Diet, Metabolic Biomarkers, and Sensorineural Hearing Loss: A Mendelian Randomization and Mediation Analysis

doi:10.21203/rs.3.rs-4887944/v1

Download PDF

Article

Diet, Metabolic Biomarkers, and Sensorineural Hearing Loss: A Mendelian Randomization and Mediation Analysis

https://doi.org/10.21203/rs.3.rs-4887944/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Sensorineural hearing loss (SNHL) is influenced by genetic and environmental factors. This study explores the causal relationships between dietary choices, circulating metabolic biomarkers, and SNHL using Mendelian Randomization (MR) analysis. We utilized summary statistics from genome-wide association studies (GWAS) on dietary choices and metabolic biomarkers as exposures, with SNHL as the outcome. Two-sample MR analysis and mediation analysis were conducted to identify causal links and metabolic pathways. The MR analysis revealed that fish, oily fish, healthy food, and vegetable consumption significantly reduce the risk of SNHL. The mediation analysis further indicated that these protective effects are mediated by the ratios of DHA, omega-3, and polyunsaturated fatty acids to total fatty acids, with DHA mediating approximately 10.57% of the protective effect of oily fish consumption. Significant causal pathways link dietary choices and metabolic biomarkers to SNHL, suggesting that dietary modifications, particularly increasing the intake of fish, healthy foods, and vegetables, could serve as effective preventive strategies against SNHL. This study provides valuable insights into the interplay between diet, metabolism, and hearing health, guiding future preventive strategies.

Sensorineural hearing loss (SNHL)

Mendelian Randomization (MR)

dietary choices

circulating metabolic biomarkers

omega-3 fatty acids

docosahexaenoic acid (DHA)

Sensorineural hearing loss (SNHL), results from damage to the inner ear's sensory cells or the auditory nerve pathways to the brain ¹. This condition can arise from various causes, including congenital factors, trauma, infections, ototoxic drugs, and age-related changes ². Epidemiological studies indicate that SNHL is relatively common. The global prevalence of hearing loss is increasing, with the World Health Organization (WHO) predicting that by 2050, nearly 2.5 billion people will experience some degree of hearing loss, with at least 700 million requiring hearing rehabilitation ³. Additionally, unsafe listening practices pose a risk of permanent hearing loss to over one billion young people worldwide ⁴.

SNHL significantly impacts both the quality of life for individuals and public health. Socially, hearing loss hampers effective communication, leading to reduced social interactions and strained personal relationships ¹. Psychologically, it can contribute to loneliness, anxiety, and depression. Professionally, hearing impairment can limit career opportunities, particularly in fields requiring acute auditory abilities. Safety-wise, those with hearing loss are at higher risk of accidents due to their inability to hear warnings or alarms. Cognitively, studies have linked hearing loss to declines in cognitive function and an increased risk of dementia ². For children, SNHL can hinder educational development and learning. Economically, the costs associated with treatment and rehabilitation, including medical expenses and hearing aids, add a significant financial burden ². Additionally, the growing number of individuals with hearing loss necessitates increased allocation of public health resources to address this challenge effectively.

Dietary choices profoundly impact overall health, including hearing health. A balanced diet provides essential nutrients such as proteins, fats, carbohydrates, vitamins, and minerals necessary for the body's proper functioning. Specifically, certain dietary patterns and nutrients have been linked to the maintenance and protection of hearing. Omega-3 fatty acids, predominantly found in fish like salmon, mackerel, and tuna, benefit heart health and can improve blood flow, positively affecting inner ear health ⁵. Vegetables and other healthy foods are rich in vitamins and minerals, such as vitamins A, C, E, and zinc, which are crucial for maintaining normal ear function ⁶. Antioxidants in fruits and vegetables help reduce damage from free radicals, protecting hearing ⁷.

Studies have shown that adhering to healthy dietary patterns, such as the Mediterranean or DASH diets, can lower the risk of hearing decline by 25–30% ⁸. These diets are rich in carotenoids, beta-carotene, folate, and long-chain omega-3 fatty acids, all of which contribute to hearing preservation. Avoiding unhealthy foods high in fats and sugars and managing blood glucose levels are also critical, as conditions like diabetes and hypertension are risk factors for hearing loss ⁴. Thus, dietary choices play a significant role in hearing health, highlighting the importance of a nutrient-rich diet in preventing SNHL and promoting overall auditory well-being.

Circulating metabolic biomarkers are molecules measurable in the blood that reflect the state of the body's metabolic processes and serve as indicators of health and disease. These biomarkers include proteins, lipids, metabolites, and hormones, whose levels vary under different physiological and pathological conditions. For example, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol are closely associated with cardiovascular health and heart disease risk, respectively ⁹. Blood glucose levels are crucial for diagnosing and managing diabetes, while high triglyceride levels are linked to an increased risk of metabolic syndrome and cardiovascular diseases ¹⁰.

In terms of hearing health, circulating metabolic biomarkers play a significant role. Omega-3 fatty acids, such as DHA and EPA, have been shown to positively impact brain and hearing health, potentially slowing age-related hearing decline ⁵. Antioxidant vitamins C and E help protect inner ear cells from oxidative stress, thus maintaining hearing function ⁶. Essential minerals like magnesium and zinc are vital for normal ear function; magnesium helps protect against noise-induced hearing damage, while zinc is crucial for maintaining cochlear function ¹¹. Elevated levels of inflammatory markers like C-reactive protein (CRP) may indicate inflammation in the ear, potentially affecting hearing. Abnormal lipid and lipoprotein levels can also be associated with hearing loss; high triglyceride levels may affect blood supply to the ear, consequently impacting hearing ¹². Therefore, circulating metabolic biomarkers are significant in regulating overall health and particularly important in the context of hearing health.

