Cerebrovascular Burden and Its Association with Ménière’s Disease: A Case-Control Study

Abstract

Background: Ménière’s disease (MD) lacks a universally accepted pathogenesis model. Recent research has revisited the vascular hypothesis. This study aims to compare the cerebrovascular burden in patients with MD and age-matched controls, investigating the potential role of cerebrovascular dysfunction in MD. Methods: A total of 145 patients (70 MD, 75 controls) underwent magnetic resonance imaging (MRI) assessment for small-vessel disease (SVD) markers (including Fazekas and EPVS scores), cortical strokes, and baseline comorbidities. Statistical analyses were performed to compare the cerebrovascular burden between the groups, adjusting for potential confounders. Results: The MD group exhibited significantly higher mean SVD scores across various measures compared to controls (p < 0.05). This association persisted even after adjusting for age, sex, and comorbidities (ORs ranging from 1.746 to 2.495, p < 0.05). Neither the presence of cortical strokes nor comorbidities significantly differed between groups. Conclusions: This study is the first to compare cerebrovascular burden between MD patients and controls. The findings suggest that cerebrovascular dysfunction may contribute to MD incidence. Further research is needed to elucidate the relationship between cerebrovascular disease and MD, potentially leading to novel therapeutic avenues.

1. Introduction

Ménière’s disease (MD) is a multifactorial inner ear disorder characterized by episodic vestibular symptoms (e.g., vertigo), sensorineural hearing loss, tinnitus, and aural pressure [1,2]. Although considered uncommon globally [3], incidence rates vary widely [4], with higher prevalence among Caucasians [5] and increasing frequency with age [3]. A slight female preponderance exists [6], and bilateral MD occurs in up to 40% of cases [7]. Despite the increasing use of magnetic resonance imaging (MRI) and computed tomography (CT) scans [8], MD diagnosis is primarily clinical [9], following criteria established by international consensus [10].

Despite over a century of recognition, MD’s etiology and pathophysiology remain elusive. Current understanding posits MD as a complex condition, with genetic and environmental factors influencing its development and phenotypic diversity [4]. Endolymphatic hydrops (EHs) is a hallmark of MD, but not all EH patients experience symptoms [11,12].

No single theory fully explains MD’s pathogenesis, epidemiology, and symptom manifestation. However, recent epidemiological studies reveal a significantly higher prevalence of cardiovascular risk factors in MD patients, including hyperlipidemia, hypertension, diabetes, and smoking [13,14,15,16]. These factors may also correlate with increased vertigo attack frequency [15].

MD attacks are thought to be triggered by abnormal inner ear pressure and ion/fluid imbalances due to Reissner’s membrane rupture [17]. While the exact mechanism remains unclear, overexcitation of the sympathetic nervous system has been proposed, leading to vascular tone imbalance, increased permeability, and subsequent EH [18]. The role of small-vessel disease in MD, however, is less understood.

Based on these observations, a vascular hypothesis for MD has emerged [12,13,15,19], suggesting that small-vessel disease may disrupt inner ear homeostasis and contribute to EH. Some researchers even consider MD part of a broader cerebrovascular syndrome [12,15]. However, no study has directly compared end-organ damage between MD patients and controls.

This study aims to address this gap by comparing signs of vascular damage on cerebral imaging between MD patients and age-matched controls. We also analyze baseline vascular risk factors in both groups to assess their potential contribution to MD pathogenesis.

2. Materials and Methods

2.1. Study Design, Settings, and Participants

This retrospective case-control study was conducted as a joint initiative between the Otorhinolaryngology and Neuroradiology Departments at Unidade Local de Saúde de Santo António, Porto, Portugal. The study design adhered to the Helsinki Declaration’s ethical principles and was approved by the local ethics committee. Figure 1 depicts the overall methodological approach.

Case Selection: Cases of definite MD were selected from the ORL Department’s Otoneurology consultation records between 2016 and 2023. Inclusion criteria for cases included the following: diagnosis of definite MD according to the classification criteria [10]; age ≥ 18 years; available MRI on clinical records; absence of other explanatory lesions on MRI.

