Abstract
Migrainous vertigo (MV) is accepted as a common cause of episodic vertigo. The peripheral or central vestibular localization of the deficit as well as the pathophysiology is unclear. This prospective study was designed to assess the clinical features of MV and to search for the localization of the vestibular pathology. Thirty-five patients with MV, 20 patients with migraine and 20 healthy volunteers were studied. Comprehensive neurotological tests were performed between attacks. None of the normal controls or the patients with migraine had ocular motor deficits or caloric test abnormalities. Three patients in the MV group showed saccadic pursuit (8.6±), in one of whom saccadic hypometria was also present. Caloric test results revealed unilateral caloric hypofunction in seven patients (20±). Static posturography results revealed increased sway velocity when the eyes were closed or the platform was distorted in patients with MV. These findings during the symptom-free period revealed that peripheral vestibular dysfunction was more common than a central deficit.
Introduction
Over the last few decades the clinical association of migraine and vertigo has been supported by both clinical and laboratory studies assessing vestibular function (1–13). Though there is a growing amount of literature suggesting the existence of the clinical entity of ‘migrainous vertigo’ (MV) (14–16), the 2004 International Headache Society (HIS) classification (17) does not include such a category. Neuhauser et al. (9) have proposed diagnostic criteria for definite and probable MV based on the history of migrainous symptoms that are temporally related to recurrent vertigo.
MV has been reported to occur at any age (3, 4, 6). A female preponderance is well known (4–6, 9). It occurs more often in patients with migraine without aura (MoA) than in those with migraine with aura (MA) (2, 5, 6). In most patients migraine begins earlier in life than MV (6, 9). Patients report spontaneous (3–6) or positional vertigo different from benign positional vertigo (2, 5, 6, 18). Head motion intolerance suggesting a vestibular problem has been reported in 31 and 77% of cases (3, 4). The duration of the attacks can vary between patients. Vertigo with the typical duration of a migraine aura, i.e. 5–60 min, has been reported in 10–30% of patients (6, 9). Many patients experience attacks with and without headache (3, 5, 9). Hearing loss and tinnitus have been reported rarely (4, 5).
Caloric test abnormalities and ocular motor deficits indicating peripheral or central vestibular system dysfunction during the symptom-free interval have been reported (1–8, 10). Only one study has evaluated the oculographic findings during the acute attack (11), as well as the eight patients of another study examined during the vertiginous episode (6).
Patients and methods
The study was conducted in the specialized neurotology clinic in Ege University Medical School Department of Neurology, Izmir, Turkey. The study protocol was approved by the local ethics committee. Informed consent was obtained from all the participants. From January 2004 to January 2006 patients with the diagnosis of migraine and MV were assessed. The diagnosis of migraine was based on the IHS criteria (17). The diagnosis of MV was based on the criteria of Neuhauser et al. (9).
Thirty-five patients fulfilling the criteria of MV and 20 patients with migraine but no vestibular symptoms were included in the study. Twenty healthy non-migrainous volunteers (mostly accompanying the patients) without a history of a vertiginous attack or balance disorder served as the control group. In order not to influence the vestibular function testing, subjects with medical, neurological or orthopaedic conditions were excluded.
All patients had a detailed neurotological examination between migraine attacks, which included examination of stance and gait in addition to examination of eye movements and positional tests consisting of the Dix–Hallpike and roll manoeuvres. The eye movements were recorded monocularly by video-oculography during these tests (Visual Eyes 4 channel VOG; Micromedical Technologies, Chatham, IL, USA). Spontaneous eye movements in darkness with the non-recorded eye fixating and without fixation, saccadic, pursuit and optokinetic eye movements were recorded. All patients underwent pure-tone audiometry and bithermal caloric tests using air calorics. ICS air caloric stimulator model NCA-200 (ICS, Schaumburg, IL, USA) was used for caloric tests with an air flow of 8 l/min at 25°C and 50°C within 60 s. Maximum slow phase velocity (SPV) was determined using the ICS velocity computer system. Caloric testing was evaluated for side difference (a 25% difference being considered significant) and bilateral hypofunction (maximal SPV of nystagmus for cold plus warm caloric stimulus should not exceed 12°/s).
Static posturography was performed using the NeuroCom System Version 8.0.3 (NeuroCom International Inc., Clackamas, OR, USA). Mean centre of gravity (COG) sway velocity (°/s) was recorded on a static platform with eyes open and closed and on foam with eyes open and closed.
The diagnosis of peripheral vestibular dysfunction was made when there was semicircular canal paresis on caloric tests with the absence of signs of central nervous system involvement. Central vestibular dysfunction was diagnosed when there was pure vertical or torsional spontaneous nystagmus with < 50% decrease of SPV with fixation, normal caloric test results and ocular motor abnormalities such as gaze evoked nystagmus, saccadic pursuit, saccadic dysmetria and optokinetic nystagmus asymmetries.