Mendelian Randomization (MR) is a statistical method that leverages genetic variations as instrumental variables to assess causal relationships between environmental exposures and health outcomes ¹³. Rooted in Mendel's laws of inheritance, particularly the law of independent assortment, MR assumes that alleles at different loci assort independently during gamete formation, akin to random assignment in controlled trials. This method offers several advantages in causal inference. Firstly, MR overcomes confounding factors. Since genetic variations are determined at conception, they remain unaffected by lifestyle or environmental factors, thereby reducing confounding biases in causal analysis. Secondly, MR addresses reverse causation issues, helping to clarify whether an exposure leads to an outcome or vice versa. This is particularly useful in complex epidemiological studies ¹⁴.

MR is also ethically and practically advantageous when randomized controlled trials (RCTs) are unfeasible. For instance, it is impractical or unethical to randomly assign individuals to smoke or drink alcohol for extended periods. MR uses genetic variants, typically single nucleotide polymorphisms (SNPs), as proxies for these exposures, providing a feasible alternative ¹⁵. Additionally, MR's utility extends to leveraging large-scale genome-wide association studies (GWAS) data, making research more efficient and cost-effective ¹⁶. It captures lifelong exposure effects, offering insights into long-term health impacts that traditional studies might miss. MR reduces biases inherent in observational studies, such as confounding and reverse causation, enhancing the robustness of causal inferences.

Overall, MR complements traditional observational studies and RCTs by providing a powerful tool for evaluating the causal effects of modifiable risk factors on health outcomes. The primary objective of this study is to explore the causal relationships between dietary choices and sensorineural hearing loss (SNHL) using Mendelian Randomization (MR). Additionally, the study aims to examine the mediating effects of specific circulating metabolic biomarkers on these relationships, providing a comprehensive understanding of the pathways through which diet impacts hearing health.

2.1 Study Design

We obtained published GWAS summary data on dietary choices, circulating metabolic biomarkers, and sensorineural hearing loss (SNHL) to conduct two-sample Mendelian Randomization (MR) analyses. These analyses aimed to evaluate the causal relationships between these exposures and SNHL. The study design is illustrated in Fig. 1. TSMR relies on three key assumptions: (1) there is a strong correlation between genetic variants and the exposure factors (dietary choices and metabolic biomarkers); (2) the genetic variants are independent of confounding factors; (3) the genetic variants affect the outcome (SNHL) only through the exposure factors, not through other pathways. Additionally, two-step and multivariable MR (MVMR) analyses were conducted to identify the mediation effects of specific metabolic biomarkers in the relationship between dietary choices and SNHL ^17,18. This study adheres to the STROBE-MR guidelines and does not require additional ethical approval as it involves reanalysis of previously published data.

Figure 1. Study Design and Analytical Framework

The figure illustrates the study design and analytical framework used to investigate the causal relationships between dietary choices, circulating metabolic biomarkers, and sensorineural hearing loss (SNHL) using two-sample Mendelian Randomization (MR) and mediation analysis.

2.2 Data Sources

Summary statistics for dietary choices, circulating metabolic biomarkers, and sensorineural hearing loss (SNHL) were obtained from large-scale genome-wide association studies (GWAS).

2.2.1 Dietary Choices

Data on dietary choices were sourced from two comprehensive GWAS. The first study, conducted by May-Wilson et al., involved 161,625 participants from the UK Biobank, who rated their liking for 139 specific foods on a 9-point scale. This study identified 1,401 significant associations for food preferences, revealing genetic determinants and correlations with distinct neurophysiological traits. The second study, conducted by Pirastu et al., analyzed 38 specific food consumption traits in 445,779 participants from the UK Biobank, identifying 283 genetic markers associated with dietary intake. Together, these studies provide a robust dataset of 187 food liking traits and 38 food consumption traits, accessible from the GWAS Catalog (GCST90094772-GCST90094871 and GCST90096892-GCST90096929, respectively) ¹⁹.

2.2.2 Circulating Metabolic Biomarkers

Summary statistics for circulating metabolic biomarkers were obtained from a large-scale GWAS conducted by Karjalainen et al. (PMID: 38448586). This study focused on a mixed-ancestry population predominantly of European descent, involving 136,016 participants from 33 cohorts. Using nuclear magnetic resonance spectroscopy, the study quantified 233 metabolic traits and identified more than 400 independent loci associated with these traits. Probable causal genes were assigned to two-thirds of these loci through manual curation. These comprehensive data can be accessed from the GWAS Catalog (GCST90302020-GCST90302152) ²⁰.

2.2.3 Sensorineural Hearing Loss (SNHL)

Data for SNHL were derived from the FinnGen biobank analysis (round 10). This study involved 397,841 participants, including 35,488 cases and 362,353 controls, predominantly of European descent. A total of 19,345,470 SNPs were analyzed. All SNHL cases were identified using ICD-10 codes H90.3, H90.4, and H90.5 from hospital records ²¹. The details of the data sources are summarized in Table 1.

2.3 Selection of Instrumental Variables

A locus-wide significance threshold of P < 5×10^− 8 was used to screen SNPs related to the exposure factors in dietary choices and circulating metabolic biomarkers, ensuring a strong association between the instrumental variables and the exposures. To avoid weak instrument bias, the F statistic for each IV was calculated using the formula:

$$\:F=\frac{{R}^{2}\times\:\left(n-k-1\right)}{k\times\:(1-{R}^{2})}$$

where R² is the portion of the exposure variance explained by the IVs, n is the sample size, and k represents the number of IVs. IVs with F < 10 were excluded. Linkage disequilibrium (LD) analysis was performed with r² < 0.001 and a clumping window > 10,000 kb to ensure independence among the selected SNPs. Palindromic SNPs with intermediate allele frequencies were excluded to avoid potential strand orientation ambiguity ²².

