Hearing loss is one of the most common chronic conditions in the elderly and has become a major public health issue worldwide(Reference Gratton and Vazquez1–Reference Lin, Niparko and Ferrucci3), given its consequences on quality of life(Reference Ishine, Okumiya and Matsubayashi4). In the USA, between 2001 and 2008, 17 % of women and 39 % of men aged 60–69 years suffered from hearing loss ( ≥ 25 decibels (dB))(Reference Lin, Niparko and Ferrucci3), while in France, in the period between 1998 and 2000, 22 % of men and women aged 60–74 years reported suffering from a hearing impairment(5). Hearing loss increases as a function of age, with poorer hearing in men compared with women(Reference Roth, Hanebuth and Probst2). The public health burden of age-related hearing loss is expected to increase drastically in the coming decades with the ageing of the population(Reference Gratton and Vazquez1, Reference Roth, Hanebuth and Probst2). Numerous underlying factors have been suggested regarding hearing loss, including exposure to noise and toxic substances, genetic factors, CVD, diabetes and obesity(Reference Bovo, Ciorba and Martini6–Reference Vaughan, James and McDermott9).
An independent role of nutritional factors in audition has been suggested in several studies. In a cross-sectional study involving fifty-five women aged 60–71 years, low intakes of folates and vitamin B12 and decreased serum concentrations of these vitamins were shown to be associated with impaired hearing(Reference Houston, Johnson and Nozza10). In addition, in a randomised controlled trial of folate supplementation involving 728 older men and women, the supplementation slowed the age-related decline in hearing acuity regarding speech frequencies(Reference Houston, Johnson and Nozza10, Reference Durga, Verhoef and Anteunis11). Hypotheses behind these associations involve an impact of folate and vitamin B12 on the nervous system as well as on vascular function with a potential link with homocysteine concentration(Reference Houston, Johnson and Nozza10–Reference Gopinath, Flood and Rochtchina12). The role of antioxidants in the management of hearing loss has also generated considerable interest over the past few years. A protective effect of antioxidants on several types of hearing impairment including age-related, noise-induced and drug-induced hearing loss has been suggested(Reference Darrat, Ahmad and Seidman13). Two cross-sectional studies with subjects older than 50 years showed positive associations between auditory function and intake of specific antioxidants, in particular, vitamin C(Reference Spankovich, Hood and Silver14), vitamin E(Reference Spankovich, Hood and Silver14, Reference Gopinath, Flood and McMahon15), riboflavin(Reference Spankovich, Hood and Silver14), Mg(Reference Spankovich, Hood and Silver14) and lycopene(Reference Spankovich, Hood and Silver14). However, antioxidant intake did not modify the risk of incident hearing loss(Reference Gopinath, Flood and McMahon15). Except for these two population-based studies, most other research has been experimental(Reference Darrat, Ahmad and Seidman13), and evidence of an association between antioxidant intake and hearing function is still lacking. Finally, vitamin A intake was inversely associated with hearing impairment in 50-year-old subjects, while it did not modify the risk of incident hearing loss(Reference Gopinath, Flood and McMahon15). On the other hand, retinol intake worsened hearing performance in 49- to 99-year-old subjects(Reference Spankovich, Hood and Silver14). Finally, serum retinol was inversely associated with hearing impairment in cross-sectional studies of 65-year-old adults(Reference Michikawa, Nishiwaki and Kikuchi16). Overall, the impact of retinol on hearing is unclear and has been insufficiently evaluated in the literature.
Hence, more data are needed in order to elucidate the association between micronutrient intake and hearing. Further, to our knowledge, the role of food groups containing specific micronutrients has not been investigated in the literature, apart from a recent longitudinal study showing that regular consumption of fish might protect against hearing loss in subjects aged 50 years or more(Reference Gopinath, Flood and Rochtchina17).
Identifying food groups that have an impact on hearing level (HL) could be of great importance from a public health perspective. We therefore aimed at investigating the potential association of HL with intake of specific micronutrients and food groups rich in those nutrients in a large sample of healthy volunteers.
