Introduction

ISSNHL is characterized by an abrupt and unexplained decline in sensorineural hearing, often accompanied by symptoms such as tinnitus, vertigo, nausea, and vomiting, and it is increasingly recognized as a significant public health concern globally [1,2]. This condition predominantly affects individuals aged 40 and older, with an annual morbidity rate estimated between 5 to 20 cases per 100,000 individuals [3]. The prevalence of ISSNHL ranges from 0.07% to 5.2%, exhibiting an age-related increase and significant variability across different countries within the adult population [4]. There remains relatively scant information regarding the etiology of ISSNHL, thereby complicating the development of successful treatment strategies. Primary intervention for mitigating hearing loss involves early administration of repetitive systemic steroid therapy [5]. However, the absence of a definitive etiological diagnosis poses challenges for precise management of ISSNHL. Therefore, both prompt diagnosis and timely intervention are of utmost urgency and great significance. Recent research has embarked on exploring the potential role of vitamin D in ISSNHL, focusing on associated risk factors and underlying mechanisms, which presents a promising avenue for further investigation.

Vitamin D, a fat-soluble hormone predominantly synthesized through exposure to sunlight, can also be obtained from dietary sources and multivitamin supplements, which is critical for the maintenance of bone health and exhibits a range of beneficial effects, including antiproliferative and prodifferentiative actions, metabolic regulation, and support for neuromuscular function [6,7]. The potential influence of vitamin D on auditory health has attracted scholarly interest, with research indicating a link between vitamin D deficiency and hearing impairments. Evidence suggests that vitamin D is integral to the proliferation and differentiation of neural stem and progenitor cells, which are essential for proper functioning of the auditory nerve [8]. Vitamin D deficiency may result in cochlear calcium demineralization, potentially disrupting microcirculation, and leading to hearing loss [9]. Studies have identified a higher prevalence of vitamin D insufficiency among individuals with ISSNHL, suggesting a possible association between serum vitamin D level and the onset of ISSNHL [10]. This finding is further corroborated by Kose’s research [11]. Furthermore, the role of vitamin D in the inner ear, potentially mediated through its receptors or its involvement in calcium metabolism, may contribute to sensorineural hearing loss when deficient [12]. Moreover, the influence of vitamin D on vascular cell function and its capacity to enhance the expression of anti-inflammatory cytokines suggests further mechanisms by which a deficiency could impact hearing [13,14]. However, traditional epidemiological interpretations may be complicated by various confounding factors, leading to the establishment of a causal relationship between the exposure factors and outcomes.

Mendelian randomization (MR) research design adheres to the Mendelian genetic law that “parental alleles are randomly allocated to offspring”, offers a robust framework for evaluating the causal relationship between exposure and outcome by employing SNPs as IVs [15]. This methodological approach is inherently resistant to confounding factors, and can effectively avert the influence of confounding factors and eliminate the interference of reverse causality. As the genetic variations present in individuals are randomly assigned at conception, independent of environmental or behavioral influences. As a result, MR is often regarded as a natural counterpart to traditional randomized controlled trials (RCTs) [16]. In the present study, we sought to clarify the causal relationship between 25-hydroxyvitamin D levels, a recognized indicator of nutritional vitamin D status, and ISSNHL by utilizing MR methodology. To the best of our knowledge, this study represents an inaugural investigation of the causal link between serum 25-hydroxyvitamin D levels and the risk of ISSNHL through MR analysis.