To further enhance the robustness of the selected instrumental variables and avoid reverse causality, the Steiger Test was used to ensure that the association of each SNP with the exposure was stronger than its association with the outcome. SNPs that did not pass this test were considered likely unrelated to the exposure and were excluded from the analysis.

The MR pleiotropy residual sum and outlier (MR-PRESSO) test was applied to detect horizontal pleiotropy, and the pleiotropy effect was eliminated by removing the outliers

In the reverse MR analysis, a genome-wide significance threshold of P < 5×10^− 8 was used to screen SNPs related to sensorineural hearing loss as the exposure. The same LD criteria were applied, and palindromic and incompatible SNPs were excluded. SNPs that could not be matched in the GWAS outcome statistics were also excluded.

To further ensure the validity of our IVs, we utilized the PhenoScanner website (http://www.phenoscanner.medschl.cam.ac.uk/) to exclude SNPs directly associated with the outcomes. This comprehensive selection process helps minimize potential biases and ensures the robustness of our MR analysis ²³.

2.4 Statistical Analysis

2.4.1 Two-sample Mendelian Randomization Analysis

All analyses were performed using R version 4.3.1 (https://www.r-project.org). To assess the causal relationship between dietary choices, circulating metabolic biomarkers, and sensorineural hearing loss (SNHL), we employed various Mendelian Randomization (MR) methods using the TwoSampleMR package version 0.5.7. These methods included inverse variance weighted (IVW), MR Egger, weighted median, weighted mode, and inverse variance weighted (multiplicative random effects).

The IVW method was our primary analytical approach, as it combines multiple random variables to reduce overall variance by weighting each variable based on its variance, providing the most accurate estimates when all genetic variants are valid instruments ^24,25. MR Egger adjusts for potential pleiotropy by allowing an intercept term ²⁶. Weighted median offers consistent estimates even with up to 50% invalid instruments ²⁷. Weighted mode provides robust estimates under different scenarios of invalid instruments ²⁸. The multiplicative random effects IVW model was used when heterogeneity among the SNPs was present, providing more conservative and robust estimates in such cases.

2.4.3 Reverse Mendelian Randomization Analysis

To explore whether sensorineural hearing loss (SNHL) has a causal effect on dietary choices and circulating metabolic biomarkers, a reverse Mendelian randomization (MR) analysis was performed. In this scenario, SNPs associated with SNHL were regarded as instrumental variables (IVs), SNHL as the exposure, and dietary choices and circulating metabolic biomarkers as the outcomes. The reverse MR analysis procedure was similar to that used for the forward MR analysis, employing the same set of methods: inverse variance weighted (IVW), MR Egger, weighted median, weighted mode, and inverse variance weighted (multiplicative random effects).

The reverse MR analysis was conducted to determine if there is a bidirectional relationship between SNHL and the identified exposures, providing additional insights into the potential causal pathways.

2.4.3 Mediation Analysis

Mediation analysis aims to evaluate the pathway from exposure to outcome through a mediator, helping to explore the potential mechanisms by which exposure affects the outcome. The mediation analysis in this study focused on dietary choices, circulating metabolic biomarkers, and sensorineural hearing loss (SNHL).

First, the causal relationship between dietary choices and circulating metabolic biomarkers was evaluated using two-sample MR methods to obtain beta (A). Second, MVMR was used to screen circulating metabolic biomarkers that still had a causal relationship with SNHL after correction for dietary choices to obtain beta (B) and ensure that the mediating effects on outcomes are independent of exposure. The mediation effect was calculated using a two-step MR:

mediation effect = β(A)×β(B)

The total effect of dietary choices on SNHL was obtained in the previous two-sample MR, and the direct effect was calculated as:

direct effect = (total effect − mediation effect)

The mediation proportion used the following formula:

mediation proportion = ( total effect / mediation effect )×100%

The 95% confidence intervals (CI) for the mediation effects and proportions mediated were estimated using the delta method. Based on the results, we categorized the identified mediators into different levels of evidence. When only a triangular relationship existed, representing that exposure was causally associated with outcome, mediator was causally associated with outcome, and exposure was causally associated with mediator, the identified biomarkers were considered to have potential mediation effects in the pathway from dietary choices to SNHL. If the identified biomarkers did not only exist in a triangular relationship but also had mediation effects significantly different from 0, they were considered as mediators with strong evidence.

2.4.4 Sensitivity Analysis

Sensitivity analysis was conducted to ensure the robustness of our findings. The intercept term of the MR-Egger regression model was examined to assess potential pleiotropic effects ²⁶. If the P-value of the intercept term exceeded 0.05, it suggested minimal influence of genetic pleiotropy, indicating that the IVs solely impacted the risk of sensorineural hearing loss (SNHL) through dietary choices and circulating metabolic biomarkers. Cochran's Q test was employed to investigate the heterogeneity of the IVs and its potential impact on the causal estimate ²⁵. A leave-one-out analysis was performed to assess the sensitivity of the results by conducting MR analysis repeatedly, gradually eliminating one SNP at a time.

Additionally, the MR pleiotropy residual sum and outlier (MR-PRESSO) test was used to evaluate any disparities between the outcomes of MR analysis before and after correction, identifying and correcting for outliers contributing to bias ²⁹. MR Steiger analysis was used to assess the direction of the potential causal association between exposure and outcomes, ensuring that the direction of causality was correctly inferred. Only causal pathways with no heterogeneity or pleiotropy were included in the subsequent analysis when the IVW MR method results reached a significance threshold of P < 0.05.

