Introduction
The World Health Organization estimated that 466 million people, or 6.1 per cent of the global population, were living with disabling hearing loss (HL) in 2018, and this number is expected to increase as the population ages rapidly.1 In 2019, World Health Organization estimated the annual global cost of HL to be $750 billion. This not only causes considerable economic losses but also seriously reduces the quality of human life. HL has recently been ranked as the fifth leading cause of years lived with a disability.Reference Stika and Hays2 Many large epidemiological studies have found that HL is independently associated with falls, isolation, cognitive decline, dementia, anxiety, depression, social isolation, increased rates of hospitalization and healthcare use.Reference Jiam, Li and Agrawal3–Reference Blazer and Tucci6 This emphasizes the importance of early detection to potentially delay or prevent its onset. The pathogenesis of HL involves various factors, including microcirculation disorders, viral infections such as rubella and measles, head trauma, genetic factors, noise exposure, autoimmune diseases and so forth. These factors can lead to damage or degeneration of the hair cells in the cochlea, disruption of the auditory nerve pathways or impairment of the auditory processing in the brain, ultimately resulting in a decrease in hearing function.Reference Chandrasekhar, Tsai Do, Schwartz, Bontempo, Faucett and Finestone7 Multiple studies have shown that the dietary intake of carbohydrates, cholesterol, fiber, protein, sugar, fruits, vegetables, saturated fats and trans-fats is associated with self-reported HL and requires public health interventions to prevent HLReference Rodrigo, Campos-Asensio, MÁ, Crespo and Olmedillas8–Reference Chen, Tan, Wu and Liu11; however, comprehensive guidelines on healthy eating are lacking. As the concept of dietary patterns continues to develop and evolve, the Healthy Eating Index (HEI) is widely used in different types of research, including surveillance, epidemiological and intervention studies, involving different populations in the United States. It can be indexed to guide the diet of Americans for early prevention of HL.Reference Schap, Kuczynski and Hiza12 The HEI is a measure that assesses whether a group of foods meets the Dietary Guidelines for Americans, which measures diet quality rather than quantity; that is, it assesses density rather than absolute quantity, providing a guide to the overall diet.Reference Krebs-Smith, Pannucci, Subar, Kirkpatrick, Lerman and Tooze13 There were no changes in the components or standards between HEI-2015 and HEI-2020.Reference Shams-White, Pannucci, Lerman, Herrick, Zimmer and Mathieu14 The relation between the HEI and HL is an important topic in healthcare and public health. In this context, by conducting a retrospective cross-sectional study involving 5,171 U.S. adults from the National Health and Nutrition Examination Survey (NHANES), we sought to clarify the relation between the HEI and HL in adults.
Materials and methods
This cross-sectional study utilized NHANES data from 1999 to 2012 and 2015 to 2018, conducted by the Centers for Disease Control and Prevention.15 The purpose of the NHANES was to assess the health and nutritional status of non-hospitalized Americans. NHANES collects demographic and in-depth health information through home visits, screenings and laboratory tests at mobile screening centers. The NHANES was authorized by the National Center for Health Statistics Ethics Review Committee, and all participants provided written informed consent prior to participation. The secondary analysis did not require additional approval from the Institutional Review Board.16 Data from NHANES may be obtained through the NHANES website (visited on March 1, 2022; http://www.cdc.gov/nchs/nhanes.htm). Individuals aged over 20 years who completed the interviews participated in our study. We excluded individuals who lacked data on HL, HEI, or covariates. The total sample size of adults evaluated was N = 49,312, the details of which are described in Figure 1. Only publicly available data were used in the analysis, and ethical approval was not required for this study.
Figure 1. The study’s flow diagram.
The HEI assesses diet quality and refers to how well a group of foods meets the Dietary Guidelines for Americans.17 HEI-2020 contains 13 components and scoring criteria, and this is the first time that there has been no change in the index; the score of HEI-2020 is the same as that of HEI-2015. Krebs-Smith et al. described the details of each component in the HEI-2015 update.Reference Krebs-Smith, Pannucci, Subar, Kirkpatrick, Lerman and Tooze13 The HEI-2015 score ranges from 0 to 100 and is graded as follows: Grade A (highest score): Total score between 90 and 100 points, indicating that an individual’s dietary nutrient intake is very consistent with dietary guidelines. Grade B: Total score between 80 and 89 points, indicating that an individual’s dietary nutrient intake is somewhat consistent with dietary guidelines. Grade C: Total score between 60 and 79 points, indicating that an individual’s dietary nutrient intake has room for improvement, but can still adopt a healthy eating pattern. Grade D: Total score below 60 points, indicating that an individual’s dietary nutrient intake needs significant adjustment to improve dietary quality. Grade F: Total score between 0 and 59 points, indicating that an individual’s dietary nutrient intake is severely inadequate and requires special attention and improvement. A higher score reflects better diet quality. We used Day 1 Total Nutrient Intake to calculate the 13 components of HEI-2015.