Subjects and methods
Study population
Subjects were participants in the Supplementation with Antioxidant Vitamins and Minerals (SU.VI.MAX) and SU.VI.MAX 2 (a full list of the SU.VI.MAX 2 Research Group Members is available in Appendix 1) studies. The SU.VI.MAX study (1994–2002; n 12 741) is a randomised, double-blind, placebo-controlled, primary prevention trial initially designed to evaluate the effect of daily supplementation with antioxidant vitamins (E (30 mg), C (120 mg) and β-carotene (6 mg)) and minerals (Se (100 μg) and Zn (20 mg)) at nutritional doses on the incidence of cancer and IHD(Reference Hercberg, Galan and Preziosi18, Reference Hercberg, Galan and Preziosi19). At the end of the supplementation (2002), a total of 6850 subjects, who had agreed to participate in a post-supplementation follow-up, were included in the SU.VI.MAX 2 observational study (2007–9) which sought to investigate the impact of nutrition on the quality of ageing.
The SU.VI.MAX and SU.VI.MAX 2 studies were conducted according to the Declaration of Helsinki guidelines and were approved by the Ethics Committee for Studies with Human Subjects of Paris-Cochin Hospital (CCPPRB no. 706 and no. 2364, respectively) and the Comité National Informatique et Liberté (CNIL no. 334 641 and no. 907094, respectively). Written informed consent was obtained from all participants. The present trial was registered at clinicaltrials.gov as NCT00272428.
Dietary assessment
During the SU.VI.MAX study, subjects were invited to provide a 24 h dietary record every 2 months, for a total of six records per year. Days of the week for these records were randomised and fixed for each subject so that each day of the week and all seasons were covered. Information was collected via computerised questionnaires with the use of the Minitel Telematic Network loaded with study-specific software. The Minitel was a small terminal widely used in France as an adjunct to the telephone. A validated instruction manual(Reference Le Moullec, Deheeger and Preziosi20) was used for coding food portions. It included photographs of >250 generic items (corresponding to 1000 specific foods). Subjects could choose from three main portion sizes, two intermediate or two extreme portions, for a total of seven different portion sizes. We included participants with a minimum of six 24 h records provided during the first 2 years of follow-up (Fig. 1). Mean intake of meat as a whole and meat sub-groups (red meat, poultry and game, organ meat and processed meat), seafood as a whole and seafood sub-groups (fish and shellfish), fruits and vegetables as a whole and sub-groups (fruits and vegetables) was evaluated (g/d). A French food composition table(Reference Hercberg21) was used to calculate nutrient contents. Mean intake of retinol (μg/d), β-carotene (μg/d), folate (μg/d), vitamin B6 (mg/d), vitamin B12 (μg/d), vitamin C (mg/d) and vitamin E (mg/d) was assessed.
Fig. 1 Flow chart of subjects from the Supplementation with Antioxidant Vitamins and Minerals 2 (SU.VI.MAX 2) study cohort (2007–9) included in the present analysis.
Auditory assessment
As part of the SU.VI.MAX 2 study, all participants were invited to undergo a check-up which included a clinical examination. Only participants without auditory devices could participate in the auditory assessment, which was performed in a quiet room by trained technicians. We measured pure-tone air conduction thresholds at 0·5, 1, 2 and 4 kHz using a portable diagnostic audiometer (ST20, Maico Diagnostic GmbH), first in the right ear and then followed by the left ear. Audiometric testing relied on the automated testing mode of the audiometer and was based on ascending responses using 5 dB steps. The instruments were calibrated once a year. Participants with present or previous ear disease (7 % within the year of auditory assessment) leading to full or partial deafness, unilateral or bilateral and necessitating treatment or follow-up by a medical practitioner were excluded. In addition, as unilateral hearing loss is a pathologic ear condition unrelated to ageing, subjects with a ≥ 20 dB difference in the average pure-tone hearing thresholds between the right and left ear were excluded. Finally, in order to focus on age-related rather than genetically determined hearing loss, only participants without first-degree family history of hearing diseases or hearing loss were included (Fig. 1). Mean HL was assessed as the pure-tone average of the 0·5, 1, 2 and 4 kHz air conduction thresholds for each ear, and the value for the worse ear was retained for analyses. Impaired hearing was defined as a failure to hear a 25 dB HL signal in the better ear.
