Menière's disease (MD) is a refractory otologic disorder that predominantly manifests in adults between 40 and 60 years of age (1, 2). Its typical features are recurrent vertigo spells associated with fluctuating cochlear symptoms in the affected ear (3–6). The disease may affect one ear, albeit bilaterality has been reported in 2–73% of cases (7). MD is strongly associated with endolymphatic hydrops (EH): a distention of the endolymphatic space of the inner ear due to accumulation of endolymph fluid (8). Although a certain diagnosis of EH is reserved for post-mortem temporal bone histology, the Equilibrium committee of the AAO-HNS has provided clinical guidelines to aid the diagnosis of MD in living patients (5, 9).

The most striking features of MD are rotatory vertigo and hearing loss, which show great variability among patients with respect to onset, duration and severity (10, 11). Vertigo attacks can last from several minutes to hours, and are often incapacitating due to their unpredictability and impact on quality of life (12). In initial stages of the disease, hearing loss is typically reversible after each vertigo attack, but profound and permanent hearing loss ultimately develops (13). Additionally, patients may experience a myriad of symptoms consisting of (but not limited to) tinnitus, aural fullness, or a sudden fall without loss of consciousness known as drop attacks (14–16). Migraine and autoimmune diseases are common comorbidities, reported in up to 16 and 11% of patients respectively (17–21).

Despite their strong association, the relationship between EH and MD remains unclear. In the last decade, the application of delayed gadolinium (Gd)-enhanced MRI has provided novel insights in the clinical effect of EH (22–30). The degree of EH is reported to correspond with disease duration and the degree of sensorineural hearing loss in MD patients (27, 28, 31, 32). Contrarily, no correlation has emerged between EH and tinnitus, aural fullness, or the frequency and duration of vertigo attacks (23, 30). The absence of a clear correlation has led to the belief that other mechanisms beyond EH must play a role in the production of audiovestibular symptoms.

Recent literature has introduced novel MRI parameters that may help clarify the clinical-MRI inconsistency. Increased perilymph signal intensity (SI) at 4 h-delayed Gd-enhanced 3-dimensional fluid-attenuated inversion recovery (3D FLAIR) MRI is a frequent concomitant finding in patients with EH (32–36). The underlying pathophysiological process is assumed to be impairment of the blood-labyrinth barrier (BLB), leading to increased permeability and leakage of Gd from capillaries into the perilymphatic spaces (32, 37). The increased perilymphatic enhancement is well-visualized on 3D FLAIR, due to its high sensitivity to changes in T1-shortening induced by gadolinium-based contrast media (38). Previous studies investigating this FLAIR hyperintensity have reported that the concurrent finding of EH and BLB impairment slightly increases specificity for MD compared with EH alone (34, 35). The importance of this MRI parameter is further emphasized by the notion that BLB impairment may be the sole MRI correlate in a small percentage of MD patients without EH (35, 36). Furthermore, normalization of the FLAIR hyperintensity has been observed in coincidence with improved vestibulocochlear function in several inflammatory inner ear conditions (39–41). The above-mentioned findings suggest that BLB impairment may provide additional clinical information about HED and may potentially bring new insights to the individual severity of audiovestibular symptoms.

Extensive literature from the past two decades has demonstrated that virtually all patients with clinically overt MD show EH on MRI, for which different subjective EH grading systems have been used (34, 35, 42–44). However, probably the most important gain from hydrops MRI is the demonstration of EH in clinically atypical or non-classifiable cases according to the current AAO-HNS criteria (34, 45). These findings highlight the diagnostic relevance of MR imaging, particularly in the clinical setting where the current AAO-HNS criteria appear to be insufficient—e.g., subjects with isolated cochlear or vestibular symptoms. This concept has sparked aspirations to abandon the symptom-based classification and to adapt a diagnostic system that incorporates the MRI-objectivation of EH. In 2018, Gürkov et al. proposed the classification of Hydropic Ear Disease (HED): an imaging-based classification system that is based upon the demonstration of EH and that acknowledges the wide spectrum of audiovestibular symptoms associated with this pathological substrate (8, 46). However, the clinical utility of this system relies on the value of imaging-derived parameters in this clinical spectrum.

The objective of this study was therefore to investigate the correlation between clinical features and labyrinthine abnormalities in patients with MRI-proven hydropic ear disease, with emphasis on EH and BLB-impairment.