Control Selection: Age and sex-matched controls were selected from an existing pituitary dataset from 2022 and 2023, excluding post-surgical and post-radiotherapy MRIs. Patients with pituitary pathology were chosen as controls due to their comprehensive MRI protocols, lack of intrinsic association with cerebrovascular or inner ear disease, and the ability to age and sex match with the MD group. This selection minimized potential confounding factors and allowed for a more accurate assessment of the cerebrovascular burden in MD.

2.2. Instruments, Variables, and Data Collection

Brain MRIs were used to assess and compare the cerebrovascular burden between the MD and control groups. All MRIs were blindly evaluated by an experienced neuroradiologist for the presence, location, size, and type of vascular lesions. The assessment included:

• Lacunar infarcts: Spherical or oval lesions >3 mm and <20 mm, located in basal ganglia, internal capsule, centrum semiovale, or brainstem, with specific MRI characteristics [20].
• White matter hyperintensities (WMH): Classified using the Fazekas scale (0–3) [21].
• Enlarged perivascular spaces (EPVS): Small (<3 mm) CSF-filled spaces, scored on a 0–4 scale for basal ganglia and centrum semiovale [22]. The values correspond to EPVS on a single hemisphere of the brain. In cases of asymmetry between hemispheres, EPVS were recorded in the slice with the maximum count, so that the higher score was accounted. EPVS total score for each patient was obtained by adding the scores from both regions (EPVS basal ganglia and EPVS semiovale), with a range of 0 to 8.
• Small-vessel disease (SVD) burden: Assessed using a modified 3-item SVD score (SVD-3), excluding cerebral microbleeds (CMBs) due to limited availability of gradient echo T2* sequences in our sample [23,24,25]. The SVD-3 score was calculated based on the presence of lacunes, EPVS, and WMH (Fazekas score) (

  Table 1. Small-vessel disease scoring (SVD-3) and large-vessel disease features explored in this work.

  Type of Cerebrovascular Disease	MRI Feature	Scoring	Assessment	MRI Example
  Small-vessel disease (SVD) [SVD-3 score]	Lacunes (International definition) [20]	SVD-3	≥1 scores 1 in SVD-3
  		SVD-3 low-Fazekas
  		SVD-3 low-EPVS
  		SVD-3 low Fazakas + low-EPVS		Right basal-ganglia lacunes
  	Enlarged perivascular spaces (semiquantitative scale) [22]	SVD-3	Moderate to severe (more than 10 EPVS in worst side) score 1 in SVD-3
  		SVD-3 low-Fazekas
  		SVD-3 low-EPVS	Mild (more than 1 EPVS in worst side) score 1 in SVD-3 low-EPVS
  		SVD-3 low Fazakas + low-EPVS		Right basal ganglia EPVS
  	White matter hyperintensities (WMH) (Fazakas scale) [21]	SVD-3	Fazekas ≥ 2 scores 1 in SVD-3
  		SVD-3 low-Fazekas
  		SVD-3 low-EPVS	Fazekas ≥ 1 scores 1 in SVD-3 low-Fazekas
  		SVD-3 low Fazakas + low-EPVS		Fazakas score = 3
  Larger vessel disease	Cortical stroke		≥ 1 in any location coded as “present”
  				Left middle cerebral artery territory infarct

  ). The SVD-3 scoring is as follows: 1 point for the presence of lacunes, indicated by the existence of ≥1 lacune(s); 1 point for the presence of EPVS grade 2–4 (moderate to severe), found either in the basal ganglia or semi-oval center; and 1 point for the presence of WMH defined by a Fazekas score ≥ 2. Additionally, for further exploring the statistical validity of results, three additional subscales of SVD-3 with lower cutoffs were created in order to check for differences in sensitivity of SVD burden scoring by lowering the cutoff for Fazekas (SVD-3 low-Fazekas), EPVS (SVD-3 low-EPVS) and both Fazekas and EPVS (SVD-3 low-Fazekas + low-EPVS) (Table 1).
• Cortical strokes: Larger infarcts involving cortical or subcortical tissue or large striatocapsular/subcortical lesions.

Figure 1. Methodological approach flow chart [21,22].