Covariate assessment
At baseline, age, sex, educational level (primary, secondary or university level), physical activity (irregular, < 1 h walking/d or ≥ 1 h walking/d) and smoking status (never smoked, former smoker or current smoker) were provided by a self-report questionnaire. At the first follow-up (1995–6), anthropometric measurements were obtained. Weight was measured with an electronic scale, with subjects wearing indoor clothing and no shoes. Height was measured under the same conditions with a wall-mounted stadiometer. BMI was calculated as the ratio of weight in kg to height in m2 (kg/m2). Blood pressure was measured by using a standardised procedure with a standard mercury sphygmomanometer. Blood pressure was taken once from each arm in subjects who had been lying down for 10 min. The mean of these two measurements was used for analyses. Subjects with missing values for BMI were excluded (Fig. 1). During follow-up, cases of cardio- or cerebrovascular disease were reviewed and validated by an independent expert committee.
Statistical methods
Included and excluded subjects were compared using the χ2 test or the non-parametric Wilcoxon rank sums test. Descriptive results from χ2 tests, non-parametric Wilcoxon rank sums tests or Student's t tests are reported as percentage, mean values and their standard deviations or geometric mean (95 % CI) across sex, as appropriate. Multivariate linear regression models were used to estimate the association between sex-specific quartiles of nutrient and food intake and HL. We adjusted for total dietary energy intake via the residual method(Reference Willett and Stampfer22) based on energy intake other than from alcohol. All analyses were adjusted for potential confounders of the association between food and nutrient intake and HL identified in the literature, i.e. age at hearing assessement, BMI, educational level, physical activity, supplementation group (intervention v. placebo), energy intake (excluding alcohol), alcohol intake and smoking habits and systolic and diastolic blood pressure. In addition to these variables, we fit a supplementary model adjusted for intake of total meat, seafood and fruits and vegetables, as appropriate. Results of the linear regression are presented as the adjusted mean difference of HL (95 % CI) in comparison to the first quartile of intake. A multivariate logistic regression was further performed to evaluate the association between food and nutrient intake and hearing loss ( ≥ 25 dB HL). In order to retain the sample size for the multivariate analyses, we performed imputation with the method of regression using BMI, age and sex to account for missing covariate data. Sensitivity analyses were performed after exclusion of subjects who developed diabetes or cardio- or cerebrovascular disease during the follow-up. All tests of statistical significance were two-sided and the type I error was set at 5 %. Statistical analyses were performed using SAS software (version 9.1, SAS Institute, Inc.).
Results
Subject characteristics
Of the 6850 adults included in the SU.VI.MAX 2 study, 1823 individuals aged 45–60 years at baseline were included in the present analysis (Fig. 1). Included subjects were older (P< 0·0001), more often male (P< 0·0001), more physically active (P< 0·01) and had a higher BMI (P< 0·0001), higher systolic and diastolic blood pressure (P< 0·0001) and lower educational level (P< 0·0001) compared with the excluded subjects. However, they had better HL (P< 0·01) than excluded subjects and showed no difference in hearing impairment (P>0·05). Participants with less than or equal to six 24 h records provided during the first 2 years of follow-up did not differ from those excluded due to fewer dietary records (P>0·05 for all characteristics).