In the present study, we explored the relationship between cochleovestibular symptoms and intralabyrinthine abnormalities in a large cohort of patients with HED. We demonstrated the association between EH grade and hearing loss, as well as their progressive deterioration with respect to disease duration. Similar correlations have been shown by a couple of radiological studies (23, 27, 28, 30, 33) and a histological study on temporal bones from MD patients (47). Contrarily, previous radiological studies by Xie et al. (N = 117) and Fiorino et al. (N = 18) revealed no significant correlation between the severity of sensorineural hearing loss and EH grade, as revealed by MRI after intratympanic Gd-administration in definite MD patients (48, 49). Although the precise reason for this discrepancy is unknown, possible explanations may be different patient selection and sample sizes, or the use of a different MRI protocol and/or visual grading method for EH. Interestingly, our study revealed that there are cases with severe EH and hearing loss despite having a short disease duration and vice versa. These findings indicate that in some patients the hydropic process and hearing loss deteriorate rapidly, while others have a slow and gradual progression. Similar findings have previously been shown (50). Therefore, we hypothesize that factors other than duration are associated with the progression of EH. However, we did not identify factors associated with rapid EH progression. Further study is needed to explore these temporal patterns, which could have important implications in understanding the pathophysiology of HED.

Tinnitus and aural fullness have received less attention in literature, as emphasis is often placed on more prominent symptoms such as hearing loss and vertigo. However, it was previously reported that 19% of MD patients rank tinnitus as their most severe symptom and 38% considers aural fullness as a severe problem (51, 52). The identification of an in vivo correlate would provide important insights in the individually perceived severity of clinical disease. However, in contrast to hearing loss, our data did not demonstrate a correlation between EH grade and the presence of tinnitus or aural fullness. Notably, our study design did not cover qualitative measures of tinnitus or aural fullness, such as the perceived loudness or intensity.

The etiology of vertigo spells in HED also remains elusive. A myriad of theories have previously been proposed, including mechanical pressure or chemical imbalance/intoxication within the inner ear (53–56). Herein, we did not demonstrate a significant correlation between the extent of EH and the frequency or duration of vertigo attacks. It can therefore be concluded that merely the presence of EH does not elicit vertigo, which agrees with previous studies (23, 30). Schuknecht suggested that ruptures of the membranous labyrinth are the cause of vertigo spells, which would theoretically decrease the endolymph volume (57). However, a previous case report has revealed no large changes in EH on consecutive MRIs despite ongoing vertigo attacks (58). Of note, it was previously suggested that the fluctuations in EH may potentially be too small to be detected by MRI (23). Regarding the time course of vertigo attacks, our data contradicts previous suppositions that the number of attacks diminish over time (59). We did not observe a relationship between disease duration and the frequency or duration of vertigo attacks, which supports previous findings by Havia et al. and Jerin et al. (30, 60). It is worth noting that, although data on the presence of symptoms was well-documented in electronic patient files, we were unable to retrieve information on the duration of vestibular or cochlear symptoms in 21% of patients. Also, in hindsight, the categories we used to classify the duration of vertigo attacks were rather wide. The majority of patients fell within the 2nd category (vertigo attacks lasting 20 min to 12 h) and we did not account for further subanalysis.

In literature, there are conflicting views regarding the origin and evolution of EH. A previous meta-analysis of temporal bone reports from 184 MD patients demonstrated a cochleocentric distribution, where EH begins in the cochlear apex and from there progresses to involve the saccule, utricle and semicircular canals (61). The typical occurrence of low-frequency hearing loss in early stages of MD might be explained by these histopathological findings (57). On the contrary, in most of the radiologic literature, EH predominates in the vestibule (10, 62, 63). In our study, isolated cochlear EH was a scarce finding and less prevalent than isolated vestibular EH, which supports the hypothesis that EH originates in the vestibule. It was recently reported that a 3D real IR sequence is superior to 3D FLAIR in the assessment of cochlear EH due to its ability to distinguish endolymph from surrounding bone (64). Therefore, we cannot exclude that the lower prevalence of cochlear EH in our study cohort may be due to lower sensitivity of 3D FLAIR to cochlear EH. In addition, a golden standard for the assessment and quantification of EH is currently lacking. Although 3- and 4-point visual grading systems are often utilized to assess EH, and are subject to ongoing refinement, further developments are desired to enable objective volumetric quantification of EH.

Previous studies have addressed the development of bilateral EH in MD patients and reported variable prevalence rates ranging from 21 to 75%, which is thought to increase with disease duration (10, 23). In our study, bilateral EH was present in 36 (16.7%) of patients. We did not reveal significant differences in disease duration or age between patients with unilateral or bilateral EH and we were therefore unable to confirm the hypothesis. Nevertheless, the occurrence of bilaterality has led to the supposition that EH is a systemic disease and associations with genetic factors and several comorbidities have been reported in literature, particularly migraine and autoimmune diseases (21). In our study, migraine and autoimmune diseases were present in 11.2 and 6.5% of cases, respectively. Among the cardiovascular risk factors, hypertension was the most common comorbidity and reported in 17.2% of patients (17.8% in men; 16.7% in women). In comparison, the prevalence of arterial hypertension in the general Dutch population was estimated at 29.9% in men and 15.6% in women (65). Additionally, the prevalence of migraine in Europe and a broad group of autoimmune diseases in Denmark were estimated at 37.6 and 7.6–9.4%, respectively (66, 67). Therefore, an association between EH and autoimmune diseases, migraine or hypertension cannot be substantiated from our data.