All hearing tests were performed by a trained examiner on candidates aged 20–69 years in a dedicated soundproof room at a mobile test center. Hearing thresholds were tested in both ears at seven frequencies (500, 1000, 2000, 3000, 4000, 6000 and 8000 Hz), with observed values ranging between -10 and 120 dB. The pure tone average (PTA) of speech-frequency in both ears was calculated as the average of the hearing thresholds at 0.5, 1, 2 and 4 kHz. Good ear PTA is a continuous variable; the higher the value, the worse the hearing. Low-frequency PTA calculations used hearing thresholds at 0.5, 1 and 2 kHz in the better ears, while high-frequency PTA calculations used hearing thresholds at 4, 6 and 8 kHz in the better ears. All hearing thresholds were reported as dB HL. Sensitivity analysis was performed using PTA with the bad ear rather than the good ear. In addition, good ear PTAs were classified according to clinical cutoff points defined by the 1997 World Health Organization, where a hearing level less than or equal to 25 dB HL indicates normal hearing and >25 dB HL indicates HL.18 As defined by the World Health Organization in 2021, HL is defined as the speech-frequency of good ear PTA greater than or equal to 20 dB, and less than 20 dB is normal hearing.Reference Jiang, Spira, Reed, Lin and Deal19
Multiple potential covariates were evaluated based on literature,Reference Jung, Kim, Cho, Jeon and Song20–Reference Scinicariello, Przybyla, Carroll, Eichwald, Decker and Bresse22 including age, sex, marital status, race/ethnicity, education level, household income, smoking status, physical activity, high blood pressure, diabetes mellitus (DM), history of cardiovascular disease (CVD), body mass index (BMI) and noise exposure. Race/ethnicity was classified as Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black or other races, including multiple races. Marital status was divided into married, never married, living with a partner, widowed, divorced or separated. The level of education was classified as less than 9th grade, 9–11th grade (including 12th grade with no diploma), high school graduate/General Educational Development or equivalent, college or Associate of Arts (AA) degree and college graduate or above. Smoking status, as defined in previous literature, was classified as never smoked (smoking less than 100 cigarettes), current smoker or former smoker (quitting after smoking more than 100 cigarettes). Physical activity was classified as sedentary, moderate (at least 10 min of exercise in the past 30 days that resulted in only slight sweating or a mild-to-moderate increase in breathing or heart rate) and vigorous (at least 10 min of exercise in the past 30 days that resulted in heavy sweating or an increase in breathing or heart rate). Previous medical conditions (including hypertension, DM, stroke and coronary heart disease) were determined based on questions in the questionnaire regarding whether the doctor had been informed of the condition in the past. BMI was calculated using a standardized technique based on weight and height. Noise exposure included exposure to gun noise outside the workplace, noise outside the workplace, working in a noisy environment and exposure to loud noise at work.
For statistical analysis, categorical variables were expressed as proportions (per centages), and continuous variables were expressed as mean (standard deviation) or median (interquartile distance). To compare the differences between groups, one-way analysis of variance (ANOVA) (normal distribution), Kruskal–Wallis test (skewed distribution) and chi-squared test (categorical variables) were performed. After adjusting for all covariates, linear regression was used to describe the relation between the healthy diet index and hearing. Logistic regression models were used to determine the odds ratios (ORs) and 95 per cent confidence intervals (CIs) between HEI and HL. Model 1 was adjusted for sociodemographic characteristics, including age, sex, race, marital status, household income and educational level. Model 2 was adjusted for sociodemographic characteristics, smoking status, alcohol consumption, physical activity and BMI. Model 3 was comprehensively adjusted for sociodemographic characteristics, smoking status, alcohol consumption, physical activity, BMI, CVD, hypertension, diabetes mellitus (DM) and noise exposure.
