Introduction

Meniere’s disease (MD) is an inner ear disorder associated with vertigo attacks and cochlear symptoms, such as tinnitus and hearing loss1. The pathophysiology of MD was first reported in 1938 as involving endolymphatic hydrops (EH)2,3; however, the underlying mechanism remains unknown. A significant number of MD patients report that their vertiginous symptoms are exacerbated by changes in weather conditions, leading to speculation about a direct link between the environment and inner ear physiology4,5. Several studies have explored this link, with Schmidt et al. reporting an association between lower atmospheric pressure, increased humidity, and a higher probability of vertigo attacks, tinnitus, and aural fullness6. Chen et al. have also found a correlation between lower atmospheric pressure and the onset of Meniere’s attacks, suggesting that low pressure, particularly during typhoon season, may aggravate EH4. Gürkov et al. reported a significant association between atmospheric pressure changes and the probability of MD episode5. However, a critical gap remains in our understanding of the physiological mechanism underlying these observations. Specifically, there is a lack of objective evidence linking changes in weather, particularly atmospheric pressure, to the volume of endolymphatic space (ELS), the presumed pathological substrate of MD. Most studies rely on patient-reported data on vertigo frequency and, thus, a subjective assessment and the distinction between BPPV-induced vertigo attacks, which are associated with one-third of all MD cases, is extremely unclear7.

We have been performing 3D analysis of the endolymphatic space (ELS) volume using 3T-MRI after intravenous injection of gadolinium enhancement, the inner ear MRI (i.e. MRI), in various studies, such as comparing the ELS volume in patients with MD with that in healthy controls8,9,10,11,12,13,14. Therefore, in the present study, we aimed to objectively determine whether there is a correlation between 24-hour changes in atmospheric pressure and ELS rates, as measured by i.e. MRI, in patients with unilateral MD (uMD) and control groups. We hypothesized that a decrease in atmospheric pressure would be associated with an increase in ELS volume in the affected ears of MD patients. We believe that the present results will help elucidate the pathophysiology of MD.

Materials and methods

The retrospective case-control study was approved by the Medical Ethics Committee of Nara Medical University (certificate number: 0889). Written informed consent was obtained from all subjects and/or their legal guardians. All study methods were performed in accordance with the Declaration of Helsinki, relevant guidelines, and regulations.

Patients

A total of 101 patients with successive definite uMD from 2014 to 2022 were enrolled in the present study according to the Bárány Society (Barany) criteria of 201515. Healthy controls comprised 53 patients with chronic rhinosinusitis without a history of vertigo/dizziness who kindly participated in this study. Sex, age, and laterality were assessed as patient background factors in all cases (Table 1). A chi-squared test was performed to examine the difference in sex composition between the two groups, and no significant difference was found (p = 0.4422). A Mann-Whitney U test determined no statistically significant difference in age between the uMD and control groups (P = 0.4272). Additionally, 4-tone-average hearing levels (HL) of 500, 1000, 2000, and 4000 Hz were collected from the unilateral MD patient group, and the group was divided into three groups according to hearing level on the affected side: group 1 < 40 dB; 40 dB ≤ group 2 ≤ 70 dB; and 71 dB < group 3.

Table 1 Demographics of patients with unilateral MD and control healthy group.
Full size table

Atmospheric pressure (JMA data)

In this study, atmospheric pressure data for Nara City, Japan, were extracted from the JMA database from 2014 to 202216. JMA is a Japanese public agency, and its meteorological data are publicly available and are reliable databases. The nearest publicly available data from Kashihara City, where our institution is located, is Nara City, so these data were used for convenience. Although Kashihara City and Nara City are 18 km apart, they are in the same basin and there is almost no difference in elevation, so we considered the atmospheric pressure data error minimal, though not perfect. Since the second MRI scan was performed at approximately 17:00, atmospheric pressure change values for the 24-hour period from 17:00 on the day before to 17:00 on the day of the MRI scan were used.

Evaluations

Inner ear MRI

Nakashima et al. reported that MRI performed 4 h after intravenous Gd infusion is useful for imaging EH17,18,19. Therefore, in the present study, MRI measurements were performed 4 h after a single intravenous administration (0.2 mL/kg or 0.1 mmol/kg body weight) of Gd-diethylenetriamine pentaacetate dimethylamide (Magnescope; Gerbe, Tokyo, Japan). 3-T MRI system with 32-channel array-head coils (MAGNETOM Verio; Siemens, Erlangen, Germany) was used. A sequence proposed by Naganawa et al., which revealed endolymphatic and perilymphatic fluids, was used.

