Clinical manifestations

Presentation and course

The diagnosis of vestibular migraine requires (Table 1):

• recurrent vestibular symptoms
• a history of migraine
• temporal association of vertigo with migraine symptoms
• and exclusion of other causes

Vestibular migraine has been included in the third edition of the International Classification of Headache Disorders (ICHD) in an appendix for emerging syndromes (27). Only the International Classification of Vestibular Disorders of the Bárány Society currently recognizes probable vestibular migraine (42).

Table 1. Diagnostic Criteria for Vestibular Migraine

Vestibular migraine presents with various types of vertigo as defined by the Bárány Classification of Vestibular Symptoms, including spontaneous vertigo, positional vertigo, visually induced vertigo, head motion–induced vertigo, and head motion–induced dizziness with nausea (10; 74). Some patients experience a sequence of spontaneous vertigo transforming into positional or head motion–induced vertigo after several hours or days. Positional vertigo usually persists as long as the critical head position is maintained (71; 84). Altogether, 40% to 70% of patients experience positional vertigo in the course of the disease, but not necessarily with every attack (33; 73). In a large sample of patients presenting with positional vertigo, vestibular migraine was diagnosed almost as frequently as benign paroxysmal positional vertigo (18). Head motion–induced vertigo presents with imbalance, illusory motion, and nausea aggravated or provoked by head movements (09). A persistent susceptibility to motion sickness is common in patients with vestibular migraine (51; 09; 17). Visually induced vertigo, ie, vertigo provoked by moving visual scenes such as traffic or movies, can be another prominent feature of vestibular migraine and may similarly linger on in between attacks (75; 09). As persistent vestibular symptoms affect a substantial subgroup of patients with vestibular migraine, a separate diagnostic category of chronic vestibular migraine has been proposed (17).

Nausea and imbalance are frequent but nonspecific accompaniments of acute vestibular migraine. Attacks of vestibular migraine may be severe enough to force patients to stay in bed for a day or two, where they lie still avoiding the slightest head movement.

Many patients experience attacks both with and without headache (69). Frequently, patients have an attenuated headache with their vertigo as compared to their usual migraine. In some patients, vertigo and headache never occur together (55). Along with the vertigo, patients may experience photophobia, phonophobia, osmophobia, and visual or other auras. These phenomena are of diagnostic importance because they may represent the only apparent connection of vertigo and migraine. Patients need to be asked specifically about these migraine symptoms because they often do not volunteer them. A dizziness diary can be useful for prospective recording of associated features. Auditory symptoms, including hearing loss, tinnitus, and aural pressure related to acute attacks have been reported in 20% to 40% of patients with vestibular migraine (53; 68; 69). Hearing loss is usually mild and transient, without or with only minor progression in the course of the disease. About 20% develop mild bilateral downsloping sensorineural hearing loss over the years (64).

Duration of episodes is highly variable: about 30% of patients have episodes lasting minutes, 30% have attacks for hours, and another 25% have attacks over several days. The remaining 15% have attacks lasting seconds only, which tend to occur repeatedly during head motion, visual stimulation, or after changes of head position. In these patients, episode duration is defined as the total period during which short attacks recur. At the other end of the spectrum, there are patients who may take 4 weeks to fully recover from an episode. However, the core episode rarely exceeds 72 hours (24; 55; 69).

Vertigo may (rarely) precede headache as would be typical for an aura, may begin with headache, may appear late in the headache phase, or may be unrelated to headaches (30). Thus, vestibular migraine often misses not only the duration criterion for an aura as defined by the ICHD, but also the temporal relationship to migraine headaches (37).

Asking for migraine-specific precipitants of vertigo attacks (eg, provocation by menstruation; deficient sleep; excessive stress; specific foods and sensory stimuli, such as bright or scintillating lights; intense smells or noise) may provide valuable diagnostic information. Sometimes, migraine accompaniments and typical precipitants may be missing, but vestibular migraine is still considered the most likely diagnosis after other potential causes have been investigated and appear unlikely. In this case a favorable response to antimigraine drugs may support the suspicion of an underlying migraine mechanism. However, apparent efficacy of a drug should not be regarded as a definite confirmation of the diagnosis because spontaneous improvement, placebo response, and additional drug effects (eg, anxiolytic or antidepressant) have to be taken into account.