Furthermore, potential changes in the relation between healthy dietary index and HL were assessed, including the following variables: age (20–65 years vs >65 years), sex, marital status (married vs. never married, living with a partner or other [widowed, divorced or separated individuals]) and BMI (<25 vs. 25–30, ≥30 kg/mReference Stika and Hays2). Multivariate logistic regression was used to assess the heterogeneity among the subgroups, and the interaction between the subgroups and the healthy diet index was examined using the likelihood ratio test.
All analyses were performed using the statistical package R 4.3.1 (http://www.R-project.org; R Foundation, Shanghai, China) (accessed on March 10, 2024) and Free Statistics software version 1.9. Descriptive statistics were calculated for all participants. In the bilateral test, statistical significance was set at p less than 0.05.
Results
Study population
A total of 49,312 participants, all aged 20 years and older, completed interviews. We excluded those with missing data on the HEI-2015 (n = 8,162), those with missing data on HL (n = 32,422) or those with covariates (n = 3,557). Ultimately, this cross-sectional study includes 5,171 participants from the NHANES between 1999 and 2012, and 2015 and 2018. The detailed inclusion and exclusion processes are presented in Figure 1.
Baseline characteristics
Table 1 illustrates the baseline characteristics of all participants according to the HEI-2015. We included a total 5,171 patients with a mean age of 50.9 plus or minus 18.5 years, among whom 60.9 per cent were non-Hispanic White and 52.9 per cent were men. HL was defined according to speech-frequency PTA in the better-hearing ear, with greater than 25 dB HL and greater than or equal to 20 dB HL as having HL. The overall prevalence rates of HL in the study population were 25.0 per cent and 37.6 per cent, respectively.
Table 1. Population characteristics by categories of the HEI-2015
BMI = body mass index; CVD = cardiovascular disease; DM = diabetes mellitus; GED = General Educational Development; HEI-2015 = Healthy Eating Index 2015; IQR = interquartile range; PIR = poverty income ratio.
The group with a higher HEI was observed to have a higher proportion of men; be non-Hispanic White and married; have a higher family poverty income ratio and educational level; be never smokers and mild alcohol users; have a lower BMI, greater physical activity, a lower incidence of stroke and history of CVD, no hypertension, no DM, no noise and normal hearing.
Relation between the HEI-2015 and HL
Univariate analysis demonstrated that age, sex, race, marital status, family income, educational level, smoking status, alcohol consumption, physical activity, BMI, coronary heart disease, hypertension, DM and noise exposure were associated with low-frequency HL, speech-frequency HL and high-frequency HL (see Supplementary Material Tables 4, 5 and 6, respectively). In linear multifactor analysis (Table 2), after adjusting for potentially confounding factors, there was a significant negative association between the HEI-2015 and low-frequency hearing, with a coefficient of -0.03 (95 per cent CI: -0.05 to -0.01; p = 0.001); speech-frequency hearing, with a coefficient of -0.04 (95 per cent CI: -0.06 to -0.01; p = 0.003); and high-frequency hearing, with a coefficient of -0.05 (95% CI: -0.08 to -0.02; p = 0.003). When the HEI-2015 was a continuous variable, after adjusting for potential confounders, there was a significant negative association between the HEI-2015 and low-frequency HL (Table 3), with OR values of 0.99 (95 per cent CI: 0.98–0.99; p < 0.001) and 0.98 (95 per cent CI: 0.98–0.99; p < 0.001). The OR was 0.99 in speech-frequency HL, defined as HL greater than 25 dB (95 per cent CI: 0.98–1; p = 0.008), and the OR was 0.99 (95 per cent CI: 0.98–1; p = 0.006) in high-frequency HL, defined as HL greater than or equal to 20 dB. When the HEI-2015 was analyzed using quintiles (see Supplementary Material Table 7), a significant negative association was found between the HEI-2015 and HL after adjusting for potential confounders. The adjusted OR values of HEI-2015 and F-scores (0–59), D(60–69), C(70–79), B(80–89) and A(90–100) for HL are shown in Supplementary Material Table 7. With hearing less than 20 dB as the normal value, the OR value of low-frequency hearing in B(80–89) was 0.54 (95 per cent CI: 0.34–0.88; p = 0.013). The OR values of speech-frequency band listening were 0.7 (95 per cent v CI: 0.53–0.93; p = 0.013) in C(70–79) and 0.14 (95% CI: 0.03–0.56; p = 0.005) in A(90–100). The OR values for high-frequency hearing were 0.76 (95 per cent CI: 0.61–0.96; p = 0.02) in D(60–69) and 0.66 (95 per cent CI: 0.47–0.91; p = 0.012) in C(70–79), and the OR value in B(80–89) was 0.52 (95 per cent CI: 0.3–0.91; p = 0.021). However, when hearing was less than or equal to 25 decibels, the OR value of low-frequency hearing was 0.7 (95 per cent CI: 0.51–0.96; p = 0.029) in C(70–79) and 0.54 (95 per cent CI: 0.3–0.96; p = 0.035) in B(80–89).