Methods and materials

Study design

We conducted a two-sample MR study aimed at deriving a comprehensive and reliable conclusion regarding the causal relationship between serum 25-hydroxyvitamin D levels (the exposure) and ISSNHL (the outcome). Serum 25-hydroxyvitamin D levels can be assessed through blood tests. ISSNHL, which is typically diagnosed according to the “Guidelines for Diagnosis and Treatment of Sudden Deafness (2015)”. A schematic representation of the study design is illustrated in Fig 1. This study is a bidirectional MR study, and follows three fundamental assumptions: Assumption 1: the selected instrumental variables exhibit a strong correlation with the exposure (P < 5 × 10⁻⁸); Assumption 2: there is no correlation between the IVs and the outcome; and Assumption 3: the IVs do not exert an influence on the outcome through confounding variables other than the exposure [17].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Design of the MR study on the causal relationship between serum 25(OH)D levels and ISSNHL.

https://doi.org/10.1371/journal.pone.0322898.g001

Results

SNP dressing by screening and MR analysis

Through screening for SNPs, we obtained 117 LD-independent SNPs from the 25-hydroxyvitamin D GWAS data. We examined potential confounding factors associated with these 117 SNPs and confirmed that none were related. Therefore, these 117 SNPs could be utilized in the GWAS of ISSNHL. The F-statistics for the SNPs in this study all exceeded 10, reinforcing the reliability of the research findings. Detailed information can be found in S1 Appendix.

The casual impact on ISSNHL of each IV was shown in the forest plot (Fig 2). It can be seen from the red line at the bottom of the forest plot that the combined analysis result crosses over 0, indicating that the level of serum 25-hydroxyvitamin D has no causal effect on ISSNHL.

The findings from IVW, weighted median, weighted model, simple model, and MR-Egger regression analyses are presented in Table 1. In this investigation, the IVW method yielded the primary result concerning causal effects, which indicated that serum concentrations of 25-hydroxyvitamin D do not exert a causal effect on ISSNHL. (IVW; Odds Ratio = 1.09, 95% Confidence Interval = 0.81 to 1.48, P = 0.573). The analysis results of weighted median, weighted model, simple model, and MR-Egger regression were consistent with the IVW result (Fig 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Associations between the serum 25-hydroxyvitamin D level and risk of idiopathic sudden sensorineural hearing loss.

https://doi.org/10.1371/journal.pone.0322898.t001

The results of the heterogeneity analysis showed that there was no heterogeneity in the IVW analysis, which was evaluated by Cohran’s Q test (P > 0.05, Table 2). The MR-Egger regression indicated that IVs have no pleiotropy of IVs (intercept = 0.002233724, se = 0.00670676, P = 0.739709 (> 0.05)). There was no outliers exist, and the funnel plot was roughly symmetrical (Fig 4A). The leave-one-out analysis was performed by systematically excluding individual SNPs and then assessing the cumulative effects of the remaining SNPs. The findings indicated that the results of the remaining SNPs consistently fell to the right of the invalid line following the exclusion of any single SNP. This observation substantiates that no individual SNP exerted a disproportionate influence on the MR results (Fig 4B), thereby affirming the robustness of the MR findings.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Heterogeneity testing of IVs for the serum 25-hydroxyvitamin D level.

https://doi.org/10.1371/journal.pone.0322898.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The forest plots illustrating the estimated causal effects of individual SNPs.

https://doi.org/10.1371/journal.pone.0322898.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Scatter plot of MR analysis.

The slope of each colored line represent the estimated MR effect in different models.

https://doi.org/10.1371/journal.pone.0322898.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Sensitivity analysis.

A. Funnel plot for the instrumental variables. B. Leave-one-out of sensitivity analysis. Calculate the MR results of the remaining IVs after removing the IVs one by one.

https://doi.org/10.1371/journal.pone.0322898.g004

Reverse MR analysis

Inverse MR analysis was performed with ISSNHL as an exposure factor and serum 25-hydroxyvitamin D levels an outcome. The findings from IVW, weighted median, weighted model, simple model, and MR-Egger regression analyses indicated that there was no evidence of a causal relationship between ISSNHL and serum 25-hydroxyvitamin D levels (Table 3). In this investigation, the IVW method yielded the primary result concerning causal effects (IVW; Odds Ratio = 1.00, 95% Confidence Interval = 0.99 to 1.00, P = 0.282). Heterogeneity analysis showed that there was no heterogeneity in the IVW analysis, which was evaluated by Cohran’s Q test (P > 0.05, Table 4). Pleiotropy test indicated that IVs have no pleiotropy of IVs (intercept = 0.00323687, se = 0.00233689, P = 0.30027999 (>0.05)). Detailed information can be found in S2 Appendix.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Associations between ISSNHL and the serum 25-hydroxyvitamin D level.