Figure 1. Methodological approach flow chart [21,22].

Additional variables from clinical records included the following: age at MRI; sex; comorbidities: epilepsy, neurocognitive affection, thyroid disease, chronic headache, obstructive sleep apnea, autoimmune disease, diabetes mellitus, hypertension, dyslipidemia, pulmonary disease, cardiac disease, smoking, obesity, and former chemotherapy.

2.3. MRI Specifications

The examinations were performed using either a Philips^® Achieva 3T TX (Eindhoven, The Netherlands) or a General Eletric^® Signa Explorer 1.5T MRI scanner. T2-weighted sequences (Boston, MA, USA), 3D T2-weighted fluid-attenuated inversion recovery (FLAIR) or axial T2 FLAIR, diffusion-weighted (DWI), gradient echo (in some cases), and T1-weighted sequences were used. SECTRA^® software (https://sectra.com/, accessed on 1 August 2024) was used to display the scans after multiplanar reconstructions.

2.4. Statistical Methods

Statistical analyses were performed using SPSS (IBM SPSS Statistics 29). Descriptive statistics were used to summarize the data, with appropriate tests for normality. Between-group comparisons were conducted using independent t-tests or Mann–Whitney tests. Binary logistic regression was used to adjust for potential confounders and determine the risk of MD associated with cerebrovascular burden. To address the risk of Type I error due to multiple comparisons, the Bonferroni correction was applied to adjust p-values where appropriate. Statistical significance was set at p ≤ 0.05. Statistical significance was set at p ≤ 0.05.

3. Results

3.1. Study Population

A total of 145 patients were included, 70 cases (MD) and 75 controls. 60 were males (41.4%) and 85 females (58.6%), with a mean age at MRI of 56.2 ± 13.6 years (minimum 29 and maximum 84 years). MRI T2 FLAIR sequence was tridimensional (3D) in 125 patients (86.2%) and axial in 20 (13.8%). 10 patients had cortical infarction sequalae on MRI (6.9%). Median overall Fazekas score from the sample was 1 (0–2). 27 patients had lacunes on MRI (18.6%). Microbleeds were reported in 4 out of the 38 patients were T2* was performed (10.5%). Median EPVS basal ganglia was 1 (0–2). Median EPVS semiovale was 0 (0–1). Median EPVS total was 0 (0–2). The overall incidence of comorbidities were as follows: epilepsy: 5 patients (3.4%); neurocognitive affection (including depression): 38 patients (26.2%); Thyroid disease: 21 patients (14.5%); chronic headache: 19 patients (13.1%); obstructive sleep apnea (OSA): 8 patients (5.5%); autoimmune disease: 10 patients (6.9%); diabetes mellitus: 19 patients (13.1%); hypertension: 51 patients (35.2%); dyslipidemia: 45 patients (31%); pulmonary disease: 7 patients (4.5%); cardiac disease: 10 patients (6.9%); smoking: 7 patients (4.8%). Obesity: 12 patients (8.3%); former chemotherapy: 2 patients (1.4%); mean SVD-3 score was 0.42 ± 0.7; mean SVD score low-Fazekas was 0.9 ± 0.7; mean SVD-3 score low-EPVS was 1 ± 0.9; mean SVD-3 score low-Fazekas + low-EPVS was 1.5 ± 0.9.

3.1.1. Cases (Ménière’s Disease)

A total of 70 MD patients were included. Mean time from first reported symptoms of MD to MRI was 8 ± 4 years. 5% of the included cases had bilateral MD. Mean time from first symptoms to first observation was 6.3 ± 5.5 years. Mean age of MD case group was 57 ± 13 years. Number of patients integrating each of the age categories is also described on

Table 2. Descriptive and bivariate analysis of case (Ménière’s) and control groups.