Characteristics of the 1002 men and 821 women retained for the present analysis are presented in Table 1. Men were slightly older, had a higher BMI and higher educational level than women. In addition, they were more physically active, more likely to be former/current smokers and had higher systolic and diastolic blood pressure compared with women. Impaired hearing was greater in men than in women. Men had higher food and energy intakes than women, except for fruits. However, after adjustment for energy intake, daily consumption of fruits (P< 0·0001), vegetables (P< 0·0001) and fruit and vegetables (P< 0·0001) was higher in women than in men, while intake of organ meat and shellfish was not significantly different between the groups (P>0·05; data not tabulated). Among the assessed nutrients, men had higher intakes than women. The dietary records were similarly distributed among week days (14·0 % of records were completed on a week day) and weekend days (14·9 %), although more records were completed in the winter (29·5 %) and spring (27·1 %) than in the summer (22·0 %) or autumn (21·4 %). Median sex-specific quartiles of food and nutrient intakes are given in Table 2.
Table 1 Characteristics of the study population at baseline and at the time of the hearing assessment in the Supplementation with Antioxidant Vitamins and Minerals (SU.VI.MAX and SU.VI.MAX 2) studies, 1994–2007 (n 1823) (Mean values and standard deviations; geometric means and 95 % confidence intervals; percentages)
HL, hearing level; dB, decibels.
* P values were based on χ2 tests, Student's t tests and non-parametric Wilcoxon rank sums tests, as appropriate.
† Characteristic measured at hearing assessment.
‡ Characteristic measured at baseline.
§ Excluding energy from alcohol.
∥ Based on the log-transformed variables.
Table 2 Median sex-specific energy-adjusted quartiles (Q) of intake of meat, seafood and fruit/vegetable groups in the Supplementation with Antioxidant Vitamins and Minerals (SU.VI.MAX and SU.VI.MAX 2) studies, 1994–2007 (n 1823)
* Adjusted on energy (all such values).
† Due to low consumption, individuals reporting organ meat consumption were divided into non-consumers, consumers with low intake ( < median of residuals calculated within organ meat consumers) and consumers with high intake (>median).
Association between nutrient intake and hearing level
Table 3 shows the associations between intake of vitamins and HL. The adjusted mean differences of HL in comparison to the first quartile of nutrient intake are presented. Women with higher intake of retinol and vitamin B12 tended to have better HL compared with those having lower intake. No association was found for β-carotene, folate, and vitamins B6, C and E. Results of the logistic regression analysis in women indicated a significant association between vitamin B12 and hearing loss (P= 0·03), but no associations were observed with retinol (P= 0·11) or other nutrients (all P>0·05). Among men, no significant association emerged with either linear (regarding HL) or logistic (regarding hearing loss) regression (all P>0·05). Exclusion of subjects who developed diabetes or vascular disease (n 111) during follow-up did not substantially modify the results.
Table 3 Linear regression analysis of the sex-specific association between quartiles (Q) of nutrient intake at baseline and hearing level (HL) 13 years later in the Supplementation with Antioxidant Vitamins and Minerals (SU.VI.MAX and SU.VI.MAX 2) studies, 1994–2007 (n 1823) (Mean differences and 95% confidence intervals)
Ref, reference.
* Adjusted for age (years), BMI (kg/m2), educational level (primary, secondary or university level), physical activity (irregular, < 1 h walking/d, ≥ 1 h walking/d), supplementation group (yes or no), energy intake (excluding energy from alcohol) (kJ/d), alcohol use (g/d), smoking status (never smoked, former smoker or current smoker), systolic and diastolic blood pressure (mmHg).
† Adjusted mean difference of HL (decibels hearing level) (95 % CI) in comparison to Q1 (all such values).
Association between food intake and hearing level
Table 4 shows the associations between intake of total meat, seafood, fruit and vegetable groups and HL. The adjusted mean differences of HL in comparison to the first quartile of food intake are presented. Women with a higher consumption of meat as a whole, red meat and organ meat had a better HL compared with those having lower consumption of these food groups, as shown by both models. However, no association was found for poultry/game, processed meat or other food groups (seafood and fruits/vegetables). Among men, higher intake of seafood as a whole and of shellfish tended to be associated with better HL, while no association was observed for the other food groups. Results of the logistic regression analysis indicated a significant association between intake of meat (P= 0·03), poultry and game (P= 0·01) and organ meat (P= 0·01) and hearing loss in women, while no association was observed in men (all P>0·05). Exclusion of subjects who developed diabetes or vascular disease during follow-up (n 111) did not substantially modify the results.