https://doi.org/10.1371/journal.pone.0322898.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Heterogeneity analysis.

https://doi.org/10.1371/journal.pone.0322898.t004

Discussion

Numerous studies have established a correlation between serum vitamin D levels and ISSNHL. In a study conducted by Zandi et al. [11], a cohort comprising 50 patients diagnosed with ISSNHL and 50 healthy individuals without hearing impairment, serving as a control group, was analyzed. The findings indicated a heightened prevalence of vitamin D insufficiency among individuals with SSNHL, thereby suggesting a potential link between serum vitamin D levels and the onset of SSNHL. Additionally, Szeto et al. [24] conducted a cross-sectional analysis involving 1,123 participants aged 70 years and older, which revealed that low vitamin D status was correlated with low-frequency hearing loss (LFHL) and speech-frequency hearing loss (SFHL) in the elderly population. Furthermore, a survey by Chen et al. [25] investigated the relationship between serum concentrations of 25-hydroxyvitamins D2 and D3 and hearing loss in U.S. adults, uncovering a positive correlation between serum 25(OH)D2 levels and both LFHL and SFHL within the studied group. An L-shaped relationship was also identified between serum 25(OH)D3 and LFHL and SFHL, indicating that elevated serum 25(OH)D3 levels were associated with a reduced risk of high-frequency hearing loss (HFHL) in Chen’s study.

The presence of a causal relationship between the two variables remains ambiguous. To address this uncertainty, we conducted a MR study aimed at exploring the potential causal association between them. This investigation was facilitated by the employment of publicly accessible GWAS databases, with meticulous attention paid to controlling for various confounding factors. Our findings suggest that there is no causal relationship between serum vitamin D levels and ISSNHL. This outcome stands in contrast to the findings of current observational studies. Based on this, a reverse MR analysis was further incorporated to investigate whether the occurrence of ISSNHL was correlated with the levels of serum vitamin D. Nevertheless, negative outcomes were still yielded. To the best of our knowledge, this study represents the inaugural effort to assess the causal relationship between these two variables. This result has significant clinical implications, as it suggests that the supplementation of vitamin D may not confer protective effects against sudden hearing loss. Clinicians should exercise caution when recommending vitamin D supplementation for this purpose, as it may lead to unnecessary healthcare expenditures without demonstrable benefit.

The observed discrepancies may be attributed to the inherent limitations associated with observational research methodologies. Primarily, these studies utilized cross-sectional data, which inherently restricts the ability to establish a causal relationship over time between serum 25-hydroxyvitamin D levels and ISSNHL. Furthermore, the focus of these investigations was confined to specific populations or to elderly individuals identified as being at heightened risk for hearing impairment. Additionally, the constraints of the available databases limited the capacity to control for all potential confounding variables, such as comorbid conditions like osteoporosis and fluctuations in parathyroid hormone levels, both of which may influence vitamin D levels and auditory health. Moreover, the findings may be affected by a bidirectional relationship, for instance, individuals with hearing impairments may tend to remain indoors more frequently, resulting in decreased sunlight exposure and, subsequently, vitamin D deficiency [26]. Another limitation of the studies in question is the lack of data regarding specific variables of interest, such as vitamin B-12 levels and the season during which 25-hydroxyvitamin D levels were assessed, both of which could potentially impact either hearing loss or vitamin D concentrations. Our MR study presents several advantages. The utilization of extensive genetic consortium data pertaining to serum 25-hydroxyvitamin D levels, which encompasses a considerable sample size (n = 33,996), in conjunction with ISSNHL risk data derived from a substantial number of cases (1,491) and controls (196,592), has facilitated a more rigorous examination of our hypothesis. This approach mitigates the limitations typically associated with individual-level data from studies with smaller cohorts. Furthermore, the application of stringent criteria for instrumental variables and the two-sample MR analysis has ensured the minimization of potential confounding factors.

