Serum Levels of Retinol and Other Antioxidants for Hearing Impairment Among Japanese Older Adults

Abstract

AGE-RELATED hearing loss is a common condition among older people (1). The prevalence of hearing impairment as defined by the pure-tone audiometric examination was 45.9% (aged 48–92 years) in the United States (2) and 39.1% (aged 55–99 years) in Australia (3). In Japan, one population-based study reported that half of the participants aged 60 years or older had hearing difficulties (4). Because hearing impairment can reduce quality of life in the older adults (5–7), primary, secondary, and tertiary prevention is important from the viewpoint of public health.

Although the etiology of age-related hearing loss still remains unclear (8), there are past reports of oxidative stress being associated with auditory aging (9–11), suggesting that antioxidants play a preventive role. The association between vitamin A (retinol) and the prevention of hearing impairment in particular has been discussed because the concentration of vitamin A in the inner ear has been shown to be high (12) and retinoic acid, an active metabolite of retinol, plays an important role in the normal development of the organ of Corti (13). An earlier experiment using rats showed that retinoic acid stimulates the regeneration of hair cells (14).

As far as we know, no community-based epidemiological studies on the relationship of retinol and other antioxidants with hearing impairment have been carried out. In this study, we examined hearing impairment using objective methods, measured serum levels of retinol and other antioxidants, and cross-sectionally evaluated the association of retinol and other antioxidants with hearing impairment in Japanese community-dwelling older adults.

METHODS

Study Population

This study is a community-based study in Kurabuchi Town of Takasaki City, Gunma Prefecture, which is located in central Japan. Between April 2006 and July 2007, a home visit health survey targeting all registered residents of this town aged 65 years or older was carried out by trained public health nurses and local welfare commissioners using a structured questionnaire. We identified 1,325 noninstitutionalized and functionally independent residents as eligible population. All eligible residents were invited to a detailed health assessment including audiometric examinations and blood collection at eight town community centers, and 762 residents (330 men and 432 women, 58% of those eligible) participated. Four participants declined to give blood samples and were excluded from later analysis. The remaining 758 residents made up the study population (Table 1).

The questionnaire data from the home visit health survey was used to compare the characteristics of the participants and nonparticipants in the detailed health assessment. Proportion of those aged 80 years or older was 27.3% among the participants and 34.0% among the nonparticipants (sex-adjusted p = .07). Women accounted for 56.6% of the participants and 53.4% of the nonparticipants (age-adjusted p = .17). Between the participants and nonparticipants, prevalences of subjective hearing difficulties (11.1% vs 13.2%, age- and sex-adjusted p = .74) and hearing aid usage (6.5% vs 5.3%, age- and sex-adjusted p = .14) were not different.

All participants gave written informed consent, and the Ethics Committee of the School of Medicine, Keio University (Tokyo, Japan), approved this study.

Measurement of Retinol and Other Antioxidants

Serum samples of nonfasting venous blood were collected, immediately frozen, and stored at −80°C until tested. Serum levels of retinol and other antioxidants (alpha- and gamma-tocopherol, and carotenoids including beta-cryptoxanthin, alpha- and beta-carotenes, lycopene, and lutein plus zeaxanthin) were determined by high-performance liquid chromatography (15,16). For the assays, the ranges of within-day and between-day variations were 0.3%–4.7% and 1.4%–8.1%, respectively. Because provitamin A carotenoids are important sources of retinol (17), we also included the combination of beta-cryptoxanthin and alpha- and beta-carotenes in the provitamin A family. Although differences in research methods make it difficult to directly compare our participants’ data with retinol and other antioxidant levels observed elsewhere, overall, our participants’ values seemed to not greatly differ from that of other Japanese populations (18,19).

Audiometric Examination

Pure-tone air-conduction audiometry was carried out by trained technicians in a separate quiet room with an audiometer (AA-56; RION Inc, Tokyo, Japan). To reduce the influence of ambient noise (24–29 dB SPL [A]), we used circumaural earphones. The audiometer was calibrated regularly during the study period to Japanese standards. Due to the field setting nature of the study, we tested a signal with an intensity of 30 dB hearing level (HL) at a frequency of 1 kHz and a signal with an intensity of 40 dB HL at a frequency of 4 kHz, in accordance with Japan’s Industrial Safety and Health Law’s stipulations on health examinations for workers. Hearing impairment was defined as a failure to hear a 30-dBHL signal at 1 kHz and a 40-dBHL signal at 4 kHz in the better ear.