Furthermore, considering the potential chance to increase the overall type I error during multiple comparisons, we implemented the false discovery rate (FDR) correction using the Benjamini–Hochberg procedure on the primary IVW results. A significance threshold of FDR < 0.05 indicates a significant association, whereas P_IVW < 0.05 but FDR > 0.05 implies a suggestive association ³⁰.

The instrumental variables (IVs) were screened based on previously described conditions. For all two-sample Mendelian randomization (TSMR) analyses, SNPs were selected using a genome-wide significance threshold of P < 5 × 10^− 8. These analyses included dietary choices and circulating metabolic biomarkers as exposures and sensorineural hearing loss (SNHL) as the outcome. Detailed information on the SNPs used is provided in Supplementary Table 1. The F-statistics of all SNPs ranged from 29.72 to 7610.08, indicating the absence of weak instruments. Additionally, all harmonized SNPs passed the Steiger test, confirming no evidence of reverse causation.

Positive results were initially obtained, after which the PhenoScanner website was utilized to exclude any exposure SNPs directly associated with the outcomes before performing the MR analysis again. Following this evaluation, no SNPs showed indications for removal, confirming the robustness of the selected instruments. In the reverse Mendelian randomization analysis, SNPs related to sensorineural hearing loss were screened as exposures using the same genome-wide significance threshold.

3.1 Impact of Dietary Choices on Sensorineural Hearing Loss (SNHL)

Using FDR correction (P_FDR < 0.05), several dietary factors were identified as protective against SNHL from 187 dietary preferences and 38 dietary consumptions. Significant associations were observed for fish consumption, healthy food consumption, vegetable consumption, and oily fish consumption (Supplementary Tables S1).

Fish Consumption: IVW method estimated the OR as 0.592 (95% CI = 0.467 to 0.751, P_FDR = 0.0006, P = 1.58E-05).

Healthy Food Consumption: IVW method estimated the OR as 0.655 (95% CI = 0.532 to 0.806, P_FDR = 0.001, P = 6.61E-05).

Vegetable Consumption: IVW method estimated the OR as 0.610 (95% CI = 0.473 to 0.785, P_FDR = 0.001, P = 0.0001).

Oily Fish Consumption: IVW method estimated the OR as 0.558 (95% CI = 0.386 to 0.807, P_FDR = 0.013, P = 0.0019).

Additionally, ten suggestive associations (P_FDR > 0.05 but P < 0.05) were identified, including fruit consumption and cooked vegetable consumption (Supplementary Tables S1). Non-oily fish consumption showed a significant association with SNHL, but the low SNP count and deviation in OR (0.068, 95% CI = 0.018 to 0.259) suggested potential bias and it was excluded from further analyses. Consistent results from MR Egger, weighted median, and weighted mode methods further validated these findings (Figs. 2, Fig. 4).

Figure 2. Associations Between Dietary Choices and Sensorineural Hearing Loss

3.2 Impact of Circulating Metabolic Biomarkers on Sensorineural Hearing Loss (SNHL)

In the two-sample Mendelian Randomization (MR) analysis, 31 suggestive associations (P_FDR > 0.05 but P < 0.05) were identified. Due to the lack of significant results after FDR correction, we selected three suggestive associations supported by previous literature for their relevance to health and hearing: the ratio of 22:6 docosahexaenoic acid (DHA) to total fatty acids, the ratio of omega-3 fatty acids to total fatty acids, and the ratio of polyunsaturated fatty acids to total fatty acids ^31–33. These ratios were chosen for further analysis (Supplementary Tables S2).

IVW Results for Selected Biomarkers:

Ratio of 22:6 Docosahexaenoic Acid to Total Fatty Acids: OR = 0.910 (95% CI = 0.858 to 0.965, P = 0.0015, P_FDR = 0.1181).

Ratio of Omega-3 Fatty Acids to Total Fatty Acids: OR = 0.930 (95% CI = 0.888 to 0.974, P = 0.0019, P_FDR = 0.1181).

Ratio of Polyunsaturated Fatty Acids to Total Fatty Acids: OR = 0.907 (95% CI = 0.843 to 0.976, P = 0.0094, P_FDR = 0.1181).

These findings suggest that specific dietary choices may influence levels of circulating metabolic biomarkers, which in turn could impact the risk of sensorineural hearing loss (SNHL). Further research is needed to explore these pathways and their implications for dietary recommendations and public health strategies (Figs. 3, Fig. 4).

Figure 3. Associations Between Circulating Metabolic Biomarkers and Sensorineural Hearing Loss

3.3 Causal Relationships Between Dietary Choices and Circulating Metabolic Biomarkers

Given the suggestive associations between circulating metabolic biomarkers and SNHL, we selected biomarkers that have shown significant associations with SNHL or the nervous system in previous observational studies. These biomarkers include the ratio of 22:6 docosahexaenoic acid (DHA) to total fatty acids, the ratio of omega-3 fatty acids to total fatty acids, and the ratio of polyunsaturated fatty acids to total fatty acids. We used significant dietary factors as exposures and these biomarkers as outcomes to analyze causal relationships.

Significant causal relationships were identified through FDR correction (P_FDR < 0.05):

Ratio of 22:6 Docosahexaenoic Acid to Total Fatty Acids:

Fish Consumption: Beta = 0.445 (SE = 0.064, P = 5.13E-12, P_FDR = 8.99E-11).

Healthy Food Consumption: Beta = 0.293 (SE = 0.063, P = 3.89E-06, P_FDR = 2.82E-05).

Vegetable Consumption: Beta = 0.323 (SE = 0.070, P = 4.03E-06, P_FDR = 2.82E-05).

Oily Fish Consumption: Beta = 0.653 (SE = 0.070, P = 9.21E-21, P_FDR = 3.22E-19).

