Jtournal of Neurology, Neurosurgery, and Psychiatry 1992;55:437-440 437

Migraine without : transient J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.6.437 on 1 June 1992. Downloaded from ischaemic attack or not?

Martin Dennis, Charles Warlow

Abstract to cerebral or ocular ischaemia-that is, tran- Patients with " aura without sient ischaemic attacks (TIAs). headache" can be separated from those Several authors have described patients in with transient ischaemic attacks (TIA) on whom the onset and nature of the symptoms the basis of the onset of symptoms, which suggests that they have a variant of migraine is generally gradual and which spread or but in whom there is no headache. Perhaps, the intensify over minutes or hours, and from most notable reports are those of Fisher4' 7 the typical migrainous nature of their who described "late onset migraine accom- visual symptoms. Fifty cases were pro- paniments" and Whitty8 who described spectively identified and these were com- "migraine without headache". Fisher has pro- pared with 50 age matched patients with a vided many detailed case histories that have TIA. Surprisingly, there were no sig- helped to delineate the clinical syndrome. He nificant differences in the prevalence of also stated that "the condition can be justifi- vascular risk factors and diseases in the ably regarded as benign"5 but without provid- two groups although those with TIAs ten- ing any detailed follow up on the patients. He ded to have more. Only one patient with appears to have relied upon the fact that many "migraine aura without headache" suf- of the patients had had attacks for many years fered a subsequent vascular event (a myo- when he first saw them and the few he followed cardial infarction). Three patients with up suffered no serious sequelae. Although his TIAs had and two others died from conclusion may be correct, the evidence is vascular disease. Although the results are unreliable since patients who are dead or not conclusive, they do suggest that severely disabled are unlikely to present a patients with the clinical characteristics of neurologist with a history of relatively minor "migraine aura without headache" have a events, nor will they be able to attend follow up low risk of subsequent vascular events, clinics. We therefore present the results of a lower than those with TIAs, despite per- follow up study of a series of patients, clinically haps having similar prevalence of vas- similar to those described by Fisher, with what cular risk factors. Although their risk we have called "migraine aura without head- factors should be treated, patients with ache", to describe more precisely the natural "migraine aura without headache" should history of this condition. We have compared http://jnnp.bmj.com/ be reassured and not subjected to inap- this series of patients with an age and sex propriate and potentially hazardous matched group of definite TIA patients to investigations and treatment. determine whether the prevalence of vascular diseases, risk factors, and prognosis is different in the two groups. This is what one would Focal cerebral and visual symptoms commonly expect if Fisher is correct, that "migraine aura

precede or accompany headache in classical without headache" is a benign condition from on September 23, 2021 by guest. Protected copyright. migraine. Patients who have classical migraine the point of view of not just the lack of with neurological symptoms often have similar headache, but prognosis for serious vascular neurological symptoms at other times but events. without headache. In such cases, few would disagree that the isolated neurological symp- toms are "migrainous" and they have been Patients and Methods called "migraine sine hemicrania,"' "migraine Between 1977-86 patients who were referred accompagnee,"' migraine associee", to a neurology clinic with transient neuro- Department of "metastases of migraine,"' "migraine accom- logical symptoms were prospectively classified Clinical paniments",4 5 or "acephalgic migraine".6 as having definite TIAs or variants of migraine, Neurosciences, Western General A more difficult clinical problem arises when among other diagnoses, by one of us (CW). Hospital, Edinburgh patients present with transient focal neuro- Detailed histories including the presence or M Dennis logical or visual symptoms and yet have never absence of vascular risk factors were recorded WWarlow experienced these symptoms in the context of a for all patients. Four hundred and sixty nine Correspondence to: classical Dr Dennis, Department of migraine attack, and who have no patients were diagnosed as having definite Clinical Neurosciences, associated headache. Such patients may pres- TIAs9 using a standard definition (see below). Western General Hospital, ent in Crewe Rd, Edinburgh EH4 relatively late life and have symptoms of Another 75 patients were classified as having a 2XU, UK vascular disease (for example, angina) and risk variant of migraine although at that time we Received 14 March 1991 factors (for example, hypertension) increasing had no set diagnostic criteria for these cases. At and in final revised form both 8 August 1991. the doctors' and patients' anxiety that the the end of this period the records of the 75 Accepted 14 August 1991 symptoms, neurological or visual, may be due patients were reviewed without knowledge of 438 Dennis, Warlow