Covariates

Body mass index was calculated from height and body weight measurements. Total cholesterol and hemoglobin A_1c were measured in blood samples at a national clinical laboratory under an internal/external quality control program. Blood pressure was taken by trained observers using an automatic sphygmomanometer (BP-103i; Colin, Tokyo, Japan) after the participants were given a 15-minute rest. Hypertension was defined as a systolic blood pressure of more than or equal to 140 mmHg, a diastolic blood pressure of more than or equal to 90 mmHg, usage of antihypertensive agents, or any combination of these. Depressive mode was determined by Geriatric Depression Scale 5: Participants scoring more than or equal to 2 were classified as being depressed (20). Information on smoking status, alcohol consumption, educational level, occupational noise exposure, and past–current medical history of life-threatening diseases (including cancer, stroke, and coronary artery disease) was also collected.

Statistical Analysis

The prevalence of hearing impairment according to age (evaluated with the Cochran–Armitage trend test) and sex was calculated.

The relationship of serum levels of retinol and other antioxidants with hearing impairment was assessed by dividing the participants into quartiles of retinol and other antioxidants. To divide the provitamin A family into categories, we assigned each participant a category score (1, 2, 3, or 4) for each antioxidant (beta-cryptoxanthin, alpha- and beta-carotenes) based on the quartiles and counted the sum of the category scores (range from 3 to 12). Statistical software was used to divide the participants automatically into four categories according to these scores. Odds ratios (OR) and 95% confidence intervals (CI) for hearing impairment were calculated for each of the upper three quartiles of retinol and other antioxidant levels relative to the lowest quartile. Trends across increasing quartiles of serum levels were also calculated by treating the quartiles as an integral value. For the adjusted analysis, we constructed two different logistic regression models. First, we adjusted for age (continuous), sex, smoking status (never, former, current), alcohol consumption (never, former, current), hypertension (no, yes), and total cholesterol level (continuous; model 1). Other measured covariates were not associated with exposure variables in this population and so were not included in the model. Second, we included retinol and other antioxidants (quartiles) as an addition to model 1 to determine their influence on the association between each antioxidant and hearing impairment (model 2). Because the correlation between alpha-carotene and beta-carotene was high (correlation coefficient = .71), we used the provitamin A family in model 2 rather than individual antioxidants. In each model, the ORs of hearing impairment were also calculated using continuous variables of retinol and other antioxidants.

All analyses were conducted using STATA version 9 (STATA Corp, College Station, TX).

RESULTS

The overall prevalence of hearing impairment was 19.7%, with no significant difference according to sex (21.9% in men and 17.9% in women, p = .18). The prevalence increased with age (4.3% in 65−69 years, 14.2% in 70−79 years, and 42.0% in older than 80 years; p for trend < .01).

The prevalence of hearing impairment by quartile of retinol and provitamin A carotenoids is presented in Table 2. Crude analysis showed that serum levels of retinol and provitamin A carotenoids were inversely related to the prevalence of hearing impairment. In model 2 including all antioxidants as well as the other covariates, the ORs (95% CI) for the highest quartile of retinol and the provitamin A family compared with the lowest were 0.51 (0.26−1.00) and 0.53 (0.27−1.02), respectively. Moreover, increasing retinol level was related with reduced odds of hearing impairment in a dose–response manner (p = .03). A dose–response relationship was marginally seen for the provitamin A family (p = .09). When retinol was included in the model as a continuous variable, the adjusted OR (95% CI) of hearing impairment for a 1 μmol/L increase in retinol was 0.68 (0.48−0.96).

In contrast, there was no association of non-provitamin A carotenoids and tocopherols with hearing impairment (Table 3). Adjusting for the covariates did not alter the results substantially.