Ratio of Omega-3 Fatty Acids to Total Fatty Acids:

Fish Consumption: Beta = 0.450 (SE = 0.062, P = 2.90E-13, P_FDR = 3.38E-12).

Healthy Food Consumption: Beta = 0.377 (SE = 0.046, P = 4.53E-16, P_FDR = 7.94E-15).

Vegetable Consumption: Beta = 0.317 (SE = 0.065, P = 1.10E-06, P_FDR = 9.59E-06).

Oily Fish Consumption: Beta = 0.686 (SE = 0.074, P = 1.08E-20, P_FDR = 3.78E-19).

Ratio of polyunsaturated fatty acids to total fatty acids:

Healthy food consumption: Beta = 0.247 (SE = 0.063, P = 0.0001, P_FDR = 0.0016).

Fish consumption: Beta = 0.289 (SE = 0.078, P = 0.0002, P_FDR = 0.0018).

Oily fish consumption: Beta = 0.292 (SE = 0.096, P = 0.0022, P_FDR = 0.0112).

Vegetable consumption: Beta = 0.151 (SE = 0.085, P = 0.0767, P_FDR = 0.1798).

These findings suggest that certain dietary choices are significantly associated with specific circulating metabolic biomarkers, providing insights into potential mechanisms linking diet to SNHL.

Figure 4. Causal Relationships Between Dietary Choices, Circulating Metabolic Biomarkers, and Sensorineural Hearing Loss (SNHL)

Notes: Colored dots represent the relationships: red dots indicate risk factors, green dots indicate protective factors, and gray dots indicate no clear causal relationship found.

3.4 Reverse Mendelian Randomization Analysis

In our exploration of the causal effects of sensorineural hearing loss (SNHL) on dietary choices and circulating metabolic biomarkers, we employed the Inverse Variance Weighted (IVW) method as the principal analysis in a two-sample Mendelian Randomization (MR) study. Despite conducting a comprehensive analysis and applying False Discovery Rate (FDR) correction, we did not identify any significant causal relationships between SNHL and the dietary choices or circulating metabolic biomarkers examined (Supplementary Table S3). No significant results were found, indicating that SNHL does not have a robust causal effect on these factors.

3.5 Mediation Analysis Results

The mediation analysis evaluated 12 combinations using 4 types of dietary exposures (fish consumption, oily fish consumption, healthy food consumption, and vegetable consumption) and 3 metabolic biomarkers (the ratio of 22:6 docosahexaenoic acid to total fatty acids, the ratio of omega-3 fatty acids to total fatty acids, and the ratio of polyunsaturated fatty acids to total fatty acids). Among these 12 combinations, 11 demonstrated significant mediation effects with P-values less than 0.05. Notably, the mediation effect of the ratio of 22:6 docosahexaenoic acid to total fatty acids in the relationship between oily fish consumption and sensorineural hearing loss (SNHL) was particularly pronounced, with a mediation proportion of 10.57%. This indicates that approximately 10.57% of the protective effect of oily fish consumption on SNHL can be attributed to changes in this specific metabolic biomarker. Additionally, the mediation effect involving the ratio of omega-3 fatty acids to total fatty acids was also significant, underscoring the potential pathways through which dietary factors, especially oily fish consumption, influence hearing health.

Table 1. Mediation Analysis of Dietary Choices, Circulating Metabolic Biomarkers, and Sensorineural Hearing Loss (SNHL)

Notes: Total Effect (β): Total effect of the exposure on the outcome. Direct Effect (β): Direct effect of the exposure on the outcome, excluding the mediator's effect. Mediation Effect (β) (A × B): Indirect effect of the exposure on the outcome through the mediator. Mediation Proportion (95% CI): Proportion of the total effect that is mediated by the circulating metabolic biomarker, with a 95% confidence interval. SNHL: Sensorineural Hearing Loss

3.6Sensitivity Analysis Results

For all associations, the MR-Egger intercept test excluded the notion of horizontal pleiotropy, indicating no pleiotropic bias. The MR-Egger intercept test showed no significant evidence of directional pleiotropy (P values > 0.05), suggesting minimal influence of genetic pleiotropy on the IVs' impact on sensorineural hearing loss (SNHL) through dietary choices and circulating metabolic biomarkers. Cochran's Q test was used to evaluate the heterogeneity of the IVs. Notably, the association between oily fish consumption and SNHL indicated heterogeneity. Therefore, we used the random effects model of IVW in the MR analysis for oily fish consumption and SNHL to account for this heterogeneity (Table 2).

The leave-one-out analysis confirmed that no single SNP significantly influenced the overall results, reinforcing the stability of our findings. The MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) test was used to evaluate the potential pleiotropic effects of the selected SNPs for positive results, and none of the SNPs showed indications for removal after this evaluation.

These sensitivity analyses collectively ensure the reliability and robustness of our causal inferences between dietary choices, circulating metabolic biomarkers, and SNHL. Specific results of the leave-one-out analysis and MR-PRESSO can be found in Table S4 and Supplementary file.

Table 2. Sensitivity Analysis of Exposures and Sensorineural Hearing Loss (SNHL)

Notes: Pleiotropy: MR-Egger intercept: Intercept value indicating pleiotropy. pval: P-value for MR-Egger intercept. A p-value greater than 0.05 indicates no horizontal pleiotropy. Heterogeneity: IVW: Cochran’s Q P-value using the Inverse Variance Weighted method. A p-value greater than 0.05 indicates no heterogeneity.MR-Egger: Cochran’s Q P-value using the MR-Egger method. A p-value greater than 0.05 indicates no heterogeneity.

