Meniere's disease (MD), first reported in 1861 by Prosper Ménière, is associated with impaired hearing, tinnitus, vertigo, and aural fullness.1 The syndrome is caused by disturbances in the homeostasis of the inner ear, resulting in impaired cochlear and vestibular functions. Since the first reports of endolymphatic hydrops (EH) in the temporal bones of patients with MD, many anatomical studies, lacking identification of causal abnormalities, have suggested idiopathic EH as the pathological basis of MD.2–5 Various factors, including viral infections,6,7 autoimmune disorders8,9 and endocrine malfunctions,10 have been associated with EH development. However, some studies have suggested that EH is not a direct etiological factor in the development of MD and the emergence of MD symptoms.11,12 Investigations of the role of EH in MD are still hampered by a limited understanding of EH development in vivo.

The current gold standard for EH detection is autopsy analysis of temporal bones, which generally detects developed EH and fails to analyze the progression of EH. Nakashima et al.13 first visualized the endolymphatic space in living patients using Gadolinium (Gd)-enhanced 3-Dimensional Fluid-Attenuated Inversion Recovery (3D-FLAIR) magnetic resonance imaging (MRI) of the inner ear. Since then, 3D-FLAIR MRI and 3-Dimensional real Inversion Recovery (3D-real IR) MRI have been commonly used for EH analysis.14,15 Human and animal studies revealed that 3D-real IR imaging provides more clear and consistent histopathological observations than 3D-FLAIR imaging.15,16 However, 3D-FLAIR imaging outperforms 3D-real IR under conditions of insufficient Gd delivery into the perilymphatic space following an intratympanic injection.16

This study utilized Gd-enhanced inner ear MRI to investigate the development of EH in patients with MD. Specifically, 3D-FLAIR and 3D-real IR MRI scans were used to analyze the occurrence and degree of EH in the vestibule and each cochlear turn of the inner ears of MD patients.

MD is associated with recurrent episodic vertigo, tinnitus, aural fullness, and fluctuating hearing loss.19 It is a multifactorial disorder of the inner ear characterized by endolymphatic hydrops, or a distended endolymphatic space, which is detectable in temporal bones upon autopsy analysis. Animal models have been used to explore the mechanisms underlying MD and EH.20,21 However, evidence regarding the role of EH as a causative factor of MD or an epiphenomenon during MD progression has remained inconclusive.22 To our knowledge, the in vivo development of EH in MD patients remains poorly understood. Temporal bone autopsy, considered a gold standard for the detection of EH, cannot be used to monitor the pathogenetic process or to control for such variables as disease onset.

By enabling the visualization of the boundary between the endolymphatic and perilymphatic spaces, MRI scanning following intratympanic or intravenous Gd injection facilitates the detection of EH in living patients.23–25 In this study, we performed Gd-enhanced MRI to visualize occurrence and degree of EH in different parts of inner ear, and also related the stages of patients with the severity of EH to analyze and deduce the dynamic development of EH in population.

All 178 patients with definite unilateral MD tested in this study had EH in the vestibule or the cochlea of the affected symptomatic ear. Asymptomatic EH was also detected in the contralateral unaffected ears of 32 patients. Asymptomatic EH appeared more frequently in the cochleae than in the vestibules. Previous studies utilizing temporal bone autopsy or temporal bone Gd-enhanced MRI have similarly reported asymptomatic EH. Following temporal bone analysis in children, Bachor and Karmody26 reported low rates of saccule and utricle dilation but high incidence of bulging in the cochlear duct, with 65.6% occurring in the apical turn. Recent studies have also reported asymptomatic EH in the vestibules and cochleae of unaffected ears, though at rates different from those observed in our study.25,27 These studies have suggested that bilateral EH does not necessarily correlate with clinical signs of MD on both sides. Evidence of asymptomatic EH has been used to support the view that EH is a feature, rather than a cause, of MD.11,12 However, the detection of EH via temporal bone autopsy in 100% of patients with a history of MD supports the hypothesis that EH causes MD.22