their outcome and diagnostic criteria were hospital notes were reviewed. The TIA patients J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.6.437 on 1 June 1992. Downloaded from applied so that a homogeneous group of were prospectively followed up until the end of patients with "migraine aura without head- 1986 at a specialist clinic, by their general ache" could be described. practitioner, or by letter. If patients either died Our diagnostic criteria for "migraine aura or described symptoms suggesting a serious without headache" were that: either firstly, the vascular event, their hospital notes were focal neurological symptoms should come on reviewed to confirm the diagnosis. The prog- gradually and spread or intensify over a period nosis in the two groups was compared using a of minutes, not seconds, or secondly, that the logrank analysis of Kaplan Meier survival patient describes positive visual symptoms curves. No patient in either group was lost to characteristic of classical migraine even if they follow up. come on abruptly-that is, fortification spec- tra, flashing lights, dazzles; and thirdly, that headache should be either totally absent or if present at all, should be described by the Results patient as mild, and that the patient does not The general characteristics of the patients with take analgesics or change their activity due to "migraine aura without headache" and TIAs the headache; and also fourthly, the patient are summarised in table 1. Forty nine of the should not have had similar symptoms pre- fifty patients (98%) with "migraine aura with- viously in the context of either classical out headache" described visual symptoms migraine or in association with a more severe which were binocular in 35 patients (71%), headache. monocular in five (10%) and uncertain in nine By applying these diagnostic criteria to the (18%). Most patients (35/49; 71%) described 75 patients with migraine variants, 25 were positive visual features such as flashing lights, excluded, leaving a group of 50 with "migraine fortification spectra or haloes. Fifteen patients aura without headache". The main reason for (30%) had other symptoms; most commonly exclusion was that the patients described mod- sensory (13/50; 26%), aphasia (8/50; 16%), or erate or severe headache in association with the dysarthria (3/50; 6%), and clumsiness or presenting or previous episodes. weakness of a limb or limbs (5/50; 10%). Some The 50 patients were stratified by age (into patients described dysgraphia (2/50; 4%) in five year age bands) and by sex. We then association with dysphasia, associated vertigo, randomly selected age and sex matched and feelings of remoteness. One patient lost patients from the TIA group which had been consciousness during an attack. Patients often similarly stratified. Unfortunately, there were gave histories going back over several months insufficient young female TIA patients in our or years and described multiple attacks during cohort group to provide age and sex matched that period. Only 11 (22%) patients with controls for each ofthe patients with "migraine "migraine aura without headache" and 15 aura without headache". In four cases, there- (30%) with TIAs were referred with their first fore, we randomly selected an age matched attack. The mean number ofattacks and length male patient since age is a more important of history was longer in the migraine group determinant of the risk of vascular events than than in those with TIAs. Most

commonly http://jnnp.bmj.com/ sex. patients (23/50; 46%) described their symp- For the purposes of this study, TIA is toms as coming on or intensifyfing over a period defined as "an acute loss of focal cerebral or of between six and 30 minutes. A few patients monocular function with symptoms lasting less with visual symptoms (10/49; 20%) described than 24 hours and which after adequate a more sudden onset (less than 1 minute), 14 investigation is presumed to be due to embolic (28%) had symptoms developing over a period or thrombotic vascular disease".10The patients of one to five minutes and a few patients (3/50; with "migraine aura without headache" were 6%) described symptoms coming on over more on September 23, 2021 by guest. Protected copyright. followed up by telephone in March and April than 30 minutes. Episodes lasted less than 15 1987 and asked questions about symptoms minutes in five patients (10%), 15 to 60 which suggested a diagnosis of or myo- minutes in 37 (74%) and more than one hour cardial infarction. Where patients described in 8 (16%). Ten patients (20%) experienced symptoms suggesting a serious vascular slight headache, one patient vomited and two problem, their general practitioner and hospi- others felt slight nausea in association with tal notes were reviewed to confirm the diag- their visual or neurological symptoms. Two nosis. In the majority of other patients the patients with "migraine aura without head-