Table 4 Linear regression analysis of the sex-specific association between quartiles (Q) of food intake at baseline and hearing level (HL) 13 years later in the Supplementation with Antioxidant Vitamins and Minerals (SU.VI.MAX and SU.VI.MAX 2) studies, 1994–2007 (n 1823) (Mean differences and 95% confidence intervals)
Ref, reference.
* Adjusted for age (years), BMI (kg/m2), educational level (primary, secondary or university level), physical activity (irregular, < 1 h walking/d, ≥ 1 h walking/d), supplementation group (yes or no), energy intake (excluding energy from alcohol) (kJ/d), alcohol use (g/d), smoking status (never smoked, former smoker or current smoker), systolic and diastolic blood pressure (mmHg).
† Adjusted mean difference of HL (decibels HL) (95 % CI) in comparison to Q1 (all such values).
‡ Model 2: model 1+food groups, as appropriate (for meat: groups of seafood and fruit/vegetables, and subgroups of meat as appropriate; for seafood: groups of meat and fruit/vegetables, and subgroups of seafood as appropriate; for fruit and vegetables: groups of meat and seafood, and subgroups of fruit/vegetable as appropriate).
Discussion
In the present large prospective study, a long-term, sex-specific association was observed between diet and HL. Specifically, higher intakes of retinol and vitamin B12 tended to be associated with better HL in women, while no association was found in men. In addition, food groups known to be significant sources of these micronutrients, i.e. meat as a whole, red meat and organ meat, were associated with better HL in women. Higher intake of seafood as a whole and shellfish tended to be associated with better HL in men.
Association with intake of nutrients
There are relatively limited data on nutrient intake and HL, although potential associations between HL and nutrient serum concentrations have received some attention. In the present sample, intake of vitamin B12 tended to be associated with better HL in women. The present results are in line with those of a study showing that reduced vitamin B12 intake and serum vitamin B12 concentration were associated with age-related auditory dysfunction in a sample of fifty-five females(Reference Houston, Johnson and Nozza10). However, in an intervention study involving ninety-three older adults, a short-term vitamin B12 supplementation was unrelated to improvement in hearing status in vitamin B12-deficient individuals(Reference Park, Johnson and Shea-Miller23). Observational studies conclude that there is no association between serum concentrations of vitamin B12 and age-related hearing loss either in cross-sectional(Reference Gopinath, Flood and Rochtchina12, Reference Berner, Odum and Parving24) or longitudinal(Reference Gopinath, Flood and Rochtchina12) settings, possibly due to insufficient power for analyses. Hearing loss in the elderly is believed to be mostly due to cochlear dysfunction(Reference Houston, Johnson and Nozza10, Reference Schuknecht and Gacek25), which is highly dependent on vascular supply(Reference Schuknecht and Gacek25). Homocysteine has been shown to be a risk factor for CVD(Reference Eikelboom, Lonn and Genest26); meanwhile, folic acid and vitamin B12 are known to be important determinants of homocysteine status(Reference Deshmukh, Joglekar and Lubree27, Reference Selhub, Jacques and Wilson28). Therefore, the potential association between poor vitamin B12 intake and HL could be partly mediated by unfavourable homocysteine concentrations. However, this interpretation is limited by the lack of an association between vitamin B12 intake and vitamin B12 blood concentration status, due to issues of malabsorption in the elderly(Reference Dhonukshe-Rutten, de Vries and de29), and by the fact that homocysteine has a stronger association with vitamin B12 status compared with intake(Reference Selhub, Jacques and Wilson28, Reference Dhonukshe-Rutten, de Vries and de29). Another pathway for the potential impact of vitamin B12 deficiency on audition might be through inhibition of neuron myelination in the cochlear nerve(Reference Houston, Johnson and Nozza10).