Heavy T2-weighted (hT2W) MR cisternography was used to obtain an anatomical reference for the total lymph fluid. hT2W 3D fluid-attenuated inversion recovery sequences with an inversion time of 2250 ms yielded positive perilymph fluid images (PPI), whereas hT2W 3D inversion recovery sequences with an inversion time of 2050 ms yielded a positive endolymphatic image (PEI). A hybrid of an inverted image of the endolymph-positive signal and a negative image of the perilymph-positive signal was obtained by subtracting PEI from PPI.

MRI volumetric measurements

Three-dimensional MRI images were constructed semi-automatically to evaluate ELS conditions using a work station (Virtual Place; AZE, Ltd., Tokyo, Japan). Details of the protocol are provided below8,9,13,20. In this study, we utilized a protocol where ELS region voxels exhibited negative signal values, while perilymph space voxels showed positive signal values. PPI and PEI data were transferred, and PEI was subtracted from PPI using the fusion program integrated into our workstation. In the subtracted image, the borderline between the gray and green areas on the color bar was defined as the zero value in our program. After activating the SPACE sequence image and the “PPI-PEI” image on the workstation, components of the inner ear were identified on the SPACE sequence image, referencing anatomical drawings20, using the borderlines between the inner ear and the peripheral side of the acoustic nerve, between the end of the cochlea and the vestibule, between the three ampullae of the semicircular canals and the vestibule, and between the distal side of the common crus and the vestibule. Next, the volume of the Total Fluid Space (TFS) was calculated using the automatic voxel counting function on our workstation. Similarly, the volume of the ELS was measured by counting the number of voxels showing negative signals on the “PPI-PEI” image. Finally, the percentage of the volume of the ELS to that of the TFS was calculated (defined as the ELS percentage). In this study, these measurements were performed three times, and the average was used. A series of protocols were conducted by the trained researcher.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 9.5.1 for Windows (Boston, MA, USA) and G*-Power software (version 3.1.9.7; Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany; http://www.gpower.hhu.de/)21. For patients’ backgrounds in Table 1, the χ2 and Mann–Whitney U tests were used to compare the uMD and control groups. A two-way ANOVA was performed to compare the ELS rates of the uMD (affected and healthy sides) and control groups (Figs. 1 and 2). Spearman’s correlation was used to examine the statistical correlation between atmospheric pressure changes and ELS rates (Table 2; Fig. 3). Post hoc power analyses were used to check the effect of sample size and statistical power. All reported p-values were two-sided, with values < 0.01 considered as significant.

Fig. 1
figure 1

The endolymphatic space (ELS) rates were evaluated using a 3D-analysis of 3T-MRI in the affected and healthy side of the uMD group and that of the control group. **p < 0.01, ****p < 0.0001, two-way ANOVA, Tukey’ test.

Full size image
Fig. 2
figure 2

Hearing thresholds and vestibular ELS were positively correlated in the affected side of uMD. Spearman’s correlation coefficient. CI 99%.

Full size image
Table 2 Correlation between 24-h atmospheric pressure change and ELS rates in patients with uMD.
Full size table
Fig. 3
figure 3

Correlation between 24-h atmospheric pressure change values and the vestibular ELS rate of the affected side of patients with uMD: classification by hearing level. Spearman’s correlation coefficient. CI 99%.

Full size image

Results

The ELS rates in the affected and healthy side of the uMD group and that of the control group

The ELS rates in the whole inner ear (cochlea + vestibule + SCCs), cochlea, vestibule, and SCCs are demonstrated in Fig. 1. The ELS rates for uMD on the affected and healthy sides were calculated as follows: affected side: whole inner ear 26.2 ± 13.7% (mean ± SD), cochlea 23.7 ± 16.3%, vestibule 38.0 ± 21.5%, SCCs 21.2 ± 14.6%; healthy side: whole inner ear: 18.1 ± 14.2%, cochlea:15.7 ± 15.5%, vestibule: 20.6 ± 15.4%, SCCs: 18.7 ± 14.4%. In the control group, ELS rates were as follows: whole inner ear 13.7 ± 7.6%, cochlea: 10.2 ± 6.5%, vestibule: 17.3 ± 9.9%, SCCs: 14.7 ± 11.8. As in previous studies8, in the affected side of uMD, the ELS rates of the whole inner ear, cochlea, and vestibule were significantly higher than those of the healthy side (p < 0.01, two-way ANOVA, Tukey’s test) and control (p < 0.0001, two-way ANOVA, Tukey’s test). Furthermore, on the affected side of the uMD, the ELS rates of the vestibule were significantly higher than those of the whole inner ear, cochlea, vestibule, and SCCs.