Continuous Variables	Mean $(\pm$Standard Deviation)		p-Value	Categorical Variables	Frequency (%)		p-Value
	Ménière	Controls			Ménière	Controls
Age (years) 1	$57\pm 13$	$55\pm 14$	$0.485$	Age (categories)
				<45 years	$18.6$	$30.7$	$0.092$
				45–55 years	$31.4$	$18.7$	$0.075$
				55–65 years	$20$	$22.7$	$0.696$
				65–75 years	$18.6$	$17.3$	$0.846$
				>75 years	$11.4$	$10.7$	$0.884$
Fazekas score	$1.1\pm 0.8$	$0.75\pm 0.8$	$0.013$	Gender (male)	$41.4$	$41.3$	$0.991$
				T2 FLAIR (3D)	$77.1$	$94.7$	$0.020$
				Lacunes	$24.3$	$13.3$	$0.090$
				Infarction	$10$	$4$	$0.154$
EPVS basal ganglia	$0.83 \pm 0.6$	$0.45\pm 0.5$	<0.001	Comorbidities
EPVS semiovale	$0.37 \pm 0.5$	$0.25 \pm 0.4$	$0.162$	Diabetes mellitus	$15.7$	$10.7$	$0.368$
EPVS total	$1.20 \pm 0.9$	$0.71 \pm 0.8$	$0.001$
SVD-3	$0.56\pm 0.9$	$0.29 \pm 0.6$	$0.041$	Hypertension	$35.7$	$34.7$	$0.895$
SVD-3 low-fazekas	$1.10\pm 0.8$	$0.71\pm 0.7$	$0.002$	Dyslipidemia	$35.7$	$26.7$	$0.239$
SVD-3 low-EPVS	$1.21 \pm 0.9$	$0.76\pm 0.8$	$0.002$	Smoking	$5.7$	$4$	$0.630$
SVD-3 low-fazekas + low-EPVS	$1.79 \pm 0.9$	$1.21 \pm 0.9$	<0.001	Obesity	$10$	$6.7$	$0.467$
				Obstructive sleep apnea	$5.7$	$5.3$	$0.920$
				Pulmonary disease *	$4.3$	$5.3$	$0.769$
				Cardiac disease	$7.1$	$6.7$	$0.910$
				Epilepsy	$5.7$	$1.3$	$0.149$
				Neurocognitive disease †	$25.7$	$26.7$	$0.896$
				Thyroid disease	$14.3$	$14.7$	$0.948$
				Previous chemotherapy	$1.4$	$1.3$	$0.961$
				Chronic headache	$14.3$	$12$	$0.684$
				Autoimmune disease	$4.3$	$9.3$	$0.231$
				OVERALL
				(≥2 of the listed comorbidities)	$51.4$	$38.7$	$0.123$
				(≥3 of the listed comorbidities)	$31.4$	$24$	$0.317$

¹ age was considered at date of MRI; EPVS: enlarged perivascular spaces score (EPVS); SVD-3: small-vessel disease (SVD) score; FLAIR: fluid-attenuated inversion recovery sequence; * includes asthma; † includes depression; p-value refers to results from bivariate analysis comparison between case and control groups, utilizing the independent t-test for continuous data and the Pearson chi-square for categorical variables.

. No significant differences were found concerning age, comorbidities, or gender between cases and controls in each of the age categories (see Table 2).

3.1.2. Controls

A total of 75 control patients were included. Of those, thirty-one (41.3%) had pituitary macroadenoma; twenty-one (28%) had pituitary microadenoma; ten (13.3%) had Rathke cleft cyst; seven (9.3%) had no pathology (either false incidentalomes from tomography scans or screening due to hormonal imbalances); three (4%) had Tuberculum sellae meningioma; two had empty sella (2.7%); and 1 (1.3%) had Hypophysitis.

3.2. Ménière’s Versus Controls

Table 2 depicts relevant bivariate comparisons between case and control subpopulations. There were no statistically significant differences between MD and controls regarding age or gender distribution, any specific comorbidity or number of comorbidities across subgroups (Table 2). There was also a significant higher percentage of 3D FLAIR sequence done in the control group (94.7% vs. 77.1% in MD, p = 0.02) (see Table 2).

