Idiopathic bilateral vestibulopathy.

We report the clinical features of 22 patients with acquired bilateral vestibulopathy of unknown cause. All had either absent or markedly decreased responses to both caloric and rotational testing. They presented with dysequilibrium and imbalance, worse at night; most reported oscillopsia but none had associated hearing loss or other neurologic symptoms. Nine reported prior prolonged episodes of vertigo consistent with the diagnosis of bilateral sequential vestibular neuritis. Of the remaining 13, none had exposure to known ototoxins or a positive family history. Idiopathic bilateral vestibulopathy is an important cause of progressive imbalance in adults and should be considered even though hearing is normal.

Because the function of the vestibular system overlaps that of other sensory systems, a slowly developing bilateral vestibulopathy can be relatively silent. Vertigo may be absent and the loss of vestibulospinal and vestibulo-ocular reflex function leads to nonspecific complaints of imbalance and visual distortion. In the absence of associated hearing loss, these symptoms may be incorrectly attributed to abnormal visual or motor function (particularly in the elderly). We report the clinical features of 22 patients with acquired bilateral vestibular loss (documented with caloric and rotational testing) without associated hearing loss. Two distinct groups emerged: a slowly progressive group and a sequential episodic group.
Case reports. Progressive-Case 1. For 6 years, a 65year-old man noted a slowly progressive imbalance when walking, particularly in the dark. There was no history of vertigo, hearing loss, or tinnitus. He had been treated for atypical facial pain 10 years earlier with numerous medications, including carbamazepine and phenytoin, but had never received known ototoxic drugs. He underwent an uncomplicated coronary bypass operation 1 year after the onset of imbalance.
On examination, Romberg's sign was present, and he had difficulty walking in tandem even with eyes open. There was no response to bithermal caloric stimulation or to lowfrequency rotational stimuli; response to high-frequency rotation was decreased. Hearing was normal.
Case 2. A 56-year-old man complained of progressive imbalance and difficulty focusing his vision for the past year. He could not read signs, and objects appeared to bounce up and down when he walked. For at least 6 years he had experienced brief episodes of vertigo (lasting less than a minute) usually induced by quick head movements. When the first episode occurred while driving on the freeway, he was able to pull to the side and then resume driving within minutes. Medical history was unremarkable. There was no known ototoxic drug exposure.
The only findings on neurologic examination were Romberg's sign and mildly impaired tandem walking. Head shaking did not induce nystagmus. Caloric responses were absent and rotational responses were decreased. Audiometric examination revealed a bilateral "notched" sensorineural hearing loss at 4,000 Hz characteristic of noise-induced hearing loss.

NEUROLOGY 39 February 1989
Sequential-Case 3. Three months after an uncomplicated delivery, a 32-year-old woman had several episodes of vertigo, nausea, and vomiting lasting at least 6 to 8 hours. Evaluation at that time revealed a 77% right vestibular paresis to caloric stimulation and a decreased response to sinusoidal rotation at low frequencies (figure 1). Hearing was normal. Eight months later she reported that the episodes of vertigo had decreased in frequency and severity but she noticed a more persistent nonspecific dizziness and mild unsteadiness when walking, particularly after rapid turns. Response to caloric stimulation on the left side had decreased compared with prior testing. Eighteen months later the episodes of vertigo disappeared, but she complained of unsteadiness of gait, most prominent in the dark. She had to hold onto the railing to climb stairs. She also noted oscillopsia while jogging. Caloric responses were now symmetrically decreased (peak velocities all below 5"/sec) ( figure 1A). Rotational responses were absent at low frequencies and decreased at high frequencies (figure 1B). Hearing remained normal. Case 4. A 52-year-old man developed the acute onset of vertigo that slowly cleared over the following week. Over the subsequent 6 months he had three episodes of vertigo lasting hours. Bithermal caloric examination at that time revealed almost complete absence of response on the right side. Three years later he began noticing oscillopsia and unsteadiness when walking, particularly at night. Medical history revealed only a left mastoidectomy as a child. Caloric responses (including ice water) were absent bilaterally. Audiometric examination showed only a mild bilateral symmetrical sloping sensorineural hearing loss beyond 2,000 Hz. Brainstem auditory evoked responses and CT of the brain were normal. Case 6. A 67-year-old woman complained of persistent imbalance for the past 2 years. She noted the imbalance most at night; it was associated with blurring of vision but no definite oscillopsia. She reported three severe bouts of vertigo, nausea, and vomiting (lasting for days) 15 years earlier but denied any vertigo for more than 10 years. Medical history was positive for hypertension and diffuse atherosclerotic vascular disease (femoral artery bypass 4 years prior, and a transient rightsided weakness lasting 12 hours 4 months prior).
Examination revealed Romberg's sign, impaired tandem walking, and a mild sensory neuropathy. Bithermal caloric responses were absent bilaterally. Hearing was normal. MRI of the brain was normal. Frequency (Hz)