This study utilized Mendelian Randomization (MR) to identify significant causal relationships between specific dietary choices and sensorineural hearing loss (SNHL). Fish consumption, oily fish consumption, healthy food consumption, and vegetable consumption were all found to be protective against SNHL. Mediation analysis revealed that the ratios of 22:6 docosahexaenoic acid (DHA) to total fatty acids, omega-3 fatty acids to total fatty acids, and polyunsaturated fatty acids (PUFAs) to total fatty acids significantly mediated these protective effects. For example, approximately 10.57% of the protective effect of oily fish consumption on SNHL was mediated by the ratio of DHA to total fatty acids. These findings highlight the critical role of dietary choices and specific metabolic biomarkers in maintaining hearing health and preventing SNHL.

Fish, particularly oily fish, provides high-quality protein and Omega-3 fatty acids (EPA and DHA), which improve cardiovascular health by lowering triglycerides, blood pressure, and inflammation markers, indirectly supporting hearing health by enhancing blood flow to the inner ear ³⁴. Omega-3 fatty acids also offer neuroprotective and anti-inflammatory properties ³⁵. Healthy foods, rich in fiber, vitamins, and minerals, reduce the risk of chronic diseases such as cardiovascular diseases and diabetes, which are linked to hearing loss ³⁶. Essential nutrients like magnesium and zinc in these foods are crucial for preserving ear health ³⁷. Vegetables offer dietary fiber, vitamins, minerals, and antioxidants, protecting ear cells from oxidative stress and maintaining ear function ⁷. These findings emphasize the importance of a balanced diet rich in fish, healthy foods, and vegetables in promoting hearing health and preventing SNHL.

Our study also examined the mediating effects of three specific circulating metabolic biomarkers: the ratio of DHA to total fatty acids, the ratio of omega-3 fatty acids to total fatty acids, and the ratio of PUFAs to total fatty acids. DHA is critical for brain health, visual protection, cardiovascular health, and anti-inflammatory effects ³⁸. The ratio of omega-3 fatty acids to total fatty acids is a key indicator of diet quality, associated with cardiovascular benefits, anti-inflammatory properties, improved metabolic syndrome profiles, and potential mental health benefits ³⁹. The ratio of PUFAs to total fatty acids reflects dietary health, offering protection against cardiovascular diseases, inflammation, and cognitive decline ⁴⁰.

These biomarkers significantly mediate the relationship between dietary choices and SNHL. The ratio of DHA to total fatty acids, particularly in the context of fish and oily fish consumption, showed a notable mediation effect. This aligns with previous research highlighting DHA’s role in reducing inflammation and protecting against cellular damage, crucial for maintaining hearing health ^5,41–43. Similarly, the omega-3 fatty acids ratio demonstrated significant mediation, supporting the hypothesis that these fatty acids improve inner ear blood flow and reduce inflammation, preserving auditory function.

The findings from this study have significant implications for public health and clinical practice. Promoting healthy dietary choices, such as consuming fish (particularly oily fish) and a balanced diet rich in vegetables and healthy foods, can serve as practical preventive measures against SNHL. Public health initiatives should emphasize these dietary recommendations to enhance hearing health.

In clinical practice, identifying specific circulating metabolic biomarkers, such as the ratios of DHA, omega-3 fatty acids, and PUFAs to total fatty acids, offers valuable insights into the biological pathways through which diet influences hearing ⁴³. These biomarkers can guide personalized dietary interventions aimed at preserving hearing function, particularly for individuals at higher risk of SNHL.

This study's strengths include the use of MR to infer causal relationships, leveraging genetic variations as instrumental variables to reduce confounding factors and mitigate reverse causation. Utilizing existing large-scale genome-wide association studies (GWAS) data provides a cost-effective and efficient analysis. However, several limitations must be acknowledged. The mediation effects identified, while significant, were relatively small, indicating that the dietary impact on SNHL might involve more complex pathways ^13,44. The majority of GWAS data were from European populations, limiting the generalizability of our findings to other ethnic groups. The lack of detailed demographic information, such as age and sex, restricted the possibility of conducting more granular subgroup analyses ^14,45.

Future research should include more diverse populations to enhance the generalizability of the results. Expanding studies to encompass various ethnic backgrounds can provide a broader understanding of how genetic variations impact the relationships between diet, metabolic biomarkers, and hearing health. Longitudinal studies are also needed to explore the long-term effects of dietary interventions on SNHL and to identify potential changes in metabolic biomarkers over time. Further investigations should focus on elucidating the biological mechanisms linking dietary factors and metabolic biomarkers to SNHL. Experimental studies and clinical trials can offer more concrete evidence to support the causal relationships suggested by our MR analyses. Additionally, interdisciplinary collaboration between geneticists, nutritionists, and audiologists can provide valuable insights and enhance the robustness of future research.

In conclusion, this study highlights significant causal pathways between dietary choices, circulating metabolic biomarkers, and SNHL. The protective effects of certain diets on hearing health are partly mediated through specific metabolic pathways, such as the ratios of DHA and omega-3 fatty acids to total fatty acids. These findings suggest that dietary modifications and nutritional interventions could be effective preventive strategies against SNHL. Despite some limitations, this research provides valuable insights into the complex interplay between diet, metabolism, and hearing, paving the way for future studies to further explore these associations and develop targeted approaches for hearing loss prevention.

Author Contributions

The study was designed by HY and CB. Statistical analyses were performed by HY. The manuscript was written by HY and YL. All authors contributed to the interpretation of data and commented on the manuscript. All authors contributed tog the article and approved the submitted version.

Informed Consent Statement

Patient consent was not required as the research utilized publicly available GWAS summary statistics data, which had already obtained informed consent from all participating studies in accordance with approved protocols by their respective institutional re-view boards.

Data Availability Statement

All data used in this study are available in the public repository. The code involved in the data analysis process can be obtained by contacting the corresponding author.