Thus, we sought to investigate the relationship between EH and the symptoms of MD. Our results emphasized the differences between EH in the symptomatic ear and the contralateral asymptomatic ear. Although cochlear EH was prevalent on both sides, differences in EH rates between the cochlea and the vestibule were more substantial on the asymptomatic side than on the affected side. Asymptomatic EH was detected in 31.3% of the vestibules on the unaffected side, whereas 95.5% of the symptomatic side vestibules exhibited signs of EH. The degree of EH was also significantly different between the two sides. The degree of asymptomatic EH on the contra-ateral side was much milder than that of the symptomatic EH on the affected side. In particular, the difference in EH degree in the vestibules of asymptomatic ears and those of symptomatic early-stage MD ears underscored the role of vestibular EH in the clinical symptoms of MD. Based on the analysis above, presence of EH in any part of inner ear, especially in cochlear turns, was not necessary for onset of MD symptoms, so EH should be identified as the pathological basis of MD, instead of the causal and definitive factor. However, as the severity of EH in vestibule progressively worsened, potentially in conjunction with other factors, the onset of MD symptoms was more likely to occur.21,28 This result was much consistent with Yoshida T et al.,29 in 2018. Yoshida T reported that EH was present in the cochlea of 45/52 affected ears of patients with MD (87%) and in 16/42 control ears which were not associated with MD (38%). However, in the vestibule, symptomatic EH was present in 49/52 affected ears (94%) and asymptomatic EH in 3/42 control ears (7%). Consequently, Yoshida T et al., regarded EH in the vestibule as a more specific predictor of definite MD than EH in the cochlea. Similarly, the results in this manuscript would also emphasize the importance of EH in vestibule to the onset of MD symptoms, especially the severity of EH in vestibule.

As for the susceptibility of EH, our observations revealed that, on both sides, EH occurred more commonly and at a higher degree of severity in the cochlea, especially in the apical turn, than in the vestibule. This result is consistent with the temporal bone autopsy findings of Okuno and Sando, demonstrating that EH is more commonly observed in the pars inferior (cochlea and saccule) than in the pars superior (utricle and semicircular canals) of the temporal bone.30 As to the higher susceptibility of EH in cochlea, we could refer it to the distribution of dark cells in the inner ear. Kimura RS31 reported that dark cells in the vestibule are mainly localized in the ampullae, utricle and common crus, which was not detected in the saccule, saccular duct, ductus reuniens and main parts of the canals by the combined use of macroscopic observations, light microscopy and electron microscopy. Due to the characteristics of the distribution of vestibular dark cells, and the morphology and function of Bast's valve, the endolymph in he saccule should be generated most likely from the dark cells of the stria vascularis in cochlea. Thus, we emphasized the importance of the stria vascularis in cochlea in the pathogenesis of MD, which may produce redundant endolymph.

Multiple studies have provided evidence for the susceptibility of the apical turn to significant EH.32,33 Nageris et al. reported that varying degrees of displacement of the basilar membrane towards the scala tympani in the apical turn, which was not observed in other turns of the cochlea, may be related to atrophy of the spiral ligament.32 Valk et al. reported that, in the guinea pig, a rupture of Reissner's membrane in response to destructive acute endolymphatic hydrops most often occurs in the apical turn of the cochlea.33 As for the higher susceptibility of significant EH in apex in the cochlea, we are not sure that it is the cause of hearing loss in early MD, which typically starts with low-frequency tones, even though low-frequency tones are consistent with the location of the basilar membrane in apex turn.19

Based on the observed relationships between MD stage and EH distribution and degree, we propose that EH undergoes dynamic development in patients with MD. Considering that the incidence of asymptomatic EH was much lower in the vestibule than in the cochlea and was very similar between the cochlear turns, we propose that EH initially develops in the cochlea. During the asymptomatic stage, EH is likely mild in both the cochlea and the vestibule. With disease progression, the severity of EH gradually increases, especially in the vestibule, and the clinical symptoms of MD become apparent. Considering the prevalence of EH in the apical and middle turns of the cochlea and the high EH correlation observed for every two adjacent anatomical parts, we propose that EH originates in the CAT, with EH severity decreasing from the apical to the basal turn, and then from the cochlea to the vestibule.

Limitations should be noted in the interpretation of this study. Although this study demonstrated the occurrence and the distinct features of asymptomatic EH on the unaffected side of some MD patients, it has not assessed the incidence of asymptomatic EH in the general population and the potential progression of asymptomatic EH into MD.

Dynamic development of EH appears to be directional, likely being initiated in the cochlea. The degree of EH severity gradually decreases from the CAT to the vestibule, with significant EH more likely to appear in the CAT and the CMT. Moreover, occurrence of EH in the vestibule is significantly correlated with symptomatic MD.

National Natural Science Foundation of China, 81570917, 81600802, 81771009.

The authors declare no conflicts of interest.