Correlations between the atmospheric pressure change and the ELS rates

Next, the correlations between the 24-h atmospheric pressure change from the day before the i.e.MRI scanning day and ELS rates in patients with uMD are demonstrated in Table 2. The results of the healthy control group are presented in Supplement 1. There was no correlation between atmospheric pressure change and any of the ELS rates in the whole inner ear, cochlea, vestibule, or SCCs.

Correlation between hearing level and ELS rates

Therefore, based on the finding that the ELS rate of the vestibule was significantly higher than the ELS rate of other parts of the uMD-affected side, we underscored on the correlation between hearing thresholds and the vestibular ELS rate of the affected side of the uMD. The correlation between hearing thresholds and the vestibular ELS rate on the affected side of the uMD is demonstrated in Fig. 2 (r = 0.3275, p = 0.0008). A positive correlation was found between hearing thresholds and the vestibular ELS rate. The results for the other parts of the uMD-affected side are shown in Supplement 2.

Correlation between the 24-h atmospheric pressure and the vestibular ELS rate of the affected side of patients with uMD: classification by hearing level

Next, we grouped uMD by the severity of hearing level on the affected side as follows: Affected-side hearing level in uMD: Group 1 < 40 dB, 40 dB ≤ group 2 ≤ 70 dB, 71 dB < group 3. The correlations between the 24-h atmospheric pressure change values from the day before i.e.MRI scanning and the vestibular ELS rate in each group are shown in Fig. 3. The correlation between the atmospheric pressure change and the vestibular ELS rate of the affected side of the uMD was 0.005365 (p = 0.9741) in group 1, \(\:-\)0.4305 (p = 0.0020) in group 2 and 0.3494 (p = 0.2407) in group 3. The correlation between atmospheric pressure change and the vestibular ELS rate of the healthy side of the uMD was 0.2028 (p = 0.2157) in group 1, 0.2312 (p = 0.1100) in group 2, and 0.08253 (p = 0.7884) in group 3. A negative correlation was found between atmospheric pressure change and vestibular ELS rate on the affected side.

Discussion

To the best of our knowledge, this is the first study to show a significant association between atmospheric pressure changes and the ELS volume in patients with MD. A decline in ambient atmospheric pressure was suggested to increase the endolymphatic space volume on the affected side in patients with moderate uMD.

In 2007, Nakashima et al. reported that EH in patients with MD was visualized using MRI and enhanced by intratympanic administration of Gd22. Subsequently, a method for EH visualizing using 3T MRI 4 h after intravenous Gd injection was established17,18,23. Cho et al. validated the use of inner ear MRI to visualize EH in patients with MD by comparing it with histopathological changes in the inner ear. The results revealed that despite the limitations of MRI technology and image processing, the EH ratios measured by MRI were similar to those of the temporal bone specimens, indicating the reliability of inner ear MRI in diagnosing MD24. Ito et al. characterized EH in patients with MD via volumetric measurement of ELS using 3-T MRI and compared it to healthy controls. They found that the percentage of ELS in the affected ear of patients with MD was significantly higher than that in controls and in the healthy side of patients with MD8. The present study observed the same trend, especially in the vestibular region; the ELS rate was notably higher on the affected side of the uMD.

Previous studies have suggested a correlation between atmospheric pressure and vertigo. Schmidt et al. reported that lower atmospheric pressure is associated with a higher probability of vertigo attacks and higher levels of vertigo, tinnitus, and aural fullness in patients with MD. Additionally, higher humidity increases the probability of vertigo attacks6. Chen et al. reported that lower atmospheric pressure correlated with the onset of Meniere attacks and suggested that low atmospheric pressure in the summer may aggravate EH, especially when accompanied by typhoons in the northwest Pacific region4. Gürkov et al. reported a significant association between atmospheric pressure changes and the probability of MD episode5. However, there are no reports on the relationship between atmospheric pressure and endolymphatic space volume. Although the frequency of vertigo attacks relies on self-reported patient information and is not accurate7, endolymphatic space volume evaluated by i.e. MRI is a highly accurate objective assessment.