The present results indicated no association between HL and folate or vitamin B6 intake, which are also cofactors of homocysteine metabolism. In the literature, folate intake was significantly associated with hearing function in an observational study(Reference Houston, Johnson and Nozza10), while folic acid supplementation slowed the decline in hearing acuity(Reference Durga, Verhoef and Anteunis11). In addition, folate status has been either negatively associated with hearing impairment(Reference Gopinath, Flood and Rochtchina17, Reference Lasisi, Fehintola and Yusuf30) or has showed no relationship(Reference Berner, Odum and Parving24). Differences across results might be partly due to cross-sample variations in nutrient intake or blood status, and/or folate assessment methods(Reference Puwastien, Pinprapai and Judprasong31). The potential audio-protective effect of vitamin B6 has received little attention in the literature. We are aware of only one study that reported no association between vitamin B6 blood status and hearing(Reference Durga, Anteunis and Schouten32), consistent with the present findings.
The role of antioxidants in the management of hearing loss has generated considerable interest over the past few years. Although antioxidants were shown to be protective against several types of hearing impairment, including those related to age, most studies have been experimental(Reference Darrat, Ahmad and Seidman13). The present findings indicated no association between HL and vitamins C and, E and β-carotene. Consistent with the present findings, no association was found between intake of vitamin C and β-carotene and hearing loss prevalence or incidence(Reference Gopinath, Flood and McMahon15). In addition, vitamin E intake was negatively associated with the prevalence of age-related hearing loss, but showed no association with its incidence(Reference Gopinath, Flood and McMahon15). In another study, a negative association was found between sensorineural hearing loss and intake of vitamins C and E, while no association was observed for β-carotene(Reference Spankovich, Hood and Silver14). Research with cancer patients showed that subjects who achieved the highest plasma concentrations of vitamins C and, E and Se after supplementation had significantly less loss of high-tone hearing during chemotherapy(Reference Weijl, Elsendoorn and Lentjes33). Finally, serum levels of β-carotene were negatively associated with the prevalence of hearing impairment in older adults(Reference Michikawa, Nishiwaki and Kikuchi16).
The present results showed that retinol intake tended to be associated with a better HL in women. Similarly, in a recent study, vitamin A intake was inversely associated with the prevalence of hearing loss in a sample of men and women(Reference Gopinath, Flood and McMahon15). However, in the same study, vitamin A intake did not modify the risk of incident hearing loss, while it was suggested that retinol intake could even worsen hearing performance in older adults(Reference Spankovich, Hood and Silver14). Increased serum retinol was shown to prevent hearing impairment in older adults(Reference Michikawa, Nishiwaki and Kikuchi16). The role of retinol in preventing hearing impairment is not yet clearly understood. Mechanisms of action may involve its antioxidant properties, although there are conflicting data in the literature about retinol being antioxidant(Reference Bjelakovic, Nikolova and Gluud34). Other potential mechanisms of the retinol-audition link include the inhibition of a c-Jun N-terminal kinase signal pathway known to be involved in apoptosis(Reference Ahn, Kang and Kim35) or enhancement of hair cell renewal(Reference Lefebvre, Malgrange and Staecker36).
Association with intake of food
Potential associations between intake of various food groups and HL have thus far received little attention in the literature. To our knowledge, the present study is the first to show an association between meat intake, in particular, organ meat and red meat, and hearing function. Meat, especially organ meat, is the richest dietary source of both retinol and vitamin B12, and helps explain the observed association between these nutrients and HL. In addition, the present results indicated that seafood, in particular, shellfish, tended to be associated with HL in men. This supports data from the literature indicating that regular consumption of fish protected against hearing loss(Reference Gopinath, Flood and Rochtchina17). Fish rich in n-3 fatty acids could have a role in maintaining healthy auditory function(Reference Gopinath, Flood and Rochtchina17).