Statistical analyses were done by Statistical Package for Social Sciences 14.0 for Windows (SPSS Inc, Chicago, IL, USA). One-way variance analysis was used. Homogeneity of variances was tested by Levene statistics. When the variances were homogeneous the differences between the mean values of the groups was tested by F statistics. If the differences were significant the binary analysis of the groups was performed by Benferroni test. When the variances were not homogeneous, Welch was used instead of F and Dunett T3 was used instead of Benferroni tests. All the tests were performed at α = 0.05 level of significance (P < 0.05).
Results
Eighteen women and two men aged 17–51 years (mean 34 years) who had no migrainous or vertiginous symptoms served as the control group. There were 20 patients with migraine (19 women, one man) not associated with vestibular symptoms aged 19–50 years (mean 34.3 years). A history of migraine was present for 1–30 years (mean 12.7 years). In 18 patients MoA and in two MA (with other neurological symptoms) was present. Headache was holocranial in 13 and hemicranial in seven patients. There were 35 patients with MV (all women) aged 14–53 years (mean 36.8 years). A history of migraine was present for 2–25 years (mean 13.1 years). In 29 patients MoA and in six MA was present. In 12 patients headache was holocranial, whereas in 13 it was hemicranial. A history of vertigo had been present for 6 months to 10 years (mean 3.75 years). The mean age of onset of vestibular symptoms was 33.5 years (range 14–51 years). In all patients except two, migrainous headache attacks had been present before the onset of vertigo. In two they had started at the same time. The mean period between the onset of migraine headache and the onset of vertiginous symptoms was 8.5 years (range 0–20 years). In all patients vertigo attacks were characterized by spinning of the environment or of the patients themselves. Head motion intolerance was a major complaint in 25 (71.4%) patients during the vertiginous episode. In one patient, in addition to recurring constant vertigo attacks, positional vertigo was reported. In 16 patients the duration of the vertigo attacks was < 1 h (1–60 min), in 15 it was 1–24 h and in four it was > 24 h. In 14 patients vertigo attacks preceded the headaches, in eight they occurred after the headache and in 13 they were reported to start simultaneously. In nine patients vertigo attacks not temporally associated with headache were also present in addition to the attacks associated with headache. None of the patients described auditory symptoms during these vertiginous episodes. The demographic features and headache characteristics of patients with migraine and MV are given in Table 1. Vertigo characteristics of patients with MV are given in Table 2. The neurological examination of all patients was normal except for the patient describing positional vertigo attacks who had typical positional nystagmus on right head hanging position. Spontaneous or gaze-evoked nystagmus was not recorded in any group by video-oculography. None of the patients with migraine had ocular motor abnormalities such as saccadic inaccuracy, saccadic pursuit or impaired optokinetic nystagmus. However, within the MV group there were three patients with saccadic pursuit (8.6%), in one of whom saccadic hypometria was also present. Video-oculographic recording of saccadic pursuit during three different velocities of target movement of a patient with MV can be seen in Fig. 1. Caloric test results were normal in all patients within the migraine group, whereas in the MV group there were seven patients (20%) with unilateral caloric hypofunction. None of these patients had any history consistent with vestibular neuritis. Video-oculographic findings of patients with migrainous vertigo are given in Table 3.
Video-oculographic recording of smooth pursuit in a patient with migrainous vertigo. Pursuit eye movements are completely replaced by saccades when the speed of the target is increased.
The demographic features and headache characteristics of patients with migraine and migrainous vertigo
Vertigo characteristics of patients with migrainous vertigo
Video-oculographic findings of patients with migrainous vertigo
Mean COG sway velocity values in normal controls, patients with migraine and with MV can be seen in Fig. 2. When the values recorded on a firm surface with eyes open were compared with each other, no statistically significant difference could be found between the three groups. The values recorded on a firm surface with eyes closed revealed a statistically significant increase in sway velocity in patients with MV when compared with the normal controls (P = 0.004). Recordings performed on foam with eyes open showed a statistically significant increase in sway velocity in patients with MV when compared with both normal controls (P = 0.001) and migraine (P = 0.02). The same was true for the recordings on foam with eyes closed. Increased mean COG sway velocity was recorded in the MV group when compared with normal controls (P = 0.001) and patients with migraine (P = 0.003). The values recorded in the migraine group were not different from the normal controls on foam with eyes either open or closed. The highest values of mean COG sway velocity on foam were recorded in patients with unilateral peripheral vestibular loss.