In most patients, the general neurologic and otologic examination is normal in the symptom-free interval. Neuroophthalmological evaluation may reveal mild central deficits such as persistent positional nystagmus and saccadic pursuit, particularly in patients with a long history of vestibular migraine (64; 54). Interictal head-shaking nystagmus was observed in 50% of patient swith vestibular migraine (66).

About 10% to 30% of patients with vestibular migraine have unilateral hypoexcitability to caloric stimulation whereas lateral video head impulse testing is normal in almost all patients (15; 84). About 10% have directional preponderance of nystagmus responses (15; 66). Such findings, however, are not specific for vestibular migraine as they can be found also in migraine patients without vestibular symptoms and in many other vestibular syndromes. Persistent positional nystagmus recorded after removal of visual fixation appears to be a more specific finding in the asymptomatic interval occurring in about half of the patients (84). Interestingly, patients with vestibular migraine become nauseous after caloric testing much more often than patients with nonvertiginous migraine. Bilateral loss of vestibular function of unknown origin is linked to migraine, but the relationship to vestibular migraine is unclear (47). Vestibular evoked myogenic potentials (VEMPs) can be abnormal in vestibular migraine, but the type and frequency of abnormalities from various studies are heterogeneous (12; 25; 84).

During the acute phase of vestibular migraine, most patients have central spontaneous nystagmus of mostly low velocity, positional nystagmus, or a combination of the two (73; 60; 83). About one third of patients with spontaneous nystagmus have up- or downbeating nystagmus whereas two thirds have horizontal nystagmus (83). Positional nystagmus is usually persistent and mostly horizontal, beating either towards the lower ear (geotropic nystagmus) or towards the upper ear (apogeotropic nystagmus) (60; 39; 84). Occasional patients have a peripheral type of spontaneous nystagmus and a unilateral deficit of the horizontal vestibulo-ocular reflex (73). Imbalance is a regular finding during acute attacks, whereas transient mild to moderate hearing loss is less common and easily missed (73).

Prognosis and complications

A follow-up study 9 years after initial diagnosis found that almost 90% of patients were still symptomatic with recurrent vertigo (64). Both cochlear symptoms during attacks (49%) and mild bilateral sensorineural hearing loss with a downsloping pattern on audiometry (18%) were more common on follow-up than at initial presentation. Mild ocular motor abnormalities in the asymptomatic interval become more common during the course of the disease (64; 54). Bilateral vestibulopathy may rarely develop in the course of vestibular migraine (79). Common comorbidities include psychiatric disorders in at least 50% of the patients (35; 09) and motion sickness in almost half of the patients (69; 80). Patients with vestibular migraine report cognitive dysfunction more often than controls (62), whereas comprehensive testing in 19 patients and 21 controls failed to objectify this (Demirhan and Celebisoy).

Clinical vignette

A 40-year-old woman presented to a dizziness clinic because of episodic vertigo, which had started 3 years previously and which had subsequently recurred about four times per year. Attacks often started in the morning with intense spinning of the visual surround, nausea, vomiting, diarrhea, and inability to walk unaided. She had never experienced auditory symptoms or headaches during the attacks, but she had noted sensitivity to light on most occasions. The vertigo was aggravated by changes of head position. After 2 days, the spinning and nausea usually subsided, but the patient needed another 2 weeks to recover completely. During an acute episode, the patient was admitted to a hospital where a spontaneous nystagmus to the left and a moderate caloric hypoexcitability of the right labyrinth were documented; these signs had resolved when she was examined 1 week later. Other investigations, including cranial MRI with diffusion-weighted imaging, audiogram, Doppler ultrasound of the extracranial carotid and vertebral arteries, lumbar puncture, and a cardiological workup, were normal. She was treated with betahistine for suspected Meniere disease, but this did not prevent or modify subsequent attacks.

On questioning, she revealed a history of migraine for more than 20 years until 4 years ago with severe throbbing left hemicranial headaches, nausea, and sensitivity to light and noise. These headaches had often started on the first day of menstruation just like her current vertigo attacks. Neurologic and vestibular examination was normal. No further tests were ordered. The patient was diagnosed with vestibular migraine, and prophylactic treatment was initiated. For acute attacks, she was advised to take a vestibular suppressant. Later, the patient sent a card expressing her gratitude for being diagnosed with a benign condition after a long odyssey through the health care system.

Biological basis

Etiology and pathogenesis

The cause of vestibular migraine is unknown, just as in migraine. Genetic, neurochemical, and inflammatory mechanisms have been proposed.