Regarding measurements of cerebrovascular disease, there were numerous statistically significant differences between MD and controls, namely the following: higher mean Fazekas score in MD (1.1 ± 0.8 in MD vs. 0.75 ± 0.8 in controls, p = 0.013); higher EPVStotal in MD (1.20 ± 0.9 in MD vs. 0.71 ± 0.8 in controls, p = 0.001);

Higher mean SVD-3 scoring in MD (p < 0.05 irrespective of the cut-off used). Lower cutoffs for scoring WMH (Fazekas) and EPVS associated with more significant differences between MD and controls (see Table 2). Figure 2 depicts the relationship map of the SVD-3 measurement with stronger association (SVD-3 low-Fazekas + low-EPVS).

To check whether this associations persisted in patients with similar age, an additional analysis comparing groups was performed within the same age categories. Results showed specific significant differences between groups under 45, between 45 and 55 years and over 75 years old (

Table 3. Descriptive and bivariate analysis of case (Meniére’s) and control groups across age categories.

Age Category (Years)
Variable	<45			45–55			55–65			65–75			>75
	MD	Control	p-Value	MD	Control	p-Value	MD	Control	p-Value	MD	Control	p-Value	MD	Control	p-Value
Gender (male)	$38.5\%$	$39.1\%$	$0.968$	$31.8\%$	$42.9\%$	$0.501$	$50\%$	$41.2\%$	$0.623$	$46.2\%$	$38.5\%$	$0.691$	$50\%$	$50\%$	$1$
Lacunes	$0\%$	$0\%$	NC	$0\%$	$0\%$	NC	$21.4\%$	$29.4\%$	$0.613$	$61.5\%$	$30.8\%$	$0.116$	$75\%$	$12.5\%$	$0.012$
Infarction	$7.7\%$	$0\%$	$0.177$	$0\%$	$0\%$	NC	$7.1\%$	$5.9\%$	$0.887$	$23.1\%$	$7.7\%$	$0.277$	$25\%$	$12.5\%$	$0.522$
(≥2 of the listed comorbidities)	$30.8\%$	$4.3\%$	$0.028$	$40.9\%$	$28.6\mathrm{\%}$	$0.452$	$50\%$	$58.8\%$	$0.623$	$69.2\%$	$61.5\%$	$0.680$	$87.5\%$	$75\%$	$0.522$
(≥3 of the listed comorbidities)	$23.1\%$	$4.3\%$	$0.086$	$22.7\%$	$14.3\%$	$0.533$	$35.7\%$	$41.2\%$	$0.756$	$38.5\%$	$30.8\%$	$0.680$	$50\%$	$50\%$	$1$
Fazekas score	$0.62 \pm 0.5$	$0.26 \pm 0.4$	$0.047$	$0.77\pm 0.7$	$0.57 \pm 0.5$	$0.347$	$1.1 \pm 0.6$	$1 \pm 0.9$	$0.982$	$1.62 \pm 0.9$	$1.08\pm 0.6$	$0.086$	$2 \pm 0.8$	$1.38 \pm 0.9$	$0.160$
EPVS total	$0.92 \pm 0.9$	$0.26 \pm 0.5$	$0.008$	$1.1 \pm 1$	$0.43 \pm 0.6$	$0.031$	$1.43 \pm 0.9$	$0.9 \pm 0.9$	$0.154$	$1.08 \pm 0.6$	$1.4 \pm 0.9$	$0.316$	$1.87 \pm 1.2$	$0.87 \pm 0.8$	$0.083$
SVD-3	$0$	$0$	NC	$0.14 \pm 0.5$	$0$	$0.285$	$0.5 \pm 0.8$	$0.59 \pm 0.9$	$0.779$	$1 \pm 0.8$	$0.61 \pm 0.6$	$0.197$	$2 \pm 1.2$	$0.5 \pm 0.5$	$0.006$
SVD-3 low-fazekas	$0.62 \pm 0.5$	$0.26 \pm 0.4$	$0.047$	$0.68 \pm 0.6$	$0.57\pm 0.5$	$0.551$	$1.14 \pm 0.8$	$0.9 \pm 0.8$	$0.489$	$1.61 \pm 0.5$	$1.15 \pm 0.8$	$0.094$	$2.12 \pm 1$	$1\pm 0$	$0.006$
SVD-3 low-EPVS	$0.62 \pm 0.5$	$0.26 \pm 0.4$	$0.047$	$0.77 \pm 0.6$	$0.36 \pm 0.5$	$0.033$	$1.21 \pm 0.8$	$1.12\pm 0.9$	$0.758$	$1.85 \pm 1.1$	$1.39 \pm 0.6$	$0.198$	$2.38 \pm 0.9$	$1.12\pm 0.8$	$0.013$
SVD-3 low-fazekas + low-EPVS	$1.23 \pm 0.9$	$0.48 \pm 0.5$	$0.003$	$1.32\pm 0.7$	$0.93\pm 0.8$	$0.160$	$1.9 \pm 0.8$	$1.6 \pm 1$	$0.310$	$2.46 \pm 0.8$	$2.08 \pm 0.8$	$0.214$	$2.62 \pm 0.5$	$1.62 \pm 0.5$	$0.002$