Table 2. Association between the HEI-2015 and hearing
CI = confidence interval; HEI-2015 = Healthy Eating Index 2015.
Table 3. Association between the HEI-2015 and hearing loss
BMI = body mass index; CI = confidence interval; CVD = cardiovascular disease; DM = diabetes mellitus; OR = odds ratio; PIR = poverty income ratio; Ref: reference.
Model 1 was adjusted for sociodemographic variables (age, sex, race, marital status, PIR, education). Model 2 was adjusted for sociodemographic variables (age, sex, race, marital status, PIR, education), smoking status, alcohol, physical activity, BMI. Model 3 was adjusted for sociodemographic variables (age, sex, race, marital status, PIR, education), smoking status, alcohol consumption, physical activity, BMI, CVD, hypertension, DM, noise.
Stratified analyses based on additional variables
Stratified analysis was performed on several subgroups to assess the potential impact of the relation between the HEI-2015 and HL. When stratified by sex, age, BMI and marital status, no significant interactions were observed between the subgroups. Given the sample size, a p-value of less than 0.05 for marriage and BMI may not be statistically significant (see Supplementary Material Figures 2, 3 and 4).
Discussion
We found that the HEI scores were associated with HL, and subgroup analyses showed stable results.
Spankovich’s findings support the link between healthier diets and lower high-frequency thresholds in adults.Reference Spankovich and Le Prell23 Adherence to healthy eating patterns has been linked to a reduced risk of HL in women.Reference Curhan, Wang, Eavey, Stampfer and Curhan24 An Australian cross-sectional study found that the overall diet quality was associated with concurrent vision and HL.Reference Gopinath, Schneider, Flood, McMahon, Burlutsky and Leeder25 Previous studies have linked a healthy diet to HL, which is similar to our findings.
Some studies have suggested that dietary supplement use is positively associated with hearing improvement at all frequencies.Reference Kang, Choi, Kim and Choi26 Studies by Rosenhall et al. have shown that fish are beneficial for hearing, while eating “junk food” rich in low-molecular-weight carbohydrates is detrimental.Reference Rosenhall, Idrizbegovic, Hederstierna and Rothenberg27 A diet high in saturated fat, cholesterol and carbohydrates is a risk factor for HL, and conversely, increasing antioxidants in the form of protein, zinc, magnesium, selenium, iron, iodine, fruits, vegetables, polyunsaturated fatty acids (omega-3) and vitamins A, C and E can prevent HL from developing.Reference Rodrigo, Campos-Asensio, MÁ, Crespo and Olmedillas8, Reference Kim, Sim, Kim and Choi28–Reference Jung, Kim and Yeo30 Ji Eun Choi et al. believe that a high intake of seeds, nuts, fruits, seaweed and vitamin A has protective effects on hearing, and dietary antioxidants or anti-inflammatory foods may help reduce the occurrence of HL.Reference Choi, Ahn and Moon31 Xinmin Wei believes that the dietary intake of magnesium and calcium is associated with a lower risk of HL.Reference Wei32 Many scholars have studied the relation between diet and HL; some diets are protective factors, whereas others are risk factors.