Acknowledgments

We thank the teams of the GWAS Catalog, FinnGen, and UK Biobank for making the summary data publicly available. We would also like to acknowledge the principal investigators of the studies for their contributions and for making their data accessible for this research.

Conflicts of Interest

The authors declare no competing interests.

Payne, T. & Wong, G. Hearing loss: Conductive versus sensorineural. InnovAiT 15, 218–225 (2022).
Liu, F., Han, B., Zhou, X., Huang, S. & Huang, J. Research progress on the treatment and nursing of sensorineural hearing loss. Frontiers in Neuroscience 17, 1199946 (2023).
Prevalence and risk factors of sensorineural hearing loss in patients with systemic hypertension. International Journal of Otorhinolaryngology and Head and Neck Surgery 8, 517–517 (2022).
Gopineti, L., Paulpillai, M., Rosenquist, A. & Van Bergen, A. H. Prevalence of sensorineural hearing loss in children with palliated or repaired congenital heart disease. Cureus 12, (2020).
Martinez-Vega, R. et al. Long-term omega-3 fatty acid supplementation prevents expression changes in cochlear homocysteine metabolism and ameliorates progressive hearing loss in C57BL/6J mice. The Journal of Nutritional Biochemistry 26, 1424–1433 (2015).
Alvarado, J. C., Fuentes-Santamaría, V., Gabaldón-Ull, M. C. & Juiz, J. M. An oral combination of vitamins A, C, E, and Mg++ improves auditory thresholds in age-related hearing loss. Frontiers in Neuroscience 12, 527 (2018).
Péneau, S. et al. Intake of specific nutrients and foods and hearing level measured 13 years later. British journal of nutrition 109, 2079–2088 (2013).
Dawes, P., Cruickshanks, K. J., Marsden, A., Moore, D. R. & Munro, K. J. Relationship between diet, tinnitus, and hearing difficulties. Ear and hearing 41, 289–299 (2020).
Wang, X., Gao, Y. & Jiang, R. Diagnostic and predictive values of serum metabolic profiles in sudden sensorineural hearing loss patients. Frontiers in Molecular Biosciences 9, 982561 (2022).
Miao, L., Wang, B., Zhang, J., Yin, L. & Pu, Y. Plasma metabolomic profiling in workers with noise-induced hearing loss: a pilot study. Environ Sci Pollut Res 28, 68539–68550 (2021).
Zhang, X. et al. Plasma metabolomics analyses highlight the multifaceted effects of noise exposure and the diagnostic power of dysregulated metabolites for noise-induced hearing loss in steel workers. Frontiers in Molecular Biosciences 9, 907832 (2022).
Llano, D. A., Issa, L. K., Devanarayan, P., Devanarayan, V. & Initiative (ADNI), A. D. N. Hearing loss in Alzheimer’s disease is associated with altered serum lipidomic biomarker profiles. Cells 9, 2556 (2020).
Causal Associations Between Dietary Habits and Cardiovascular Disease: A Mendelian Randomization Study. British Journal of Nutrition 1–21 (2023) doi:10.1017/s000711452300140x.
Sanderson, E. et al. Mendelian randomization. Nat Rev Methods Primers 2, 6 (2022).
Navarese, E. P. et al. Independent Causal Effect of Remnant Cholesterol on Atherosclerotic Cardiovascular Outcomes: A Mendelian Randomization Study. ATVB 43, (2023).
Cai, D. et al. A causal relationship between irritability and cardiovascular diseases: a Mendelian randomization study. Frontiers in Cardiovascular Medicine 10, 1174329 (2023).
Burgess, S. & Thompson, S. G. Multivariable Mendelian randomization: the use of pleiotropic genetic variants to estimate causal effects. American journal of epidemiology 181, 251–260 (2015).
Sanderson, E., Davey Smith, G., Windmeijer, F. & Bowden, J. An examination of multivariable Mendelian randomization in the single-sample and two-sample summary data settings. International journal of epidemiology 48, 713–727 (2019).
Pirastu, N. et al. Using genetic variation to disentangle the complex relationship between food intake and health outcomes. Preprint at https://doi.org/10.1101/829952 (2019).
Karjalainen, M. K. et al. Genome-wide characterization of circulating metabolic biomarkers. Nature 628, 130–138 (2024).
Kurki, M. I. et al. FinnGen: Unique genetic insights from combining isolated population and national health register data. MedRxiv 2022–03 (2022).
Burgess, S., Thompson, S. G. & Collaboration, C. C. G. Avoiding bias from weak instruments in Mendelian randomization studies. International journal of epidemiology 40, 755–764 (2011).
PhenoScanner, V. A database of human genotype-phenotype associations. (2021).
EPIC- InterAct Consortium et al. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. Eur J Epidemiol 30, 543–552 (2015).
Burgess, S., Butterworth, A. & Thompson, S. G. Mendelian Randomization Analysis With Multiple Genetic Variants Using Summarized Data. Genetic Epidemiology 37, 658–665 (2013).
Bowden, J., Davey Smith, G. & Burgess, S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. International journal of epidemiology 44, 512–525 (2015).
Bowden, J., Davey Smith, G., Haycock, P. C. & Burgess, S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genetic Epidemiology 40, 304–314 (2016).
Hartwig, F. P., Davey Smith, G. & Bowden, J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. International journal of epidemiology 46, 1985–1998 (2017).
Verbanck, M., Chen, C.-Y., Neale, B. & Do, R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nature genetics 50, 693–698 (2018).
Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series B: Statistical Methodology 57, 289–300 (1995).
Shahidi, F. & Ambigaipalan, P. Omega-3 Polyunsaturated Fatty Acids and Their Health Benefits. Annu. Rev. Food Sci. Technol. 9, 345–381 (2018).
Zhang, Y.-P. et al. DHA, EPA and their combination at various ratios differently modulated Aβ25-35-induced neurotoxicity in SH-SY5Y cells. Prostaglandins, Leukotrienes and Essential Fatty Acids 136, 85–94 (2018).
TANAKA, Y. et al. A stronger suppression of ear swelling that was sensitized with 2, 4-dinitro-1-fluorobenzene was observed in mice fed docosahexaenoic acid (DHA) enriched phospholipid diets than those fed DHA enriched triacylglycerol diets. Journal of Japan Oil Chemists’ Society 49, 75–79 (2000).
Swanson, D., Block, R. & Mousa, S. A. Omega-3 fatty acids EPA and DHA: health benefits throughout life. Advances in nutrition 3, 1–7 (2012).
Simopoulos, A. P. The importance of the ratio of omega-6/omega-3 essential fatty acids. Biomedicine & pharmacotherapy 56, 365–379 (2002).
Tuomilehto, J. et al. Prevention of Type 2 Diabetes Mellitus by Changes in Lifestyle among Subjects with Impaired Glucose Tolerance. N Engl J Med 344, 1343–1350 (2001).
Role of Magnesium in Health and Disease. Journal of Experimental Sciences 4, 32–43 (2013).
Lin, P. Y. et al. Omega-3 fatty acids and blood-based biomarkers in Alzheimer’s disease and mild cognitive impairment: A randomized placebo-controlled trial. Brain, Behavior, and Immunity 99, 289–298 (2022).
Huerta, A. E., Laiglesia, L. M., Martínez-Fernández, L. & Moreno-Aliaga, M. J. Role of Omega-3 Fatty Acids in Metabolic Syndrome. in Omega-3 Fatty Acids (eds. Hegde, M. V., Zanwar, A. A. & Adekar, S. P.) 189–202 (Springer International Publishing, Cham, 2016). doi:10.1007/978-3-319-40458-5_14.
Ilse, A. & Amanda, K. Impact of DHA on Metabolic Diseases from Womb to Tomb. Marine Drugs 12, 6190–6212 (2014).
A High-Fat Diet Induces Low-Grade Cochlear Inflammation in CD-1 Mice. International Journal of Molecular Sciences 23, 5179–5179 (2022).
Associations between diet and both high and low pure tone averages and transient evoked otoacoustic emissions in an older adult population-based study. Journal of The American Academy of Audiology 22, 49–58 (2011).
Biomarkers of fish oil omega-3 polyunsaturated fatty acids intake in humans. Nutrition in Clinical Practice 29, 63–72 (2014).
What can Mendelian randomization contribute to biological anthropology? American journal of biological anthropology (2023) doi:10.1002/ajpa.24750.
Unbiased Estimation and Asymptotically Valid Inference in Multivariable Mendelian Randomization with Many Weak Instrumental Variables. (2023).