MD—Ménière disease; NC—not computed (statistics not possible since there is a constant in both groups—0 zero).

).

3.3. Cerebrovascular Burden and Ménière’s Risk: A Model

To assess the potential impact of elevated cerebrovascular burden (SVD-3) on the incidence of MD, a supplementary analysis was conducted employing a binary logistic regression (

Table 4. Results for logistic regression analysis for outcome “Ménière” against baseline predictor variables (age, sex, and comorbidities) and SVD-3 score.

Independent Variable Predicting “Ménière”	β 1	SE β 2	OR (Exp β) 3	[95% CI] for OR	p-Value
Age	$-0.017$	$0.017$	$0.983$	$0.952–1.015$	0.298
Sex	$-0.046$	$0.348$	$0.955$	$0.483–1.888$	0.894
≥2 of the listed comorbidities	$0.584$	$0.386$	$1.792$	$0.841–3.821$	0.131
SVD-3 score	$0.557$	$0.278$	$1.746$	$1.012–3.012$	0.045

OR—odds ratio; [95% CI]—lower and upper bound of 95% confidence interval; ¹ β stands for unstandardized regression coefficient; ² standard error for unstandardized regression coefficient; ³ “Exp(β)”, or the odds ratio, is the predicted change in odds for a unit increase in the predictor. The “exp” refers to the exponential value of β.

). Results indicated a significant association between MD and SVD-3: the odds of MD increased by 79% (95% CI (1.012,3.012)) for a one unit increase in SVD-3 score. The logit (Li) = B0 + B1X1 + B2X2 + B3X3 + B4X4 formula obtained was Li = 0.431 − (0.017 × age in years) − 0.046 × sex (1 if male or 0 if female) + 0.584 × comorbidities (1 if ≥2 or 0 if <2) + 0.557 × SVD-3 score.

In order to bring similar formulae for SVD-3 altered cut-offs, an additional analysis using the same logistic regression method was employed (

Table 5. Results for multiple logistic regression analysis for outcome “Ménière” against baseline predictor variables (age, sex, and comorbidities) and SVD-3 scores with altered cut-offs.

Independent Variable Predicting “Ménière”	β 1	SE β 2	OR (Exp β) 3	[95% CI] for OR	p-Value
Age	$-0.028$	$0.017$	$0.972$	$0.940-1.005$	$0.094$
Sex	$-0.22$	$0.363$	$0.803$	$0.394-1.636$	$0.545$
≥2 of the listed comorbidities	$0.592$	$0.393$	$1.807$	$0.836-3.906$	$0.132$
SVD-3 low-Fazekas	$0.914$	$0.293$	$2.495$	$1.406-4.430$	$0.002$
Independent variable predicting “Ménière”	β 1	SE β 2	OR (Exp β) 3	[95% CI] for OR	p-Value
Age	$-0.032$	$0.018$	$0.968$	$0.935-1.003$	$0.07$
Sex	$-0.111$	$0.358$	$0.895$	$0.443-1.806$	$0.757$
≥2 of the listed comorbidities	$0.728$	$0.402$	$2.071$	$0.942-4.554$	$0.070$
SVD-3 low-EPVS	$0.804$	$0.249$	$2.234$	$1.371-3.641$	$0.001$
Independent variable predicting “Ménière”	β 1	SE β 2	OR (Exp β) 3	[95% CI] for OR	p-Value
Age	$-0.038$	$0.018$	$0.963$	$0.963-0.997$	$0.033$
Sex	$-0.247$	$0.369$	$0.781$	$0.379-1.609$	$0.503$
≥2 of the listed comorbidities	$0.694$	$0.404$	$2.001$	$0.907-4.416$	$0.086$
SVD-3 low-Fazekas + low-EPVS	$0.887$	0.234	$2.428$	$1.536-3.840$	$<0.001$