Genetics. Clustering of vestibular migraine in some families suggests that a genetic mechanism may increase susceptibility (59). A positive family history of migraine was reported by 70% and a family history of vertigo by 66% of the patients in a large prospective series (69). The locus of familial vestibular migraine has been linked to chromosomes 22q12 and 5q35, but no specific genetic defect has been identified (40; 04).

Genetic defects of a voltage-gated calcium channel have been identified as the cause of familial hemiplegic migraine and episodic ataxia type 2. As both of these paroxysmal disorders can have vertigo and migraine headache as prominent symptoms, a defective gene in the same region is a candidate mechanism for vestibular migraine. So far, however, no such genetic defect could be identified (72).

Cortical mechanisms. The neural mechanisms of vestibular migraine are still obscure. The variability of symptoms and clinical findings, both during and in between attacks, suggests that migraine interacts with the vestibular system at various levels (26). Spreading depression, which presumably causes or is coincident with the migraine aura, may play a role in patients with short attacks (21). Spreading depression is a cortical phenomenon, which may produce vestibular symptoms when it reaches the multisensory vestibular cortex in the posterior insula and at the temporoparietal junction. Also, patients with vestibular migraine have altered visual-vestibular interactions at the cortical level as demonstrated by elevated vestibular thresholds after visual motion exposure (08). Cortical volumetry using 3-Tesla MRI showed decreased gray matter volume in several regions of the vestibular cortex in patients with vestibular migraine compared to healthy controls (86).

Release of neurotransmitters and trigemino-vestibular interaction. Several findings during the acute stage of vestibular migraine, including complex positional nystagmus, cannot be explained by cortical dysfunction but point to brainstem or cerebellar involvement. Some of the neurotransmitters that are involved in the pathogenesis of migraine, such as calcitonin gene-related peptide (CGRP), serotonin, noradrenaline, and dopamine, are also known to modulate the activity of central and peripheral vestibular neurons and could contribute to the pathogenesis of vestibular migraine (21; 05; 85). There is widespread distribution of CGRP-expressing fibers and CGRP receptors in the brainstem and cerebellum, not only in those areas involved in migraine, but also in the vestibular nuclei (28). Several animal experiments underline the role of CGRP in the central vestibular system. The density of CGRP as expressed by CGRP immunoreactivity in the vestibular nuclei increases in rats after exposure to a vestibular rotatory stimulus (82). In a chronic migraine model with recurrent intraperitoneal administration of nitroglycerin, an upregulation of CGRP in the trigeminal nucleus caudalis and vestibular nuclei has been observed in rats (85). Furthermore, rats in the migraine model group showed imbalance as compared to a control group. Knockdown of CGRP by application of a viral vector prevented imbalance after exposure to nitroglycerine. Likewise, the application of the CGRP receptor antagonist olcegepant prevented allodynia and imbalance in the chronic migraine model (70).

In a human experimental model of vestibular migraine, trigeminal activation by painful electrical stimulation of the forehead produced spontaneous nystagmus in migraine patients but not in controls, indicating that those with migraine have a lowered threshold for crosstalk between trigeminal and vestibular brainstem structures (50).

Inflammation. Plasma extravasation from dural vessels causing transient meningeal inflammation is considered to be a key mechanism in migraine. In mice, serotonin-induced plasma extravasation was observed not only in the dura mater, but also in the inner ear (34). Vestibular and nociceptive pathways show neurochemical similarities and share central pathways for perception, interoception, and affect (05). However, plasma levels of inflammatory mediators were not found to be elevated during acute attacks of vestibular migraine and during prophylactic treatment as compared to healthy controls (32).

Endolymphatic hydrops. Endolymphatic hydrops was identified by gadolinium-enhanced MRI in several patients with vestibular migraine and auditory symptoms (58), though the prevalence is rare in contrast to Meniere disease (81). It remains unclear whether these patients were misdiagnosed and have in fact a variant of Menière disease (with migraine symptoms) or whether vestibular migraine can be associated with hydrops.