In our MR analysis, we investigated the causal relationship between serum 25-hydroxyvitamin D concentrations and ISSNHL. One potential concern raised is the influence of seasonal variations on serum vitamin D levels, as sunlight exposure, a primary source of vitamin D synthesis, varies significantly across seasons. Seasonal fluctuations in vitamin D levels are well-documented, with higher concentrations typically observed during summer months due to increased ultraviolet B (UVB) radiation exposure, and lower levels during winter months. These variations could theoretically impact the estimation of vitamin D’s causal effect on ISSNHL if not properly accounted for. However, a key strength of our MR approach is its robustness to such environmental confounders, including seasonal variations. MR leverages genetic variants as instrumental variables, which are randomly assigned at conception and are not influenced by seasonal or environmental factors. This methodology is predicated on the assumption that genetic variations influence health outcomes exclusively through designated exposure factors. The random assignment of IVs at conception effectively positions the MR analysis as a “natural randomized controlled trial,” serving as a robust mechanism for inferring causal relationships. Since the genetic determinants of serum 25-hydroxyvitamin D levels are fixed and independent of temporal or seasonal changes, our MR estimates are less susceptible to bias from these fluctuations. This inherent feature of MR provides a more reliable estimate of the causal effect, even in the presence of seasonal variability in vitamin D levels [27].

It also should be noted that hearing loss manifests differently across various frequency ranges, including low, mid, and high frequencies, each potentially associated with distinct pathological mechanisms. Although our study did not stratify the analysis by frequency ranges, the findings provide valuable insights into the overall causal relationship between serum 25-hydroxyvitamin D concentrations and ISSNHL. ISSNHL typically involves multiple frequency ranges, and our results suggest that vitamin D levels may not exert a significant causal effect on the condition as a whole. However, future research should explore whether vitamin D levels differentially impact hearing loss across specific frequency ranges, as this could reveal more nuanced relationships and inform targeted interventions.

Our MR results estimate the LATE of serum 25-hydroxyvitamin D on ISSNHL risk, representing the causal effect among compliers whose vitamin D levels are modifiable by the genetic instruments. This subgroup-specific effect may not equate to the average causal effect in the general population, as genetic variants typically exert smaller physiological effects compared to clinical interventions. The monotonicity assumption underlying LATE, while supported by our selection of unidirectional vitamin D-associated SNPs from GWAS, remains untestable and warrants cautious interpretation. Notably, non-compliers and effect heterogeneity across subpopulations could limit generalizability. These considerations align with broader limitations of MR for estimating population-wide effects but do not invalidate the utility of LATE for identifying modifiable biological pathways. Future studies using vitamin D supplementation trials or gene-environment interaction analyses may help bridge this translational gap.

Emerging evidence highlights vitamin D as a key immunomodulator. As demonstrated by Murdaca et al. [28], 25(OH)D deficiency impairs the balance between pro-inflammatory and regulatory immune responses, particularly through suppression of Th17-mediated pathways and enhancement of Treg cell differentiation. Notably, the immunomodulatory effects of vitamin D may be partially mediated by gut microbiota. Murdaca et al. [29] revealed that vitamin D receptor signaling promotes colonization of anti-inflammatory commensal bacteria, while vitamin D deficiency exacerbates dysbiosis and gut barrier dysfunction. This ‘gut-immune axis’ could theoretically extend to the inner ear microenvironment, where microbial metabolites or systemic inflammation triggered by dysbiosis might precipitate cochlear inflammation and vascular compromise. Although our MR analysis focused on serum 25(OH)D, the observed association with ISSNHL might be amplified through microbiota-immune interactions. For instance, vitamin D sufficiency could maintain gut barrier integrity, thereby reducing bacterial translocation and subsequent endotoxin-induced cochlear microvascular inflammation - a hypothesized mechanism in ISSNHL pathogenesis.