Figure 1. Serial caloric (A) and rotational (B) responses in case 3. For bithermal caloric testing, the lower end of the normal range for peak slow-phase velocity is 5"Isec. For rotational testing, the normal mean 2 standard deviation is given for each frequency by the closed squares and vertical bars.
Methods. All patients underwent a complete history and neurologic examination. The history focused on possible . . known causes of bilateral vestibulopathy including a detailed drug history. Vestibular function was assessed with bithermal caloric and rotational stimuli. Complete details of the test procedure are published elsewhere.' In brief, caloric and rotational responses were recorded using electro-oculography with the patient's eyes open in darkness (constant mental alerting). Recordings were also made for (1) spontaneous and positional nystagmus, (2) visual tracking (saccades and smooth pursuit), and (3) fixation suppression of vestibular nystagmus.

Results. Clinical features.
All patients had documented bilateral vestibular loss without associated hearing loss. None were exposed to known ototoxic agents. All complained of disequilibrium when walking (worse at night) and most reported either visual blurring or oscillopsia with head movements (particularly noticeable while walking or running). Examination re-vealed only mild unsteadiness of gait most evident on tandem walking. All were able to stand in the Romberg position with eyes open, but were unstable with eyes closed.
Two subgroups were identified based on a sequential course with previous prolonged episodes of vertigo (hours to days) or a slowly Progressive course. Several patients in both groups reported brief episodes of vertigo (seconds) induced by quick head movements or position change. In the sequential group the mean age of onset was 57.0 years (range, 39 to 78) with a mean duration of symptoms of 4.3 years (range, 1 to 19), while in the progressive group the mean age of onset was 56.7 years (range, 34 to 84) with a mean duration of 6.7 years (range, 1 to 16). The male : female ratio was 6 : 7 in the progressive group and 5 : 4 in the sequential group.
Associated disorders. Associated illnesses and their frequency of occurrence in the two groups were as follows (sequential = S, progressive = P): hypertension-2S, 3P; myocardial infarction-lS, 1P; transient cerebral ischemia-2s; migraine-& 1P; diabetes mellitus-lS, 1P; Paget's disease-1P; and breast carcinoma-1P. Since all of these disorders are common in this age group, their association with the vestibular loss may be coincidental. One of the patients with transient cerebral ischemic attacks had recurrent episodes of vertigo along with brainstem and cerebellar symptoms. Angiography documented complete occlusion of one vertebral artery and 80% occlusion at the takeoff of the other. The recurrent episodes spontaneously stopped. The subsequent development of bilateral vestibular loss might be attributed to bilateral ischemia of the labyrinths but hearing remained unaffected. One patient had Paget's disease involving the femur and pelvis; CT of the skull was normal.
Oculornotor function testing. All patients eventually had bilateral decreased response to bithermal caloric testing (ie, the peak slow-phase velocity for all responses was less than 5"/sec). Four patients in the sequential group had a documented unilateral caloric hypoexcitability earlier in their course (eg, cases 3 and 4) (figure 1). One patient in the progressive group and two in the sequential group had a typical benign paroxysmal positional nystagmusz on rapid positional testing and one in each group had a direction-changing static positional nystagmus in the lateral positions (eyes open in darkness). Rotational responses were symmetrically impaired in all patients (figure 2). As previously described' the gain was decreased maximally at low frequencies. Phase lead was increased at low frequencies but was normal at higher frequencies. Saccades, smooth pursuit, and fixation supression of vestibular nystagmus were normal in all 22 patients. Auditory testing. Nine patients (4 progressive, 5 sequential) had a bilateral symmetrical high-frequency sensorineural hearing loss consistent with either noise exposure or aging (presbycusis). One patient in the sequential group had a long-standing (since infancy) unilateral sensorineural hearing loss (presumably due to mumps) that did not change with the onset of vestibular symptoms.