Epidemiology

The lifetime prevalence of vestibular migraine has been estimated at 1% in a population-based study (57). In patients with unclassified recurrent vertigo, the prevalence of migraine ranges from 60% to 80% (16). Conversely, vestibular migraine was diagnosed in 13% of unselected patients presenting to a headache clinic (19) and in 30% of a migraine cohort (37). In the general population, migraine headaches and vertigo coincide about three times more often than expected by chance alone (57). The 1-year prevalence of benign paroxysmal vertigo of childhood, a pediatric variant of vestibular migraine, was 2.6% in a population-based study and accounted for 14% of diagnoses in a systematic review of pediatric patients with dizziness (01; 22). The more specific category of vestibular migraine appears to be the most common diagnosis in children with vertigo (24%) (38; 22). In adult patients, vestibular migraine usually manifests itself several years after the onset of migraine headaches (69), but a recall bias neglecting attacks in childhood or adolescence may contribute to this finding.

Differential diagnosis

The differential diagnosis of vestibular migraine includes other disorders with recurrent vestibular symptoms. Distinction is mainly based on exploration of attack duration, provoking factors, and accompanying features.

Menière disease. Menière disease presents with vertigo attacks lasting between 20 minutes and 12 hours. Auditory symptoms, including hearing loss, tinnitus, and aural fullness are required for the diagnosis. These symptoms are often transient in the early stages of the disease and become permanent later. Both hearing loss and tinnitus preferentially affect the low-frequency range (45). Mild hearing loss, tinnitus, and aural fullness may also occur in vestibular migraine, but hearing loss does not progress to profound levels (64). Furthermore, when hearing loss develops in vestibular migraine, it is typically bilateral, whereas symmetrical evolution of hearing loss is very rare in Menière patients. Migraine is more common in patients with Meniere disease than in healthy controls (63). Patients with features of both Meniere disease and vestibular migraine have been repeatedly reported (63; 53; 68). Similarly, migraine headaches, photophobia, and even migraine auras are common during Meniere attacks (63; 13; 46). The pathophysiological relationship between vestibular migraine and Menière disease remains uncertain. Endolymphatic hydrops, the hallmark of Menière disease, may be detected on contrast enhanced MRI in a small subset of patients with vestibular migraine and auditory symptoms (58). In the first few years after onset of symptoms, differentiation of vestibular migraine from Meniere syndrome may be particularly challenging as Menière disease can be monosymptomatic with vestibular symptoms only in the early stages of the disease.

Benign paroxysmal positional vertigo (BPPV). Vestibular migraine may present with purely positional vertigo, thus, mimicking BPPV. Direct observation of nystagmus during the acute phase may be required for differentiation. In vestibular migraine, positional nystagmus is usually persistent and not aligned with a single semicircular canal. Symptomatic episodes tend to be shorter with vestibular migraine (minutes to days rather than weeks) and more frequent (several times per year with vestibular migraine rather than once every few years with BPPV) (71). Clinicians should be aware that a history of migraine increases the risk for BPPV migraine 2-fold (20).

Transient ischemic attacks (TIAs). A differential diagnosis of vertebrobasilar TIAs must be considered, particularly in elderly patients. Suggestive features include onset after the age of 60 years, total history of attacks of less than 1 year, vascular risk factors, sudden onset of symptoms, and angiographic or Doppler ultrasound evidence for vascular pathology in the vertebral or proximal basilar artery.

Vestibular paroxysmia. Vestibular paroxysmia presents with brief attacks of vertigo, lasting less than a minute, which recur many times per day. Attacks are usually prevented by carbamazepine. Vestibular paroxysmia is caused by vascular compression of the vestibular nerve.

Episodic ataxia type 2. Vestibular migraine shares several clinical features with episodic ataxia type 2. In both disorders, a history of migraine and a positive family history for episodic vertigo are often present. Episodic ataxia type 2 is a rare autosomal dominant paroxysmal disorder of early onset characterized by episodes of incoordination and truncal ataxia. In contrast to vestibular migraine, onset after the age of 20 years is exceptional in episodic ataxia type 2. Attacks are commonly triggered by physical and emotional stress and typically last hours. In about half of the patients, at least one of the following can be found: vestibular symptoms with vertigo, nausea and vomiting during attacks, generalized weakness during attacks, gradual progressive baseline ataxia, and a history of migraine. Between attacks, the vast majority of patients present with gaze-evoked nystagmus, and a third present with spontaneous or positional downbeat nystagmus (31). These interictal ocular motor signs are an important key to differentiate vestibular migraine from episodic ataxia type 2 as they are inconsistent and subtle in the former and prominent in the latter. Genetic testing is commercially available for episodic ataxia type 2 and identifies a mutation in the CACNA1A gene in about 60% of patients.