While our research yielded noteworthy findings, it is crucial to recognize certain limitations. Firstly, the study’s participant pool was predominantly composed of individuals of European descent, indicating the need for additional data collection and analysis to ascertain whether the results can be extrapolated to other ethnic groups. Secondly, hearing loss was operationally defined as a pure tone average exceeding 25 dB in either ear at low frequencies, specifically across 500, 1,000, and 2,000 Hz, as well as at higher frequencies of 4,000, 6,000, and 8,000 Hz. It is important to note that the frequency ranges utilized in this study were not exhaustively analyzed. This was due to the lack of detailed frequency-specific data in the datasets used for our MR analysis. As hearing loss can manifest differently across frequency ranges, future studies should aim to investigate whether serum 25-hydroxyvitamin D concentrations have frequency-specific effects on hearing loss, to provide deeper insights into the role of vitamin D in auditory health and help identify potential subgroups that may benefit from vitamin D supplementation. Thirdly, despite the robustness of our MR analysis to seasonal variations, we acknowledge that future studies could further explore the impact of seasonal changes on serum vitamin D levels and their potential association with ISSNHL. For instance, longitudinal studies measuring vitamin D levels across different seasons could provide additional insights into how temporal fluctuations might influence the risk of ISSNHL. Incorporating such data into future MR analyses or other causal inference frameworks could help refine our understanding of the relationship between vitamin D and hearing health. Fourthly, we acknowledge that conventional observational methods like OLS regression offer distinct advantages for probing nonlinear relationships and adjusting for measured confounders. However, these approaches are ill-suited to our research context for two reasons. First, the publicly available GWAS data lack individual-level measurements of critical covariates. Second, and more fundamentally, the bidirectional relationship between vitamin D status and hearing-related disability_whereby severe ISSNHL may limit sunlight exposure, creating reverse causality_renders OLS estimates vulnerable to substantial bias. Mendelian randomization provides a principled alternative by leveraging genetic variants as instrumental variables, naturally adjusting for lifelong confounding through Mendel’s laws of inheritance. While this approach requires careful evaluation of pleiotropy, as addressed through sensitivity analyses, it avoids the insurmountable endogeneity problems inherent to observational vitamin D studies. Furthermore, the research underscores the significant role of genetic factors in the onset of ISSNHL and vitamin D deficiency, suggesting a complex interplay of origins for ISSNHL. In other word, while we did not find evidence for a causal relationship between vitamin D and sudden hearing loss, it remains plausible that vitamin D could influence auditory function through mechanisms such as immune modulation or inflammatory response, which may not be captured in our analysis. Additionally, environmental and social factors, which are non-genetic in nature, also play a substantial role in these conditions, however, these elements were not comprehensively addressed in this study. Moreover, Due to the lack of individual-level data for the exposure and outcome samples, we are unable to assess potential overlaps between these samples which may leads to bias of our finding.

Conclusion

Our MR-based investigation did not reveal any substantial evidence of a significant causal relationship between genetically determined serum 25-hydroxyvitamin D levels and the risk of ISSNHL among individuals of European ancestry. Nevertheless, there remains a pressing need for additional longitudinal studies employing larger datasets and more heterogeneous populations to investigate the causal mechanisms involved.

Supporting information

Acknowledgments

The authors thank all the participants and support of the Central Laboratory members of The Affiliated Panyu Central Hospital, Guangzhou Medical University and member of Department of laboratory medcine, Third affiliated hospital of Sun Yat-sen University.