Most previous studies on diet and hearing have focused on the analysis of single nutrients, with the limitation that nutrient intake is correlated, making it difficult to isolate the effects of one nutrient from those of others. Single-nutrient analyses, which do not consider biochemical interactions between nutrients, also increase the possibility of false-positive associations.Reference Dawes, Cruickshanks, Marsden, Moore and Munro33 As a supplement to single-nutrient analysis, dietary pattern analysis can better capture the synergistic and cumulative effects of total dietary intake on health outcomes.Reference McGuire34
The HEI assesses whether a group of foods meets the American Dietary Guidelines for Americans. The HEI can be applied to any group of foods by assigning a score to each component and comparing its density to the relevant criteria. HEI scores have been widely used in numerous studies on diet quality in populations, the relation between diet quality and health outcomes, the impact of interventions on diet quality and economic and food-context-based studies.Reference Krebs-Smith, Pannucci, Subar, Kirkpatrick, Lerman and Tooze13 The HEI includes eating more fruits, vegetables, legumes, whole grains, nuts, dairy products, seafood, plant proteins and fatty acids, and less saturated fat, sugar and sodium. There are many mechanisms by which this healthy eating pattern prevents HL, including the prevention of microvascular and macrovascular damage to cochlear blood flow, inhibition of oxidative damage and reduction of inflammation.Reference Curhan, Wang, Eavey, Stampfer and Curhan24 A lower high-fat diet in a high-quality diet may reduce induced oxidative stress, mitochondrial damage and apoptosis in the inner ear, which has a protective effect on inner ear cells.Reference Tang, Hwang, Yang, Hsu, Wu and Liur35, Reference Le Prell, Gagnon, Bennett and Ohlemiller36 Inadequate blood supply to the cochlea can lead to hypoxia and ischemic injury, oxidative stress, mitochondrial dysfunction, cellular damage and peripheral and central auditory neurodegeneration.Reference Yamasoba, Lin, Someya, Kashio, Sakamoto and Kondo37 Dietary nutrients of polyunsaturated fatty acids and multivitamins, such as vitamins B12 and C, calcium and selenium, may have antioxidant and neuroprotective effects.Reference El Soury, Fornasari, Carta, Zen, Haastert-Talini and Ronchi38 They may also protect against neurodegeneration of auditory nerve fibers and central auditory pathways.Reference Curhan, Wang, Eavey, Stampfer and Curhan24 Dietary guidance for patients with potential HL is provided by examining the effects of a healthy diet index on HL.
This study has several limitations. First, the cross-sectional design was a major limitation, and no causal relation could be inferred from this study. Second, the use of self-reported 24-hour food recall data is limited because they are prone to overestimation or underestimation. Finally, more participants were excluded because of the lack of data on any covariates that could have affected the results. Despite these limitations, our study has several strengths. The use of a large, nationally representative database to assess diet quality is a major strength of the present study.
We found an association between the HEI-2015 score and HL. Therefore, attention should be paid to the relation between dietary quality and HL. Future cohort studies or randomized controlled trials are needed to confirm this relation.
What is already known
• Hearing loss (HL) is a major global health burden linked to cognitive decline, depression and reduced quality of life, with modifiable risk factors understudied
• Prior studies suggest associations between single nutrients (e.g., vitamins, fatty acids) and HL, but evidence on overall dietary patterns remains limited
• The Healthy Eating Index (HEI) is a validated tool for assessing diet quality, yet its relationship with HL in adults is poorly characterized
What this paper adds
• First large-scale study demonstrating a significant inverse association between HEI-2015 scores and HL across low-, speech- and high-frequency thresholds in U.S. adults
• Higher HEI-2015 scores (indicating better diet quality) correlate with reduced odds of HL, particularly in quintile-based analyses (e.g., OR = 0.52 for high-frequency HL in the 80–89 score group)
• Provides actionable evidence for integrating dietary guidelines into HL prevention strategies, highlighting diet quality as a modifiable lifestyle factor
• Strengthens the rationale for future longitudinal studies to explore causal links between diet and auditory health
Conclusions
We found that higher HEI scores were associated with better hearing and that a healthy diet may also help reduce the risk of HL.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022215125000635.
Acknowledgments
The authors extend their gratitude to all NHANES participants and staff members whose dedication made this study possible. Special thanks are extended to Dr. Liu Huanxian from the Department of Neurology, Chinese PLA General Hospital (PLAGH), for his invaluable expertise in statistical methodology, consultation on study design rigor and critical review of the manuscript.
Disclosure of interest
The authors declare no conflict of interest.
Funding
This work was supported by Zhejiang Provincial Traditional Chinese Medicine Science and Technology Project (2023ZF070).
Open access