OR—odds ratio; [95% CI]—lower and upper bound of 95% confidence interval; ¹ β stands for unstandardized regression coefficient; ² standard error for unstandardized regression coefficient; ³ “Exp(β)”, or the odds ratio, is the predicted change in odds for a unit increase in the predictor. The “exp” refers to the exponential value of β.

). Like in previous bivariate analyses, significance increased with lower cutoff setup for SVD-3 (lower Fazekas and/or EPVS). The odds of MD increased by 149% (95% CI (1.406, 4.430)) for a one unit increase in SVD-3 low-Fazekas score. The formula obtained from regression was Li = 0.573 − (0.028 × age in years) − 0.22 × sex (1 if male or 0 if female) + 0.592 × comorbidities (1 if ≥2 or 0 if <2) + 0.914 × SVD-3 score. The odds of MD increased by 123% (95% CI (1.371, 3.641)) for a one unit increase in SVD-3 low-EPVS score. The formula obtained from regression was Li = 0.688 − (0.032 × age in years) − 0.111 × sex (1 if male or 0 if female) + 0.728 × comorbidities (1 if ≥2 or 0 if <2) + 0.804 × SVD-3 score. The odds of MD increased by 143% (95% CI (1.536, 3.840)) for a one unit increase in SVD-3 low-Fazekas + low-EPVS score. The formula obtained from regression was Li = 0.559 − (0.038 × age in years) − 0.247 × sex (1 if male or 0 if female) + 0.694 × comorbidities (1 if ≥2 or 0 if <2) + 0.887 × SVD-3 score.

This model results show a positive association between SVD-3 and MD after adjustment for potential confounders (age, sex, and comorbidities) and irrespective of the Fazekas and EPVS cut-off used to score SVD-3.

4. Discussion

MD remains an enigmatic pathology, with its underlying pathophysiology still unclear despite years of research [26]. This lack of understanding poses diagnostic challenges and contributes to significant patient morbidity due to difficulties in early detection [26]. While EH is consistently observed in MD [15], it is not sufficient to explain the entire disease process.

Our study sought to address this knowledge gap by investigating the “brain–ear” hypothesis, which posits a link between MD and cerebrovascular disease [12]. Previous research has highlighted a potential association between MD and cerebral ischemia-related conditions, such as migraine [12], as well as a higher prevalence of cardiovascular risk factors in MD patients [13,14,15,16]. However, no prior study has directly compared end-organ vascular damage between MD patients and controls.

To explore this, we compared signs of vascular damage on cerebral imaging between MD patients and age-matched controls. Our results demonstrate a significant association between MD and increased cerebral small-vessel disease (SVD), as measured by the SVD-3 score. This association persisted even after adjusting for potential confounders like age and sex, reinforcing the potential role of vascular dysfunction in MD pathogenesis.

The findings of our study support the notion that MD may be part of a broader cerebrovascular syndrome, as proposed by Foster [12,13] and Rego et al. [15]. The observed link between cerebral SVD and MD may be mediated by shared risk factors and pathophysiological mechanisms [27]. Cardiovascular risk factors are known to compromise microvasculature, induce oxidative stress, and damage the blood–brain and blood–labyrinth barriers, potentially affecting both the brain and the inner ear [28,29].

Intriguingly, while comorbidities are associated with more severe or therapy-resistant MD [15,30,31,32], we found no significant difference in their prevalence between MD patients and controls. This could be attributed to our limited sample size or other unaccounted factors, such as genetic predisposition or acquired vulnerability to “brain–ear” SVD. Emerging data suggests that MD may be a systemic oxidant disease [33], with excessive free radical generation and oxidative stress contributing to both EH/MD development and cerebral SVD [33].