There was no significant correlation between atmospheric pressure changes and endolymphatic space volume changes in healthy controls or patients with MD cases as a whole. In healthy individuals, the pressure equalization mechanism between the inner and middle ear is thought to function properly making them less susceptible to barometric pressure changes25. This could be interpreted as the preservation of endolymphatic homeostasis in response to atmospheric pressure changes in healthy controls. Based on these findings, endolymphatic homeostasis can be maintained in mild cases of MD. Conversely, in severe cases, this homeostatic mechanism may be disrupted and no longer affected by changes in the surrounding environment26. Therefore, we divided patients with Meniere’s disease into three groups according to disease severity, examined the correlation between atmospheric pressure and volume changes, and found a significant negative correlation in the moderate group 2. A post hoc analysis was used to examine the effects of sample size and statistical power with regard to these results in Supplement 3. The statistical power values were very low except for the affected side of group 2, and sample size was not considered to influence the result of no significant correlation. On the affected side of group 2, statistical power value was 0.7403697, which was not an inadequate sample size. We considered the results to be confident enough to ensure a significant negative correlation. Although this negative correlation was only moderate, we consider it clinically relevant because it demonstrated a correlation between atmospheric pressure change and ELS, a quantitative indicator. This result could partially explain the mysterious phenomenon of vertigo onset occurring when weather conditions worsen, which has been reported for a long time4,5,6.

To support these results, the following points need to be considered. The endolymphatic sac is thought to be involved in endolymphatic fluid absorption, ion transport, and immune responses27. It is necessary to consider how changes in atmospheric pressure affect endolymphatic sac function. The stria vascularis plays an essential role in maintaining the composition of endolymph fluid28. The effect of atmospheric pressure changes on the function of stria vascularis should be considered. Future directions could include in vitro studies examining the pressure sensitivity of endolymphatic sac cells and in vivo studies measuring inner ear fluid pressure in response to controlled atmospheric pressure changes.

The relationship between MD and autonomic nerves is characterized by significant autonomic dysfunction, particularly involving sympathetic hyperactivity. Research indicates that patients with Ménière’s disease often exhibit altered autonomic nervous system (ANS) responses, especially before vertigo attacks, suggesting a strong link between stress, sympathetic activation, and the onset of symptoms. Ishis et al. reported that before a Ménière’s attack, there is marked sympathetic overactivation, with significant suppression of parasympathetic activity29. The ANS plays a critical role in regulating physiological responses to environmental changes, including atmospheric pressure fluctuations. Future research should investigate the role of ANS reactivity in mediating the relationship between atmospheric pressure changes and the onset of MD symptoms.

The present study has some limitations.

  1. 1.

    Owing to database limitations, we checked atmospheric pressure changes at the single point closest to where the patient underwent an MRI scan. MRI was performed while the patient was hospitalized for a week for neuro-otological examination, and the patient’s exposure to atmospheric pressure at the location where the MRI was performed was not transient. However, the weather conditions at the location where the patient stayed before admission for examination were not considered. Additionally, the JMA data for Nara City may not always accurately represent atmospheric pressure at the MRI site.

  2. 2.

    We were unable to capture endolymphatic space volume changes over multiple MRI scans of the same patient. The design of this study was to collect data on the relationship between the very fragmentary weather conditions around the day each patient underwent MRI and the endolymphatic space volume and to analyze the correlation.

  3. 3.

    The patients’ general condition, including their ANS, on the day of the MRI was not considered.

  4. 4.

    Although none of the patients had vertigo attacks during their hospitalization, the time elapsed since their last vertigo attack, frequency of attacks, and duration of attacks were not considered.

  5. 5.

    The relationship between vertigo attacks and the endolymphatic space volume was not considered in this study.

To address all of these issues, a prospective study design should be considered using a wearable device that can continuously measure the accurate atmospheric pressure, humidity, and other weather data for the location where the patient is located for the MRI scan. In addition, wearable devices that can continuously measure heart rate, blood pressure and other parameters would be effective in evaluating the relationship with autonomic nervous system function. Furthermore, to investigate the relationship with vertigo attacks, it would be effective to collect data around the day of MRI imaging using wearable videoculography.

Conclusions

This study suggests that there may be a correlation between changes in atmospheric pressure and vestibular volume in the affected ear in patients with MD.