Difference by sex
In the present study, HL was associated with various nutrients and food groups, mostly in women. No association was observed in men, apart from the role of seafood (especially shellfish) intake, which tended to be associated with HL. Reasons behind these differences can be various. First, the prevalence of age-related hearing loss is clearly sex dependent. Cochlear and sensorineural functions of the ears are more affected in men in the course of ageing and might be less sensitive to nutrition factors. Second, as women had lower intakes of nutrients and food groups than men, they are therefore more likely to benefit from an increase in intake of beneficial dietary components. The present analyses indicated no interaction by sex, which supports the idea that mechanisms behind the association between diet and hearing might be similar in men and women. It was, however, important to present the models by sex, as previously emphasised(Reference Dobie37).
Strengths and limitations
Strengths of the present study include its large sample of community-dwelling subjects and its prospective design. There are indeed very few epidemiological data on the association between intake of food and nutrients and hearing function. Dietary data reflected midlife exposure because they were collected when the participants were aged 45–60 years, which was 13 years before the assessment of auditory function. The use of repeated 24 h dietary records resulted in relatively accurate dietary data, especially concerning nutrient intake in well-educated subjects(Reference Resnicow, Odom and Wang38). In the present study, hearing function was assessed using standardised audiometric methods. Finally, the exclusion of subjects with personal or family history of hearing impairment allowed us to more accurately identify subjects with age-related hearing loss. It is important to note that the level of sensitivity to nutrition and the type of food involved in hearing protection might differ according to the pathogenesis of hearing loss.
The main limitation of the present study was the absence of audiometric assessment at baseline. Further, generalisability of the findings is somewhat limited as we selected a subsample of the SU.VI.MAX2 cohort with no personal or family history of hearing loss. These results should therefore be confirmed in other populations. In addition, participants in a long-term cohort initially recruited for a randomised controlled trial are likely to be particularly health conscious and to have high functional capacity levels. The relatively young age of the population was another limitation potentially leading to an underestimation of the association between diet and HL. Finally, although a wide range of covariates was assessed during follow-up, we cannot exclude the possibility that other important confounders were omitted, such as ototoxic medication, habitual noise exposure at work, medical conditions other than those taken into account and genetic factors.
Conclusion
The present study documented a long-term association between diet in midlife and HL assessed 13 years after baseline. Intake of retinol and vitamin B12 tended to be associated with a better HL in women. In addition, intake of food groups known to be significant sources of these two micronutrients (i.e. meat as a whole, red meat and organ meat) was also associated with a better HL in women. The associations were less significant in men in whom only intake of seafood, particularly shellfish, tended to be associated with a better HL. Several direct or indirect mechanisms are likely to be involved in the association between diet and hearing function. Further research is required to confirm these results and to better elucidate the mechanisms behind the link between diet and hearing.
Acknowledgements
We thank all scientists, dietitians, technicians and assistants who helped carry out the SU.VI.MAX 2 study. We especially thank N. Arnault, who coordinated data management and performed the analyses, R. Mehroug, who coordinated operational aspects of the study and G. Monot, who coordinated the computing aspects. We are grateful to the volunteers and clinicians who carried out the clinical examinations. S. P. conducted the literature review and drafted the manuscript. E. K.-G., C. J., P. D., V. A. A., S. H. and P. G. were involved in interpreting results and editing and critically reviewed the manuscript. E. K.-G., P. G. and S. H. were responsible for developing the design and protocol of the study. The SU.VI.MAX2 study was supported by the Agence Nationale de la Recherche (ANR-05-PNRA-010), Direction Générale de la Santé (Ministry of Health), Médéric, Sodexo, Ipsen, Mutuelle Générale de l'Education Nationale and Pierre Fabre. Mederic and Mutuelle Générale de l'Education Nationale are French health insurance organisations complementary to the national health insurance system. Ipsen and Pierre Fabre are private pharmaceutical companies. Sodexo is a food-catering company that sponsored events between the researchers and study participants. S. P., C. J., P. D., V. A. A., P. G. and E. K.-G. report no conflicts of interest. S. H. participated in the examiner board of the Union Pour la Promotion des Industries Conserve Appertisée.
