular disorders, epilepsy and other life-threatening neurological conditions, and it is therefore important to understand the key characteristics of MA. The total duration of MA is one of those important features.
In the second edition of the International Classification of Headache Disorders (ICHD-II) (1), individual symptoms of non-hemiplegic MA (NHMA) are considered to be of typical duration when they last between five and 60 minutes, whereas hemiplegic migraine aura can be longer. In our clinical experience, NHMA symptoms often last longer than one hour. While NHMA lasting for more than one hour was coded as 1.2.2 “migraine with prolonged aura” according to ICHD-I (2), this term and code was dropped in ICHD-II. Long-lasting aura can still be coded as persistent aura without infarction – code 1.5.3, but it then needs to last for more than one week. Otherwise when lasting between 60 minutes and seven days, MA can be coded as “probable migraine with aura” – code 1-6-2 (1), and this terminology is proposed for ICHD-III. The use of the term probable implies an uncertainty that the symptom can be MA. In view of our clinical experience, we find this an unhelpful characterization of longer auras.
In this systematic review, we searched and analyzed all articles including data about the duration of NHMA in order to provide a scientific basis for the classification of MA in terms of its duration. These findings have been reported in preliminary form (15th International Headache Congress, Berlin, June 2011 (3)).
Methods
We performed a systematic literature search identifying articles reporting the typical or prolonged duration of NHMA. We also performed a comprehensive literature search in order to identify all population-based studies or case series in which clinical features of NHMA, including but not restricted to aura duration, were reported, in order to gain a complete coverage of the available scientific data about aura duration. Our primary aim was to provide an overview of the scientific data related to the distribution of NHMA duration and in particular the prevalence of overall NHMA as well as individual symptoms lasting for more than one hour. Conference abstracts were considered if they included detailed information on the pertinent outcomes.
Literature search
The last search was performed in January 2011, with both PubMed and EMBASE using the key words “migraine aura” and “migraine with aura” combined with the words “duration,” “prolonged,” “characteristics,” “features,” “atypical” and “typical.” Articles in English, German and Italian were considered. We also considered articles from the reference list of studies that were found to be relevant as well as literature that was known to be relevant by the authors. Moreover, we considered the bibliography of ICHD-II (1).
Data extraction
Three investigators examined the abstracts found in the literature search. Whenever the title or abstract suggested that relevant data could be part of the publication, the entire manuscript was examined. Relevant data from accepted articles were abstracted for the following data categories: publication information (authors, years), population (number of patients, gender distribution, age, age of aura onset, recruitment strategy), aura features (prevalence and duration of each aura's symptoms, duration of the whole aura, presence of headache during aura, frequency of aura attacks), study methodology and the applied MA definition (diagnostic criteria). Three reviewers were involved in the review of the abstracts (MV, MA, TS). Each abstract was reviewed by two reviewers. In case of disagreement, a consensus agreement was reached by involving the third person. This was the case in the review of seven abstracts.
Case definitions
We did not consider articles exclusively relating to familial or sporadic hemiplegic migraine, basilar-type migraine, persistent auras or symptomatic (secondary) MA. We also did not consider reports of single cases as we wanted to assess the distribution of the MA duration in a population of patients affected by NHMA and not obtain anecdotal data. Case series were considered if they included 10 or more patients.
Results
The search strategy identified 1751 published studies (see Figure 1). We did not find any article exclusively focusing on the prevalence of prolonged aura in NHMA sufferers, although we found 25 articles with relevant data.
Flowchart of the review process.
Available material
We had to exclude two clinical trials in which acute (sumatriptan) or preventive (topiramate) medications were assessed as a treatment for MA (4,5). In those articles, only patients with typical aura, <60 minutes duration of individual symptoms, were recruited. In another study, lamotrigine was used to treat 47 selected patients with severe and disturbing aura (6). Among them there were eight patients with prolonged aura. We considered there could be selection bias. For similar reasons, we excluded a study in which the authors reviewed the characteristics of 30 patients with a diagnosis of migraine with prolonged aura (7) and another study that included only patients with atypical aura (8). We excluded a report of personal observations on visual aura (VA) as the data were not systematically presented (9). We had to exclude nine studies in which aura duration was recorded and in which, although the study did not focus on hemiplegic migraine, a proportion of included patients (1.1% to 18%) (10–18) had motor aura symptoms and those patients could not be separated from the NHMA patients concerning the duration of either each individual aura symptom or the aura duration as a whole. On the other hand, from one study of 47 patients (19), involving 40 patients with typical NHMA, one patient with familiar hemiplegic migraine and six patients with basilar-type aura were included as it was possible to separate these groups of patients with respect to the duration of the aura (“in five cases of migraine with typical aura, the aura sometimes, lasted longer than 60 minutes”).