Table 1 General characteristics andprevalence ofvascular diseases and riskfactors in patients with "migraine aura without headache", TIA controls, and all the TIA patients referred during the study period Migraine aura without TIA headache controls AU TIAs' n =50 n =50 n =469 Mean age (SD) 48-7 (13-3) 49-0 (13-7) 62-1 (12) No of Men (%) 30 (60) 34 (68) 317 (68) Hypertension (%) 10 (20) 13 (26) 198 (42) Ischaemic heart disease (%) 4 (8) 4 (8) 99 (21) Peripheral Vascular disease (%) 3 (6) 5 (10) 76 (16) Diabetes (%) 1 (2) 2 (4) 25 (5) Current smokers (%) 15 (30) 20 (40) 221 (47) Migraine aura without headache: transient ischaemic attack or not? 439

on a Table 2 A comparison ofprognosis in patients with "migraine aura without headache" and TIAs with p value based J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.6.437 on 1 June 1992. Downloaded from logrank analysis Observed Expected Observedl Group n strokes strokes Expected TIA 50 3 1-5 2-0 lp = 0 04 Migraine aura without 50 0 1-5 0.0 headache Observed vasc deaths/ Expected vasc deathsl Observed! Group n strokes/MI strokes/MI Expected TIA 50 5 3 0 1-7 lp = 0-05 Migraine aura without 50 1 3-0 0 3 headache MI = myocardial infarction.

ache" gave a history of classical migraine years Discussion before but their aura had been quite different We have described a group of patients who from the symptoms described during the pre- presented with transient focal neurological or senting episode. Ten other patients gave histo- visual symptoms and a benign prognosis. ries of past headache without other features of Because of the lack of severe headache the migraine. diagnosis of TIA was considered but because The prevalence of vascular diseases and risk of the gradual onset or nature of the symptoms factors in the patients with "migraine aura described, these patients were classified as without headache" and control TIAs was not having "migraine aura without headache". significantly different although the TIA Since there are no investigations to confirm or patients did have a higher prevalence of all refute the diagnosis of "migraine" or of TIA, vascular disease and risk factors other than one has to rely on indirect evidence to decide ischaemic heart disease (see table 1). The whether it is reasonable to classify patients in control TIAs were younger than the whole this way. For example, if patients with group of TIA patients referred to the clinic "migraine aura without headache" had a much over the study period and almost certainly as a lower prevalence of vascular diseases and risk result had a lower prevalence of vascular factors than patients with TIAs and also had a diseases and risk factors. Seventeen of the 50 lower risk of developing subsequent vascular patients with TIAs had amaurosis fugax, eight complications, it would be clinically useful to of these were less than 50 years old, a group separate these groups of patients and one known to have few risk factors and a good might deduce that the underlying pathogenesis prognosis. Forty seven (94%) of the TIA was different. It would, of course, be more patients had an onset of symptoms over only a difficult to say that these patients definitely had few seconds. In three patients (6%) symptoms migraine, one could only say that whatever the came on between 1 and five minutes. These mechanism of the symptoms, the clinical three were obviously less classical in their syndrome has a good prognosis and sounds