Psychiatric dizziness syndromes. Anxiety and depression may cause dizziness and likewise complicate a vestibular disorder. Anxiety-related dizziness is characterized by situational provocation, intense autonomic activation, catastrophic thinking, and avoidance behavior. About 50% of patients with vestibular migraine have comorbid psychiatric disorders (36; 35), and vice versa, there is an increased prevalence of migraine and vestibular migraine in patients with persistent postural-perceptual dizziness (65). The contribution of migrainous and psychiatric mechanisms may be difficult to disentangle in patients with chronic variants of vestibular migraine (26; 02).

Migraine induced by vestibular activation. Caloric stimulation often triggers migraine attacks within 24 hours in patients with migraine, which shows that, in susceptible individuals, migraine attacks can be a secondary effect of vestibular activation rather than its cause (52). The high rate of headaches and other migraine symptoms during Meniere attacks can possibly be explained by this mechanism. Thus, migraine symptoms during vertigo attacks do not prove a diagnosis of vestibular migraine and consideration of other potential causes remains mandatory.

Diagnostic workup

History will usually provide more clues for the diagnosis than vestibular testing because there are no abnormalities that are specific for vestibular migraine. Clinical examination of the vestibular, auditory, and nervous systems in the asymptomatic interval may show mild abnormalities but serves mainly to exclude other disorders producing permanent damage (49). In patients with a clear-cut history, no additional vestibular tests are required. However, vestibular testing in the interval can be useful to reassure both patient and doctor that there is no severe abnormality, such as a complete canal paresis, which would rather suggest another diagnosis. Audiometry helps to differentiate vestibular migraine from Meniere disease. MRI is required in patients presenting with central abnormalities and no previous history of similar attacks.

Management

Drug treatment. In many patients, vestibular migraine attacks are severe, protracted, and frequent enough to warrant acute or prophylactic treatment. Unfortunately, current treatment recommendations are based on anecdotal experience and expert opinion rather than on randomized clinical trials (48; 14). One small, controlled study on the use of zolmitriptan for acute vestibular migraine was inconclusive (56). There is limited evidence for prophylactic treatment from a single randomized trial using 10 mg flunarizine (not marketed in the U.S.), which showed a significant decrease in frequency and severity of attacks compared to placebo (43). Metoprolol was ineffective in a randomized controlled trial involving 130 patients (07). Systematic reviews of treatment trials found insufficient evidence for efficacy of any acute or prophylactic treatment (76; 77). Accordingly, no specific drug can be favored, and recommendations remain tentative (67). Beyond the usual migraine prophylactic agents, acetazolamide and lamotrigine have been advocated.

The role of CGRP antagonists for treatment of vestibular migraine is still unclear as adequate trials are lacking. An uncontrolled case series showed improvement of vestibular symptoms in most patients, but placebo responses and spontaneous recovery (regression to the mean) couldn’t be excluded (67). To close the current evidence gap, future treatment trials for migraine may combine headache-related and vestibular endpoints. Also, publicly funded trials are needed as industry funding usually wanes after a drug has been approved and even more so after patents have expired. An international expert group has formulated shortcomings in the current understanding, diagnosis, and treatment of vestibular migraine with the intention to guide future research (49).

The selection of a particular drug will be influenced by expected side effects (eg, orthostatic hypotension with beta blockers or weight gain with valproate, flunarizine, or amitriptyline). Treatments should be started with a low dose, which is then gradually increased. Patients can monitor frequency and severity of their attacks in a diary. Treatment response should be evaluated after 3 months. A greater than 50% reduction in attack frequency is a reasonable goal. For acute attacks, conventional antiemetic drugs such as dimenhydrinate or benzodiazepines seem most appropriate. Intravenous methylprednisolone (1000 mg per day for 1 to 3 days) terminated prolonged attacks or exacerbations with daily recurrences in a small clinical series of four patients (61).

Nonpharmacological treatment. Nonpharmaceutical treatment should not be neglected and may be even more effective than drugs in individual patients. A thorough explanation of the migraine origin of the attacks can relieve unnecessary fears. Environmental modification (eg, avoidance of identified triggers, and ensuring regular sleep schedules, adequate sleep, and regular meals) has a firm place in migraine prophylaxis. Regular exercise seems to have a prophylactic effect on vestibular migraine (41). Selected patients, particularly those with persistent symptoms or anxiety between attacks, may profit from vestibular rehabilitation (03; 06). However, the overall evidence for nonpharmacological interventions is still very low (78).