NHMA
Relevant data from accepted articles. M: male; F: female; Pt: patient; HA: hemiplegic aura; MA: migraine aura; NHMA: non-hemiplegic migraine aura. TNA: transient neurological attack; TIA: transient ischemic attack; TVS: transient visual symptoms; VA: visual aura; SA: somato-sensory aura; MTA: motor aura; AA: aphasic aura; NR: not reported; NOS: not otherwise specified; R: retrospective; P: prospective, MD: medical doctor; dx: diagnosis; FND: focal neurological deficits; PFO: patent foramen ovale.
Aura definition
The applied classification criteria for the diagnosis of MA (or classic migraine) varied depending on the publication date of the original articles. In one article the ICHD-II criteria were used (21), in five articles (19,20,22–24) the ICHD-I criteria (1) were used, and in one article (25) the criteria of the Ad Hoc Committee on Classification of Headache of the National Institute of Health from 1962 (26) were applied, whereas in three manuscripts no data regarding classification criteria were available or clearly stated (27–29).
Age and aura frequency
The number of patients studied in the individual articles ranged between 11 and 362. The cumulative number of all patients in the studies that were considered was 1178. The age of onset of MA in the sample of patients was reported only in two studies and ranged from a mean of 11.9 years (± 3.1, range 4–17) (19) to a mean of 21 years old (± 12, range 5–77) (20). The frequency of MA attacks was clearly reported in two studies: In the first, 54.9% of patients suffered from less than one attack per month and 9.7% from more than three attacks per month (25), in the second, the mean of MA episodes/year/patient was reported to be 29 (ranging from less than one to 156) (22). The occurrence of headache in relation to NHMA was reported in three out of 10 studies (20,24,28).
Duration
The last two columns of the Table 1 summarize the data relating to the duration of NHMA. In five articles, there was some information on the overall duration of NHMA (19–21,23,25). Those studies reported the prevalence of MA lasting for more than one hour: The minimum percentage was 11.6% (25), whereas the maximum was 36.9% (21) of patients (see the Table 1). In six articles, the duration of single NHMA symptoms was separately recorded (20,22,24,27–29). In two articles VA, somatosensory (SA) and AA were recorded (20,29), in one article VA and SA duration were reported (24) whereas in three articles the authors reported just the duration of VA (22,27,28). Most of these studies reported the prevalence of individual NHMA symptoms lasting for more than one hour: Six percent to 10% of patients experienced episodes of VA lasting for more than one hour. The lowest prevalence of patients with VA symptoms lasting for more than one hour (3%) was reported in a study (28) in which patients described more than one episode of VA. Three percent of patients constantly experienced VA longer than one hour whereas 10% of patients had some MA longer than one hour. With respect to non-visual symptoms: SA lasted for more than one hour in 14% to 27% of patients and aphasic aura (AA) in 17% to 60% of patients (see Table 1). In another article (27), the prevalence of VA lasting between 30 minutes and 120 minutes (as upper limit) was reported to be 25% of patients. On the other hand, a prospective study in which 56 MA attacks were studied with a questionnaire in 20 patients (24) reported the median duration and range of VA with progressive onset (median 10 minutes, range 10 to 40 minutes), VA with acute onset (median 45 minutes, range 15 to 65) and SA (median 55 minutes, range three minutes to several days).
Discussion
This review identified 10 articles in which aura features, including duration, were studied in patients with NHMA. Five articles demonstrated an overall duration of NHMA lasting for more than one hour in 11.6%–36.9% of patients. Other articles reported the prevalence of at least one individual NHMA symptom lasting for more than one hour: VA was experienced by 6%–10% of patients, SA in 14%–27% of patients and AA in 17%–60% of patients. These data suggest ICHD-III should reinstate migraine with prolonged aura as a diagnostic category for patients with aura lasting longer than one hour and less than seven days. The use of the term probable MA is unhelpful and does not reflect the reality of the existence of the prolonged aura phenotype.
