Video-head impulse test results in patients with Menière’s disease related to duration and stage of disease

1. Introduction

Menière’s disease (MD) lacks a diagnostic reference standard to objectify the diagnosis. The diagnosis of MD is based upon its clinical characteristics accompanied by documented hearing loss [7] rather than the use of vestibular tests. Although the caloric test may be considered as the reference standard for assessing vestibular function, great variability in the results is found in MD, making the test unsuitable to serve as a reference standard [1, 28].

In the pursuit of objectifying MD, scientific studies combined results of vestibular tests such as the caloric test and the recently developed video-head impulse test (vHIT) [3, 20].

Although both the caloric test and the vHIT measure vestibular function, they capture distinct phenomena [24, 25]. While the caloric test uses a non-physiological low-frequency stimulus, the vHIT measures head and eye movements during physiological high-velocity rotatory head thrusts [14, 23].

Based on previous research in patients with MD we know that caloric test responses decrease most profoundly in the first decade after which responses stabilize at a fixed level of hypofunction of approximately 50% [8, 9,15]. Similarly to that of caloric testing, one would expect that abnormal VHIT results are more common in MD in the chronic stage. Maire and van Melle [21] found that in the chronic phase of the disease (>12 months), the VOR gain decreased, resulting in an abnormal vHIT result. One may argue whether the chronic stage begins after one year of vertigo symptoms since the total duration of disease is estimated to last 20 years [27]. On the other hand, Cerchiai et al. [5] found that the proportion of abnormal vHIT results was similar in patients with ‘early’ MD (5 years or less) compared to those with ‘late’ MD (more than 5 years).

Based on the disagreement between pervious study results, we aim to evaluate whether the vHIT is more often abnormal in patient with a later stage of disease than in those with an early stage, related to either duration of vertigo attacks in years or level of hearing loss. In case progression of disease is consistently related to an increase of abnormal vHIT results, this may serve as a diagnostic hallmark in the course of the disease.

2. Materials and methods

In this study, we retrospectively evaluated patients diagnosed with MD who visited a tertiary dizziness clinic from 2012 to 2015 (n = 343). Patients were included if they met the criteria for ‘definite’ MD as defined by the AAO-HNS in 1995 [7] (n = 250). No patients with ‘certain’ MD were included as we did not seek histopathological confirmation in patients with ‘definite’ MD. A single attack of vertigo accompanied by unilateral hearing loss was regarded to be clinically more compatible with (viral) labyrinthitis, therefore patients matching the criteria of ‘probable’ MD were not included in this analysis either. We excluded ‘possible’ MD as it represents a less well defined clinical entity and this population may contain vertigo related diseases as well (e.g. vestibular migraine) [18]. We categorized patients according to the duration of disease based on the duration of symptoms in years. In addition, the stage of the disease was based on the four tone average of the pure-tone thresholds at the frequency of 0.5, 1, 2 and 3 kHz of the worst audiogram in accordance with the AAO-HNS 1995 guideline [7]. Stage I was defined as a four tone average (rounded to the nearest whole number) of 25 decibel (dB) or less. Stage II represented an average hearing loss of 26 to 40 whereas patients with a Stage III suffer from an average hearing loss between 41 to 70 dB. Stage IV included patients with an average hearing loss of more than 70 dB. All data were analyzed anonymously and procedures were performed in accordance with the ethical standards of the institutional research committee and the 1964 Helsinki declaration and its later amendments.

3. Results

Within the MD population, data with regards to the vHIT and the caloric test was available in 89 (36%) patients. Patients had a mean age of 55±5 years and 42 (47%) were male. We found that 84 (94%) suffered from symptoms of unilateral MD and 5 (6%) had symptoms of bilateral MD. The mean duration of disease was 5±6.2 years. The mean vHIT gain on the left side was 0.91±0.14 and 0.99±0.2 on the right side. Based on the level of hearing loss, 3 (3%) patients met the criteria for Stage I. Fifteen (17%) patients could be classified as stage II, whereas Stage III and Stage IV comprised 56 (63%) and 13 (15%) of the patients respectively. In two patients no information on the amount of hearing loss was available. Table 1 shows the percentage of abnormal caloric tests and VHIT test results related to stage of disease. A progressive increase in the number of patients with an abnormal caloric test was seen when results were related to stage of the disease. Results of the per ear analyses are shown in Fig. 1. In 85% of the ears an unilateral vestibular hypofunction or bilateral hypofunction was identified by the caloric test. We found that in 10% of the ears, both the vHIT test and the caloric test identified an ipsilesional vestibular hypofunction.

OPEN IN VIEWER

Fig.1

Results per ear regarding ipsilesional and contralesion vHIT test results based on a unilateral caloric weakness or bilateral hypofunction.

Previous research has shown that caloric test responses in patients with MD tend to decrease in the first decade of disease [8, 23]. Similar to that of the caloric test, one would expect abnormal vHIT results to be more common in MD patients who are in a later stage of the disease. We evaluated whether vHIT abnormality is more common in patients who are in a later stage of the disease as opposed to those in an earlier stage.

The vHIT was related to the duration of symptoms in years and the stage of disease. No changes in the proportion of abnormal vHIT results were found when related to progression of disease.

A recent retrospective chart review by Cerchiai et al. [27], evaluated vHIT findings of ‘definite’ MD patients who were treated with either intratympanic gentamicin or conservatively (dietary modifications combined with acetazolamide, hydrochlorothiazide or betahistine) and related results to duration of disease. In line with our study findings, no relation was found between the VOR gain results and the duration of disease.