The association between EH and BLB impairment at 4 h-delayed Gd-enhanced MRI has previously been demonstrated (32–36). In the present study, visually increased perilymphatic enhancement was present in 157 (87.7%) of unilateral HED patients. Furthermore, the degree of BLB impairment was significantly associated with EH grade. The BLB is indicated as the essential structure for the maintenance of inner ear fluid homeostasis by selectively regulating the passage of ions between the vasculature and inner ear (68). Based on this knowledge, it has been postulated that destruction of the labyrinthine barriers causes the formation of EH (33). However, previous studies have demonstrated that increased BLB permeability may also be present in non-hydropic ears from patients with sudden sensorineural hearing loss (SSNHL) or vestibular neuritis (36, 69). Therefore, it can also be deduced that BLB impairment does not necessarily lead to the formation of EH. Regardless of their actual pathophysiological processes, both EH and BLB impairment are likely markers of disordered labyrinthine homeostasis. Interestingly, it has been demonstrated on MRI that patients with SSNHL may show improvement of symptoms in parallel with resolution of the FLAIR hyperintensity (40). This close correspondence between MRI findings and clinical features in SSNHL patients raises the question whether this might also apply to patients with EH, though the underlying pathophysiological mechanisms are likely different. However, we evaluated the isolated effect of SIR on the degree of sensorineural hearing loss and did not demonstrate a significant correlation between SIR and low or high Fletcher indexes. Thus, a putative association between SIR and hearing loss in HED patients cannot be supported from this data. Our results are in agreement with Suzuki et al., who used the SIR of the basal cochlear turn against the cerebellar hemisphere as marker of BLB permeability and found no significant correlation with hearing thresholds (70). On the contrary, Kahn et al. did report a significant correlation between BLB impairment and worse hearing levels (71). In their study, however, they performed a visual assessment of BLB integrity and did not control for the potential confounding effect of EH. In addition to hearing loss, we also did not find a significant correlation between SIR and tinnitus, aural fullness, or the frequency and duration of vertigo attacks. It is worth noting that our retrospective study design inherently did not acknowledge certain clinical factors that could have potentially influenced the MRI results, such as the time elapsed between the MRI examination and the last vertigo attack or intratympanic corticosteroid injections. Longitudinal radiological and clinical assessment could offer a better characterization of the dynamic alteration of MRI parameters in relation to treatment strategies and symptomatology.

Drop attacks are known to occur in a subset of MD patients, although the reported incidence is variable with rates ranging from 13.5 to 72%, depending on the definition used (16, 24, 72). In our study, drop attacks were present in 7.4% of unilateral HED patients. We found that drop attacks were associated with a longer disease course and higher grades of EH compared to patients without DA's. The DA group also demonstrated greater hearing loss than non-DA patients, which may be attributed to the longer disease duration and greater degree of EH. Similar observations were previously made by Wu et al. (16). Contrarily, the median SIR was not significantly different between patients with and without DA. Although the pathophysiological mechanisms of drop attacks are unknown, mechanical stimulation of the otolithic organs (saccule and/or utricle) by endolymphatic pressure gradients is currently the most widely accepted etiology (73, 74). The finding that DA occurs in patients with a greater degree of EH seems to support this theory.

The rationale behind MR imaging of EH is based on a desire to seek improved diagnostic precision and better understand how labyrinthine abnormalities affect the function of the inner ear. In addition, the possibility that imaging-derived data could serve as a surrogate marker of disease progression and therapeutic efficacy has become attractive. A critical aspect when considering the use of an MRI as a surrogate parameter is validation: the degree of EH and/or BLB impairment must show a correlation with the presence and severity of audiovestibular symptoms. Investigating the clinical picture from a morphologic point of view (i.e., Hydropic Ear Disease) allows assessment of this pathologic stage across the entire clinical spectrum. Although some correlations have been found, the degree of EH and BLB impairment incompletely characterize the whole phenotype of HED, especially regarding the incidence and severity of vertigo attacks. If a more explicit radiologic-clinical correlation can be established, then this information could potentially be used to better predict disease progression, monitor treatment response, and aid the development of novel treatment strategies.

In this retrospective study, we have investigated the clinical picture of patients with MRI-evidence of hydropic ear disease with or without BLB impairment. In our study population, the degree of EH correlates with disease duration, the severity of sensorineural hearing loss and the incidence of drop attacks. BLB is impaired in association with EH grade, but without clear contribution to the severity of audiovestibular symptoms.

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author/s.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

LP, JS, and SH contributed to the conception and design of the manuscript. LP, SH, TV, MH, HB, and CB acquired the data. JG helped shape and verify the statistical methods. LP performed the statistical analysis and drafted the manuscript. All authors contributed to the article and approved the submitted version.

Related to the research being submitted: SH: Radiology Research Fund from the Radiological Society of the Netherlands, to assess the technical feasibility of MRI-evaluated endolymphatic hydrops and allow development and optimization of imaging techniques. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication. Not related to the research being submitted: BV: grants from Oticon and Advanced Bionics paid to institution.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.