Our study also found no difference in cortical stroke incidence between groups. This suggests that the association between MD and cerebral ischemia may be more dependent on small-vessel dysfunction rather than large-vessel disease.

Several mechanisms could explain the link between MD, cardiovascular factors, and cerebrovascular disease. EH, a hallmark of MD, may be influenced by vascular dysfunction through altered endolymphatic sac blood flow and fluid balance [34]. Microvascular damage, oxidative stress, atherosclerosis, and microthrombosis can disrupt inner ear fluid homeostasis, increasing the risk of EH and MD [35,36,37]. Conversely, EH may act as a Starling resistor on inner ear vasculature, triggering ischemic episodes in individuals with low ear perfusion pressure due to “brain–ear” SVD [12].

The autonomic nervous system, which plays a crucial role in cardiovascular regulation, may also be involved in MD pathophysiology. Dysregulation of this system could affect inner ear blood flow and fluid dynamics, contributing to MD symptoms [18]. Additionally, recent research suggests a potential role for venous stasis and cerebrospinal venous insufficiency in EH and MD development [38,39,40].

Interestingly, current treatments for MD aim to promote inner ear vascularization and reduce fluid volume [41]. Vasodilators like betahistine and nimodipine have shown efficacy in MD [41]. Dietary salt reduction and diuretic use, both impacting cardiovascular risk by lowering blood pressure, are also employed [15].

This study has limitations, including its cross-sectional design, which prevents establishing causality, and its relatively small sample size from a single center, limiting generalizability. Further research is needed, ideally with a larger, multicenter cohort, to validate these findings. The radiological assessments were performed by a single experienced neuroradiologist. While this approach ensured consistency in the evaluations, future studies may benefit from including independent assessments by multiple radiologists to enhance inter-rater reliability. The lack of T2* sequence in a significant proportion of the sample prevented the inclusion of CMBs in the SVD-3 score. Additionally, the retrospective nature of the study and reliance on clinical records may have led to incomplete documentation of comorbidities. Regarding the control sample, it is important to note that while pituitary pathology itself may not directly contribute to cerebrovascular disease, some underlying conditions (e.g., certain hormonal imbalances) or treatments associated with pituitary disorders could theoretically have a minor impact on vascular health. However, these potential effects are considered to be minimal compared to the primary advantages of using this control group. Moreover, while this study utilized conventional MRI sequences to assess cerebrovascular burden, future research could incorporate additional neuroimaging modalities such as CT, CT angiography, and CT perfusion to provide a more comprehensive evaluation of cerebrovascular disease [42]. Additionally, future research could investigate the impact of cerebral SVD burden on MD severity, incorporating audiological features and frequency of attacks. Further exploration of other potential contributing factors, such as vascular compression syndromes or anatomical variations in the inner ear, is also warranted.

Our study has several notable strengths. It is the first to demonstrate a potential association between cerebral microvascular lesions and MD. Furthermore, it provides relevant equations to estimate the risk of MD based on MRI-identified microvascular burden (Section 3.3). By highlighting a potential common vascular “brain–ear” disorder, our results contribute to the ongoing debate regarding the cardiocerebrovascular pathway for MD.

5. Conclusions

This study presents novel evidence suggesting a potential association between cerebral microvascular disease and MD. Findings highlight a significantly higher burden of small-vessel disease (SVD) in MD patients compared to controls, even after adjusting for potential confounders. This suggests that cerebrovascular dysfunction may play a role in the pathophysiology of MD.

The exact mechanisms underlying this association remain to be elucidated. However, our results underscore the importance of further investigating the “brain–ear” hypothesis and the potential contribution of cardiovascular risk factors to the development and progression of MD.

This research, if confirmed by larger, prospective studies, has important implications for future clinical practice and is of interest to a wide range of healthcare professionals, including ENT specialists, neurologists, imaging specialists, and other allied healthcare workers involved in the management of patients with MD. Emphasizing the importance of identifying and addressing cerebrovascular risk factors in these patients could potentially improve outcomes and quality of life for individuals with MD.