Mean centre of gravity sway velocity in posturography is increased in patients with migrainous vertigo on foam with eyes both open and closed when compared with normal controls and patients with migraine.
Pure tone audiometry was normal in 70 patients. Four patients (two in the MV, one in the migraine group and one in normal controls) demonstrated bilateral high-frequency sensorineural hearing loss consistent with presbyacusis. There was one patient with unilateral conductive hearing loss in the MV group.
Discussion
In our study 20 migraine patients and 35 MV patients were assessed prospectively. The demographic and clinical features of the patients with MV given in Tables 1 and 2 were concordant with previous reports (2–6, 8, 9).
Neurootological tests during the symptom-free period have documented peripheral or central vestibular dysfunction in patients with MV (2–4, 6, 8). Studies on patients with migraine unselected for vestibular symptoms have shown similar results (1, 10). Dieterich and Brandt (6) have reported a very high incidence (65%) of pathological ocular motor findings indicating permanent brainstem or cerebellar dysfunction, whereas peripheral vestibular deficits have been reported as 8.3%. Cutrer and Baloh (3), on the other hand, have reported the frequency of persisting peripheral vestibular deficits as 21%, the highest of all the studies. We had three patients (8.6%) within the MV group who persistently showed saccadic pursuit, and in one of those three saccadic hypometria was also present on behalf of a central origin. On caloric tests, unilateral vestibular hypofunction was found in seven patients (20%), indicating a peripheral defect. None of our patients with migraine not associated with vestibular symptoms had any finding indicating a peripheral or central dysfunction. In the study of Toglia (1), 80% of the migraine patients had a peripheral labyrinthine dysfunction. In Harno's series (10), 17% had unilateral hypoexcitability to caloric stimulation, but almost one-third had inaccuracies in their saccadic eye movements indicating vestibulocerebellar dysfunction. There have been few case reports on clinical findings during an acute episode of MV (6, 19, 20) indicating central vestibular dysfunction. Von Brevern et al. (11) studied 20 patients during the acute attack and have documented findings consistent with central vestibular dysfunction in 50% and peripheral vestibular dysfunction in 15%. None of our patients could be evaluated during the acute attack.
There are few posturographic studies on patients with migraine or with MV (4, 10, 21, 22). They have demonstrated an excessive reliance on somatosensation for upright balance (4, 22). The findings in our group were the same. Mean COG sway velocity recorded on a firm surface with eyes open was not different from values recorded in normal controls or patients with migraine. However, recordings performed on foam with eyes both open and closed revealed increased sway velocities in patients with MV when compared with normal controls and with migraineurs. This was concordant with the caloric test results showing peripheral vestibular deficits in 20% of the MV patients. The results of the patients with migraine in these two conditions were not different from those of the normal controls. Harno et al. (10) have described abnormal static posturography results in patients with migraine especially tested with eyes open, which were attributed to the reliance of the patients on other sensory input than visual. However, this was not the case in our migraine group.
Evidence of peripheral abnormalities in posturography is thought to occur when the vision and platform distorted equilibrium scores are more abnormal than when not distorted. Abnormalities in posturography alone would not be enough to influence the interpreter in either direction, but may increase the confidence in diagnosis when used in conjunction with oculography.
Several mechanisms have been proposed for the pathophysiology of MV. A spreading depression affecting brainstem structures has been suggested by Dieterich and Brandt (6), who have found mainly central vestibular disorders in their patient group. Baloh (23), on the other hand, has proposed that sudden episodes of vertigo associated with migraine could be explained on the basis of vasospasm of the internal auditory artery in support of the peripheral vestibular dysfunction hypothesis. However, as both peripheral and central vestibular deficits have been observed in patients with MV, mechanisms involving both structures seem more plausible. One might be the role of neuropeptides, especially calcitonin gene-related peptide (CGRP) released during migraine attacks. They have a neuromodulatory role both in the peripheral and central vestibular system (3). Ion channel defects may be the other mechanism involving both peripheral and central vestibular structures. It appears to be the most promising hypothesis because of the clinical overlap between MV and the paroxysmal disorders familial hemiplegic migraine and episodic ataxia type 2, which result from mutations in the calcium channel gene CACNA1A (24). However, the search for mutations in the CACNA1A1 gene was negative in several patients with familial migraine with vertigo (25). Another study on candidate genes for MV has revealed the same results (26).
In our study MV seemed to be ear-related in at least 20% of cases. There seemed to be permanent damage of the vestibular end organ revealed by caloric hypofunction during the symptom-free period. This supports the internal auditory artery vasospasm theory more than transient hair cell depolarization due to a defective ion channel or CGRP release.
Acknowledgements
The authors thank to Assistant Professor Timur Köse from the Department of Medical Statistics for the statistical analysis of data.