Table 1 and 2 are available in the Supplementary Files section.

No competing interests reported.

SupplementaryMaterialslegends.docx
Supplementaryfile.zip
TableS1.ImpactofDietaryChoicesonSensorineuralHearingLoss.xlsx
TableS2.ImpactofCirculatingMetabolicBiomarkersonSensorineuralHearingLoss.xlsx
TableS3.ReverseMendelianRandomizationAnalysis.xlsx
TableS4.pressoresultsummary.xlsx
TableS5STROBEMRchecklist.docx
Table1.MediationAnalysisofDietaryChoicesCirculatingMetabolicBiomarkersandSensorineuralHearingLoss.xlsx
Table 1. Mediation Analysis of Dietary Choices, Circulating Metabolic Biomarkers, and Sensorineural Hearing Loss (SNHL) Notes: Total Effect (β): Total effect of the exposure on the outcome. Direct Effect (β): Direct effect of the exposure on the outcome, excluding the mediator's effect. Mediation Effect (β) (A × B): Indirect effect of the exposure on the outcome through the mediator. Mediation Proportion (95% CI): Proportion of the total effect that is mediated by the circulating metabolic biomarker, with a 95% confidence interval. SNHL: Sensorineural Hearing Loss
Table2.SensitivityAnalysisofExposuresandSensorineuralHearingLoss.xlsx
Table 2. Sensitivity Analysis of Exposures and Sensorineural Hearing Loss (SNHL) Notes: Pleiotropy: MR-Egger intercept: Intercept value indicating pleiotropy. pval: P-value for MR-Egger intercept. A p-value greater than 0.05 indicates no horizontal pleiotropy. Heterogeneity: IVW: Cochran’s Q P-value using the Inverse Variance Weighted method. A p-value greater than 0.05 indicates no heterogeneity.MR-Egger: Cochran’s Q P-value using the MR-Egger method. A p-value greater than 0.05 indicates no heterogeneity.

Download PDF

Reviewers invited by journal
20 Sep, 2024
Editor assigned by journal
18 Sep, 2024
Editor invited by journal
27 Aug, 2024
Submission checks completed at journal
23 Aug, 2024
First submitted to journal
09 Aug, 2024

You are reading this latest preprint version

Diet, Metabolic Biomarkers, and Sensorineural Hearing Loss: A Mendelian Randomization and Mediation Analysis

Status:

Version 1

Abstract

Figures

1 Introduction

2 Materials and methods

2.2.1 Dietary Choices

2.2.2 Circulating Metabolic Biomarkers

2.2.3 Sensorineural Hearing Loss (SNHL)

2.4.1 Two-sample Mendelian Randomization Analysis

2.4.3 Reverse Mendelian Randomization Analysis

2.4.3 Mediation Analysis

2.4.4 Sensitivity Analysis

3 Results

4 Discussion

5 Conclusion

Declarations

References

Table 1 and 2

Additional Declarations

Supplementary Files

Status:

Version 1