DISCUSSION

To the authors’ knowledge, this is the first community-based epidemiological study to indicate that serum levels of retinol and provitamin A carotenoids are inversely related to the prevalence of hearing impairment in an older population.

The first strength of our study is the objective measurement of retinol and other antioxidant levels in serum samples. Because only limited information is available on the levels of these compounds in foods and the Japanese diet is particularly rich in variety, objective measurements provide far more precise and accurate data than can be collected through such subjective means as questionnaires and recall. Second, the serum concentrations of retinol and other antioxidants in our participants mostly reflected their intake from daily meals (21) because 5% or less had used supplements (unpublished data from the previous questionnaire survey by the local government). Therefore, our results suggest that appropriate intervention in dietary intake of retinol and provitamin A carotenoids may be an important primary prevention factor for hearing impairment. Rather than high-dose supplementation of retinol and provitamin A carotenoids (particularly beta-carotene), which is potentially harmful (22,23), dietary control would be the more desirable intervention method.

Retinol is required for good vision, growth, reproduction, cell proliferation, cell differentiation, and the integrity of the immune system (22). In fact, decreased serum retinol (<1.05 μmol/L) can lead to compromised biologic function (24). It is also known that retinoic acid, an active metabolite of retinol, is indispensable for normal development of the Corti organ (13). Although the role of retinol in preventing hearing impairment is not clearly understood, an experimental study indicated that retinoic acid prevents noise-induced hearing loss through the inhibition of a c-Jun N-terminal kinase signal pathway known to be involved in apoptosis (25). Also, there are in vitro reports of retinoic acid and other factor mixtures (transforming growth factor-alpha and epithelial growth factor) enhancing hair cell renewal after damage caused by ototoxic drugs (14,26,27).

Since oxidative stress has been suggested to be involved in hearing impairment, antioxidants are expected to play a preventive role (9–11). Keithley and colleagues (28) noted that Cu/Zn superoxide dismutase (SOD1) protects cells from reactive oxygen species and observed that the degeneration of spiral ganglion cells and hair cells progresses earlier in sod1 null mice than in wild-type mice. Although beta-cryptoxanthin, and alpha- and beta- carotenes can be potentially converted into retinol (provitamin A activity), these provitamin A carotenoids themselves have antioxidant activity (29), which may partially explain our findings. However, further experimental and epidemiological evidence is necessary to investigate this possibility; in our study, other antioxidants such as non-provitamin A carotenoids (in particular, lycopene was the most efficient single-oxygen quenchers of the natural carotenoids (29)) and tocopherols were not associated with hearing impairment.

Despite the strengths of this study, there are several methodological issues that warrant discussion. First, the participation proportion of the eligible participants in the detailed health assessment was 58%. However, the home visit health survey revealed no difference in subjective perceptions of hearing difficulties and hearing aid usage between the participants and nonparticipants in the detailed health assessment. Second, due to the field setting nature of the study, only minimal audiometric tests were performed, which possibly led to an occasional misclassification of hearing impairments. But such misclassifications would have been nondifferential and therefore more likely to underestimate the true associations. Finally, the generalizability of the finding should be confirmed in other populations.

It should also be noted that because this is a cross-sectional study, cause-and-effect relationship cannot be inferred, and the possibilities of reverse causation or other mechanisms cannot be ruled out. For instance, hearing-impaired older adults eating lower quality food, or low serum retinol and provitamin A carotenoids may simply represent a chronic proinflammatory state. However, if this were the case, serum levels of non-provitamin A carotenoids should also be lower in hearing-impaired participants (30). Because we did not observe such association, we think it is unlikely that our findings are explained by reverse causation or by chronic inflammation.

In conclusion, increased serum levels of retinol and provitamin A carotenoids were clearly associated with a decreased prevalence of hearing impairment. This study suggests that modification of dietary habits may contribute to the prevention of hearing impairment.

Funding

This study was supported by a grant-in-aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 19500692).

We thank the staff members at Kurabuchi Branch Office and Harunaso Hospital, and participants in this study. Additional thanks go to Michiko Okamoto for her valuable help.

References

Author notes