presentation but had been diagnosed as defi- like migraine. http://jnnp.bmj.com/ niteTIAs when they first presented. For the In this study the patients with "migraine purposes of the analysis in this paper they were aura without headache" had a similar preva- not subsequently reclassified. This might have lence of vascular diseases and risk factors as resulted in smaller differences between theTIA those with TIAs. Obviously with only small and "migraine aura without headache" numbers of patients in each group, only large groups. differences could reliably be demonstrated and During a mean follow up period of 4 5 years type two errors are likely. One explanation of (1H1 to 9-4 years), one patient (2%) with this similarity, if it is indeed true, might be that on September 23, 2021 by guest. Protected copyright. "migraine aura without headache" had a non- the symptoms of "migraine aura without head- fatal myocardial infarction, none died and no ache" may be related to the presence of patient suffered a stroke (table 2). One patient vascular pathology as in TIAs. It has been required coronary artery surgery for angina suggested that migraine attacks can be trig- pectoris and another had arterial surgery for gered by embolic events" but our study cannot intermittent claudication. During a mean fol- further this debate. An alternative, and in our low up period of 4-9 years (01 to 9 7 years) six view, a more likely explanation might be that TIA patients died (12%), three from malig- general practitioners and others seeing patients nancy, two from definite cardiac disease and with transient neurological symptoms with one suddenly, presumably from a vascular some of the features of migraine might refer event. Three (6%) TIA patients had a stroke, patients with vascular disease and risk factors one ofthese died subsequently ofheart disease, to a specialist neurology clinic leading to a but none suffered a non-fatal myocardial referral bias. The control TIA patients in this infarction. The prognosis of the two groups in study had less associated vascular disease and respect of their risk of stroke, myocardial fewer risk factors than unselected TIA patients infarction, and vascular death is shown in table (see table 1) almost certainly because they 2. During a mean follow up period of4-1 years, were relatively young. More encouraging was 58 (12%) of all 469 TIA patients died from a the result of our follow up study which despite vascular cause, 63 (13%) had a stroke and 16 small numbers suggested the patients with (3%) had a non-fatal myocardial infarction.9 "migraine aura without headache" had a very 440 Dennis, Warlow

low risk of serious vascular events. This risk factors should- be treated on merit but poten- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.6.437 on 1 June 1992. Downloaded from appeared to be lower than that in the control tially hazardous investigations and treatments TIA patients, although the differences in prog- should be avoided. Generally, patients with nosis did not reach conventional levels of "migraine aura without headache" should be statistical significance; partly because of small reassured that they have a type of migraine, numbers but also because we selected younger fortunately without headache, which has a very TIA patients to be controls and by chance good prognosis, and that they have not had a included a high proportion (34%) of patients mild stroke. with amaurosis fugax who have a better prog- nosis than patients with transient cerebral ischaemic attacks.9This meant that the control 1 Bruyn GW. Migraine equivalents. In: Clifford Rose F, ed. TIA patients had a much lower risk of stroke Handbook of cliical neurology Vol 4 (48): Headache. and other vascular events compared with the Amsterdam: Elsevier, 1986:155-71. 2 Priorry P. Memoire sur une des affections designees sous le whole group of TIAs and it became more nom migraine ou hemicranie. J Univ Hebd Med Chir Prat difficult therefore to show a statistically 1931;2:5-18. sig- 3 Tissot M. Le traite de la catalepsie, de l'extase de nificant difference between them and the l'anesthesie, de la migraine, et des maladies du cerveau. "migraine aura without headache" group. The Lausanne grasset CY 1788:90-1374. 4 Fisher CM. Late life migraine accompaniments as a cause of better prognosis observed in the patients with unexplained transient ischemic attacks. Can J Neurol Sci "migraine aura without 1980;7:9-17. headache" could in 5 Fisher CM. Late life migraine accompaniments-further part be explained by a selection bias; this group experience. Stroke 1986;17:1033-42. may have contained a greater proportion of 6 O'Connor PS, Tredici TJ. Acephalgic migraine. Fifteen years e ence. Oph oy 1981;88:999-1003. "survivors" than the TIA group because they 7 Fisher CM. Cerebral ischemia, less familiar types. Clin presented having had more Neurosurg 1971;18:267-336. attacks and with a 8 Whitty CWM. Migraine without headache. Lancer greater interval since their first ever attack. 1967;ii:283-5. We have shown that patients with "migraine 9 Hankey G, Slattery J, Warlow C. The prognosis of hospital referred transient ischaemic attacks. J Neurol Neurosurg aura without headache", according to our Psychiatry 1991;54:793-802. criteria, even in the presence of 10 Warlow CP, Morris PJ. Introduction. In:Warlow CP, Morris vascular PJ, eds. Transient ischaemic attacks. New York: Dekker, disease or risk factors, have a good prognosis 1982:vii-xi. and that it is therefore worth 11 Peatfield RC. Can transient ischaemic attacks and classical distinguishing migraine always be distinguished? Headache 1987; them from TIAs. Obviously their vascular risk 27:240-3. http://jnnp.bmj.com/ on September 23, 2021 by guest. Protected copyright.