Two studies reported in a different way the distribution of aura symptom duration in their population samples. In the first (27) the prevalence of VA lasting between 30 minutes and 120 minutes was 25% of patients. Unfortunately, no data are available on the subgroup of patients with symptoms lasting between 60 and 120 minutes in this study. Importantly, these latter values are not conflicting with the data we found in other studies. Another study (24) reported the aura duration in terms of median and range of duration for visual and sensory symptoms. This study reported VA with progressive onset: median 10 minutes, range 10 to 40; VA with acute onset: median 45 minutes, range 15 to 65; SA median 55 minutes, range three minutes to several days. From this study, we could not deduce the prevalence of aura symptoms lasting for more than one hour, but it is clear that the duration of the aura symptoms, especially SA, must have easily exceeded the one-hour limit in a proportion of patients.
Some authors studied one aura attack; others reported the “most common aura” of their patients, while others provided diaries or questionnaires for patients to complete when experiencing aura episodes. Hence, the available data are not homogeneous. In three studies (19,23,28) the patients described more than one aura, and it is interesting to note that some of the patients with aura attacks lasting for more than one hour, either the whole MA or the single VA, can also experience attacks with an entirely typical duration of aura according to ICHD-II. This indicates that even in individual patients there is a spectrum in terms of aura duration. It could be argued from these data that the one-hour limit is artificial.
Retrospective observations suffer from recall bias; however, in light of very few prospective studies, retrospective articles offer some insights. Of the four prospective studies found in the literature, two studies (23,24) stated they provided questionnaires, including a question about MA duration, that the patients had to complete during at least one attack. This is probably the most reliable and valid way to obtain data about the duration of MA symptoms. A limitation of these two studies, as well as of the others except one (20), is that the sample of patients is clinic based, presenting potential inclusion bias. Interestingly, the only population-based study with such data (20) provided results in line with clinic-based studies. It seems clear from all these data that the ICHD-II description of MA as characterized by “focal neurological symptoms that usually develop gradually over 5–20 minutes and last for less than 60 minutes” excludes an important group of patients.
The literature lacks studies designed to investigate specifically the duration of NHMA. The studies we found have different limitations: i) Most of the studies are retrospective and some of the prospective studies did not use a diary to be completed during the attacks, putting them at risk for recall bias; ii) in some studies diagnostic criteria for MA were old or not clearly stated, and iii) many studies investigated clinic-based patient populations presenting potential inclusion bias.
We intend this systematic review to stimulate clinical, epidemiological and also pathophysiological research into the clinically relevant issue of prolonged aura. Such studies will allow a better definition of the atypical aura characteristics that should really worry the physician regarding a potential non-benign etiology.
Conclusion
The best data currently available suggest in 12% to 37% of the patients suffering from NHMA the whole aura can last for more than one hour. The data suggest it is essential when finalizing ICHD-III to provide an appropriate diagnostic category for these patients. The label “probable” invites repeated investigations that waste resources, and concern that seems unwarranted from both the data and clinical experience. The data strongly suggest the category of migraine with prolonged aura be re-established in ICHD-III.
Footnotes
Clinical implications
Migraine aura can be more than one hour in more than 10% of cases
The term prolonged aura is clinically useful
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest
PJG is on Advisory Boards for Allergan, Colucid, MAP pharmaceuticals, Merck, Sharpe and Dohme, eNeura, Neuraxon, Autonomic Technologies Inc, Boston Scientific, Electrocore, Eli-Lilly, Medtronic, Linde gases, Arteaus, AlderBio and BristolMyerSquibb. He has consulted for Pfizer, Nevrocorp, Lundbeck, Zogenix, Impax and DrReddy, and has been compensated for expert legal testimony. He has grant support from Allergan, Amgen, MAP, and MSD. He has received honoraria for editorial work from Journal Watch Neurology and for developing educational materials and teaching for the American Headache Society. None of this work pertains to the work presented.