Figure 2. Comparison of the mean patient and normal rotational responses at multiple sinusoidal frequencies. Vertical bars represent one standard deviation.
Discussion. Bilateral vestibular loss should be considered in any patient complaining of oscillopsia, unsteadiness of gait, or both. There may be a history of vertigo. Simmons3 found a probable diagnosis in 39 of 43 patients with bilateral absent caloric responses while Chambers et a14 found a probable diagnosis in 17 of 24 such patients. Common diagnoses in these two studies included ototoxic drugs, infection (particularly meningitis), congenital malformations, tumors, endolymphatic hydrops, and autoimmune inner ear disease. Unlike our patients, however, most of the patients in these two reports had combined auditory and vestibular loss.
There are potential pitfalls in the diagnosis of bilateral vestibulopathy using caloric testing. A caloric stimulus induces endolymph flow in the horizontal semicircular canal by creating a temperature gradient from one side of the canal to the other. Anything that interferes with heat transfer through the temporal bone (eg, cerumen in the external canal, thickened bone) can lead to artifactualy decreased responses. Furthermore, a caloric stimulus tests only the low-frequency range of the vestibulo-ocular reflex. By contrast, rotational testing uses precise physiologic stimuli over a broad frequency range; it is the best method for identi&ing and following the progress of patients with bilateral vestibular loss.' All of our patients had decreased caloric and rotational responses.
We divided our cases of idiopathic bilateral vestibulopathy into two groups based on the course: progressive and sequential. The latter began with recurrent episodes of vertigo suggesting asymmetric involvement initially (documented with serial caloric testing in four cases). Schuknecht and Witt5 hypothesized that bilateral sequential vestibular neuritis could produce this clinical course. Based on epidemiologic and pathologic evidence,6 they argued that a single prolonged attack of vertigo without associated involvement of the auditory system or central nervous system is usually due to viral vestibular neuritis. A subsequent similar attack involving the opposite side could produce symptoms and signs 274 NEUROLOGY 39 February 1989 of bilateral vestibular loss. All of our patients in the sequential group had initial episodes of vertigo that would have been consistent with a clinical diagnosis of viral vestibular neuritis. Two patients in this group also had transient cerebral ischemic episodes; one had complete occlusion of one vertebral artery and 80% blockage of the other on angiography. Although the labyrinthine artery divides into vestibular and cochlear branches within the inner ear, the complete loss of vestibular function with preserved auditory function is difficult to explain on an ischemic basis.
The pathophysiology of the vestibular loss in the progressive group is even more obscure. Inherited, autoimmune, and toxic disorders must be considered. It is possible that in some of our patients the vestibular loss was long-standing and stable but became more bothersome as they grew older. Verhagen et a17 reported three siblings with vestibular areflexia, without hearing loss, presumably on the basis of an autosomal recessive inheritance. X-rays of the temporal bone in the proband were normal. In two siblings, symptoms of imbalance and oscillopsia were identified only after detailed questioning, ie, they had no medical complaints. None of our patients reported a family history of similar vestibular symptoms. Although documented cases of autoimmune disease involving the inner ear produce a combined auditory-vestibular loss: one could hypothesize a selective antibody directed at the vestibular apparatus. We know of no experimental evidence to support such speculation, however. Some ototoxins (eg, streptomycin) are relatively selective for the vestibular system but none of our patients reported receiving known ototoxic drugs. Of course, they may have been exposed to an ototoxic agent of which they were unaware.