Zulueta-Santos et al. [31] evaluated the distribution of normal and abnormal vHIT results in all planes (i.e. evaluation of horizontal, superior and posterior canals) in 36 patients with definite unilateral MD. A rather diverse set of results was found for the affected and unaffected ear. They did not find a relation between the degree of canal function loss expressed by the vHIT and the duration of disease or hearing loss.

However, Maire et al. [15] concluded that the stage of the disease could affect the vHIT test result. They stated that in early Menière’s disease, the VOR gain is higher towards the affected side as opposed to the intact side, while the opposite is seen in patients in a later stage of the disease. Funabiki et al. [10] used the direction of nystagmus attempting to explain this dynamic change in the peripheral vestibular system. They stated that VOR gain was higher towards the affected side when an ipsilateral nystagmus was present and decreased when there was a contralateral nystagmus. Subsequently, Hozawa et al. [12] found that just prior to vertiginous periods the VOR gain was higher in the direction of the ear with MD versus the contralateral unaffected ear.

The dissociation between the vHIT and the caloric test may be explained by damage primarily to the low-frequency spectrum in the vestibular apparatus in Meniere’s disease. The caloric response represents a low frequency (0.002–0.004 Hz) rotation, whereas the vHIT response involves a more physiological, high frequency rotation, representing a frequency up to 5 Hz [14, 19].

Moreover, various studies on the vHIT have used different techniques to calculate the gain. One common method is the local sample point to point gain in which the velocity gain is calculated from a fixed interval length 60 milliseconds after the head impulses is started [3]. In our study the gain was calculated from the area of the head and eye velocity over the responses after which it was divided to yield the gain. Based on previous research on this matter, standard higher gains were found using this method from which it was recommended to use higher cut-off values [6, 30].

With respect to caloric test responses and vestibular hypofunction, previous studies showed variability in cut-off values. A recent consensus document published by Strupp et al. [29] used V_max cut-off values of less than 6 degrees per second to define vestibular hypofunction whereas a cut-off value of 35 degrees per second is used by the University Medical Centre in Maastricht. We used a relatively low cut-off value of 15 degrees per second to identify patients with a vestibular hypofunction, when comparing it to the cut-off values defined by the University Medical Centre Maastricht. A lower cut-off value results in a larger group of patients with a false positive result. In clinical practice we prefer to include patients with false positive results to minimize the proportion of false negative results, as a smaller group of patients will be deprived of further diagnostic evaluation [22].

It is proposed that different nerve fibers in the crista ampullaris are stimulated depending on the velocity of the head-movement [22]. Caloric testing would stimulate the regular afferent from the peripheral zones whereas high-velocity head movements during the vHIT are believed to stimulate irregular centrally located afferents [19]. Based on histopathological research it is suggested that MD mainly affects the peripheral zones and therefore leads to abnormality of the caloric test [25]. As the disease progresses, central damage to the crista will then cause abnormality of the VOR responses. Although the theory seems plausible, our study demonstrated various responses of normal and abnormal vHIT results in patients with both recent onset of symptoms and those with progressed disease.

To date, this is the largest study analyzing vHIT characteristics in patients with MD. Moreover, this is the first study quantifying normal and abnormal vHIT test results related to the duration and the stage of disease. Another strength of this study was the use of AAO-HNS 1995 guidelines by which our MD population was defined and staged in a standardized way.

However, there are some limitations which need to be considered when interpreting our study findings.

First, analysis could only be performed in 89 out of 250 patients (36%) due to the absence of information on both the caloric test results and the vHIT results for the remainder of the population. Therefore, selection bias may have influenced the results.

Mainly, the reduction of the sample size is based on a practical limitation. While two caloric systems were available for caloric testing, only one test system was accessible at the department for the vHIT. The reduced number in vHIT results can be considered as random missing information as no pre-selection criteria on patients were applied before performing the diagnostic tests. Therefore, it is unlikely that the results of our study are clinically not applicable.

Secondly, both tests have been evaluated in a retrospective way, between vertigo spells and without serial evaluation. A recent retrospective case series by Lee et al. [17] analyzed results of both tests during vertigo attacks in patients with MD at various stages of the disease. They concluded that vHIT results tend to fluctuate when tested during the irritative (vertigo) phase compared to the paretic phase. To fortify information on the canal function based on vHIT test results in MD, serial evaluation of individuals may provide relevant information with respect to changes of results over time.

Thirdly, only vHIT results for the horizontal canal were analyzed whereas the vertical and lateral semicircular canals may also be of importance as the whole membranous labyrinth is believed to be affected in MD. It is possible that vertical and lateral semicircular canals are differently affected by the disease and may be able to show a consistent change in vHIT results when related to disease progression.

Lastly, eight patients were treated with intratympanic injections whereas the remaining received conservative treatment (medication, Prisma® glasses). Due to the vestibulotoxic effect of gentamicin, abnormality of the vHIT is expected in all these patients. In six of these patients, information on the age of onset was available. Subanalysis revealed that these patients had varying disease duration and both normal and abnormal vHIT results were found. However, the small sample size of patients analyzed in the intratympanic group should be taken into account when interpreting these findings.

We retrospectively analyzed vHIT test results related to the stage and duration of disease, and found no relation between the two. Future, prospective, serial evaluation of individuals analyzing all semicircular canals may be able to provide information on consistent changes of the vHIT and may serve as an objective finding in the diagnostic process in MD.

Funding

This work was supported solely from institutional and/or departmental sources from the Apeldoorn Dizziness Centre, Gelre Hospital, Albert Schweitzerlaan 31, 7334 DZ Apeldoorn, The Netherlands.

Compliance with ethical standards