APPENDIX 1 Members (City) of the Research Group SU.VI.MAX 2 Study
Luc Vogt (Agen); Michèle Escande (Allauch); Jean-Marie Sérot (Amiens); Emmanuel Vasseur (Angers); Matthieu Debray (Annecy); Chantal Hussonnois (Auxerre); Martine Iehl-Robert (Besançon); Myriam le Sommer (Bordeaux); Thierry Boge (Bourg en Bresse); Josiane Rajaonarivo (Bourges); Jean Jouseau (Bourges); Monique Frison (Brest); Armelle Gentric Brest); Fabienne Leenaert (Caen); Bernard Bascou (Carcassonne); Jean-Paul Lemaire (Cavaillon); Nathalie Baptiste (Champcueil); Marie-France Maugourd (Champcueil); Anne Gibelain (Chartres); Roland Lopitaux (Clermont-Ferrand); Jacques Hild (Colmar); Henri Nachar (Avignon); Géraldine Soulié (Dax); Francine Clémenti (Dax); Patrick Friocourt (Blois); Alain Sagnier (Chambéry); Philippe Schiano (Hyères); Andréa Collet (Port-Louis Riantec); Dominique Richard (Dijon); Françoise Zandi (Gradignan); Pascal Couturier (Grenoble); Agathe Raynaud-Simon (Ivry); Pierre Lermite (La Roche sur Yon); Michel Alix (La Rochelle); Emmanuel Alix (Le Mans); François Puisieux (Lille); Cédric Gaxatte (Lille); Thierry Dantoine (Limoges); Pierre livet (Lyon); Gilles Albrand (Lyon); Pierre Haond (Lyon); Pascal Ménecier (Macon); François Pinoche (Malestroit); Sylvie Bonin Guillaume (Marseille); Marc Heim (Marseille); André Wang (Metz); Claude Jeandel (Montpellier); Yves Passadori (Mulhouse); Athanase Benetos (Nancy); Catherine Couturier (Nantes); Gilles Berrut (Nantes); Sylvie Sacher-Huvelin (Nantes); Henri Patouraux (Nevers); Patrice Brocker (Nice); Olivier Guérin (Nice); Denise Strubel (Nîmes); Florence Dupriez (Orléans); Jean-Bernard Gauvain (Orléans); Yves Wolmark (Paris); Olivier Hanon (Paris); Bernard Cassou (Paris); Philippe Déjardin (Paris, Troyes); François de la Fournière (Pau); Frédéric Woné (Périgueux); Claudine Buj-Hardy (Perpignan); Marc Paccalin (Poitiers); Bertrand Placines (Pontacq-Nay); Jean-Luc Novella (Reims); Pierre Jouanny (Rennes); Florence Martin (Rodez, Toulouse); Philippe Chassagne (Rouen); Isabelle Landrin (Rouen); Chantal Girtanner (St-Etienne); Régis Gonthier (St-Etienne); Claudie Troadec (St-Brieuc); Christophe Dourthe (Saintes); Eric Bonnin (Saintes); Jean-Paul Marot (St-Nazaire); Sylvie Rossignol (Sens); Georges Sebbane (Sevran); Pierre-Yves Cornu (Sevran); Jean-Jacques Monsuez (Sevran); Isabelle Périllat (Sevran); Robert Ratiney (Sevran); Georges Kaltenbach (Strasbourg); Yves Rolland (Toulouse); Thierry Constans (Tours); Gisèle Coz (Tréguier); Monique Ferry (Valence); Jean Pierre Aquino (Versailles); Odile Cézard (Voiron).