Extended major histocompatibility complex haplotypes in patients with multiple sclerosis
Article abstract-We derived complete haplotypes of the major histocompatibility complex for 33 patients with MS and their families. The DR2 allele and DR2-bearing extended haplotypes, in proportion, were overrepresented on chromosomes of MS patients compared with parental chromosomes not transmitted to MS offspring. We did not confirm previous reports that particular alleles at the BF locus are overrepresented in MS or that C2 hypocomplementemia is present.
These results suggest that the DR2 allele is a risk factor for MS, and not merely a genetic marker of the population of origin. The significance of this disease association has not been clarified. It is possible that DR2 is a marker for the population at risk for MS rather than a gene which influences susceptibility. In this regard, the DR2 allele is common in individuals of Northern European and, in particular, of Scandinavian origin; 43% of chromosomes from normal Scandinavian individuals carry DR2.l These are the very groups at highest risk for MS. Furthermore, studies of siblings concordant for MS have not found shared inheritance of parental major histocompatibility complex (MHC) genes beyond that expected by c h a n~e .~.~ In addition to the HLA genes, four complement genes and the gene for the enzyme glyoxalase are found within the MHC. The complement genes are inherited as single genetic units called "complotypes." Some specific complotype, HLA-B, DR allelic sets have been found to occur in linkage disequilibrium, which is to say that their frequency as a haplotype exceeds the frequency predicted by the individual frequencies of the alleles of which they are composed. These are called "extended hap lo type^."^ In the current study, the occurrence of these extended MHC haplotypes was assessed in families of MS patients.
Materials and methods. Thirty-three white individuals with definite MS (Schumacher criteria) and their relatives were studied. No family had more than one affected patient. Blood samples were collected and lymphocytes and plasma separated as previously de~cribed.~ HLA-A, -B, and -C antigens were identified by the standard NIH lymphocyte microcytotoxicity assay, and HLA-DR typing was performed using the technique of the Oxford (7th International Histocompatibility) workshop. Glyoxalase I (GLO), BF, and C2 typing and the identification of C4 polymorphism and nomenclature used for C4 is as described previ~usly.~ Complotypes are designated by their BF, C2, C4A, and C4B alleles, in that arbitrary order? Null or QO alleles are simply "0." Thus, SCOl stands. for BF*S, C2*C, C4A*QO, C4B*1.
A group of 1,713 normal-control MHC haplotypes was assembled from random unrelated white individuals from families without MHC-related disease and from independent haplotypes not occurring in patients but in families with a member or members with MHC-related disease (diseasenormal controls).e A population of normal chromosomes from the current study families, ie, parental chromosomes not transmitted to MS offspring, was used to assure ethnic matching for the disease chromosomes and is referred to as "familynormal." Statistical significance was determined by x2 analysis using either (observedexpected)2 /expected, Fisher's exact test, or 2 X 2 contingency tables.

Results.
Sixteen of 33 MS patients were DR2-positive in this group. The DR2 allele was present in 14 of 22 MS patients (64%) of Northern European origin and in none of six patients (0%) of Italian origin; DR2 has been previously reported not to be associated with MS in Italians.' Of the 66 chromosomes from these 33 individuals, complete MHC haplotypes could be assigned to 51 patient chromosomes. Ascertainment of DR allotype was not possible on eight chromosomes, and allotype assignment at the A or B locus was lacking on seven additional chromosomes.
The distribution of MHC haplotypes by the HLA-DR allele they carry is shown in the figure for patient chromosomes, ethnically matched family-normal chromosomes, and normal-control chromosomes. Each bar in the figure is segmented to show the absolute contribution of extended haplotypes for each type of chromosome. The DR2 allele was present on 29.4% of patient chromosomes compared with 12.5% of family-normal chromosomes and 15.8% of normal-control chromosomes; these Werences approach statistical significance (p = 0.031 MS versus family-normal chromosomes;p = 0.009 MS versus normal-control chromosomes).