Introduction

Age-related macular degeneration (AMD) is one of the principal causes of blindness, comprising an estimated 5% of all cases of blindness.1 Worldwide, the estimated prevalence of early and late AMD varies from 4.4 to 16.8% and 0.6 to 3.7%, respectively, depending on the study population.2, 3, 4, 5, 6 Numerous studies have identified age,7, 8 female gender,9 smoking,7, 8 blood cadmium level,10 diabetes mellitus,7, 11 alcohol consumption,12 liver cancer7 and family history13 as risk factors for AMD.

Observational studies14, 15, 16 have also investigated the relationship between AMD and dietary factors, and suggested that fruits, vegetables, vitamin C, vitamin E, β-carotene and other carotenoids and flavonoids may have an important role in the onset and progression of AMD. Recently, studies have shown associations between AMD and dietary patterns,17, 18, 19 the Healthy Eating Index Score20 and the Mediterranean diet.21 Healthier diets including high fruit and vegetable intake were associated with lower AMD risk.

Among carotenoids, lutein and zeaxanthin, the main pigment in the macular area in human retina are particularly effective for protecting the macular from oxidative damage by eliminating harmful reactive oxygen species.22, 23 β-Carotene accumulates in retinal pigment epithelium (RPE) cells and provides some protection against light-induced retinal damage.24 Flavonoids assist with blocking the accumulation of reactive oxygen species and upregulating the expression of proteins, which could protect against oxidative stress in human RPE cells.25 However, interventional studies assessing the effect of dietary supplementation on AMD progression are inconsistent.26, 27, 28

Most previous studies on the association between AMD and dietary factors have been performed in the American14, 15, 26 and European regions.2, 16, 29 Some studies have also been conducted in Japan, China and India.11, 30, 31 The prevalence of AMD in Korea is estimated to be 16.9% for early AMD and 1.8% for late AMD in the elderly.8 With an increasing elderly population, age-related health problems are becoming more important in Korea. In addition, female gender has been associated with a higher risk of AMD in many studies.5, 9 Identification of risk factors for female AMD need to understand the pathogenesis of the AMD and reduce the risk. However, there has been little research associated with AMD and the relationship between dietary intake and AMD has never been studied in the Korean population.

Thus, this study was conducted to investigate the association between AMD and dietary factors, including the consumption of fruits, vegetables, carotenoids and flavonoids in the Korean elderly women, using the Korea National Health and Nutrition Examination Survey (KNHANES) data.

Subjects and methods

Study population

This study used data from the fifth KNHANES, a cross-sectional and nationally representative survey conducted by the Korean Ministry of Health and Welfare from 2010–2012. KNHANES comprises a series of cross-sectional, stratified and multistage probability surveys of the civilian, non-institutionalized Korean population. The survey consists of a health interview survey, a health examination survey and a nutrition survey. The response rate of these surveys was 81.9, 80.4 and 80.0% in 2010–2012, respectively. Detailed information of each survey is available on the website (http://knhanes.cdc.go.kr).32

This study included all women aged ⩾65 years (n=2711) in the database. We excluded 159 participants without 24 h dietary recall data; 559 without fundus photograph; 366 diagnosed with diabetes mellitus or diabetic retinopathy, or reported taking medications for diabetes; 33 reported a history of liver, stomach, colorectal, uterine or cervical cancer; 573 had undergone ocular surgery; and 13 with implausible energy intakes (<500 or >5000 kcal). Thus, a total of 1008 participants were eligible for further analysis.

This study was approved by the Institutional Review Board of the Korea Centers for Disease Control and Prevention (2010-02CON-21-C, 2011-02CON-06-C, 2012-01EXP-01-2C). All subjects in the study participated voluntarily and provided informed consent.

General characteristics and other covariates

The general questionnaire and health interview data were used to obtain the demographic and socioeconomic characteristics of the participants, including age, body mass index (BMI), reproductive (number of pregnancies and duration of oral contraceptive pills) and menopausal (menopausal age, postmenopausal period and duration of hormone replacement therapy) factors, residential area, education level, family income, smoking status, alcohol consumption and dietary supplement use.

BMI was calculated as weight (kg)/(height (m))2. Residential area was categorized as urban or rural. Education level was categorized as less than high school or high school and above. Family income was assessed with 4 categories (low, middle-low, middle-high or high). Smoking status was assessed with three categories (current smoker, ex-smoker or non-smoker). Alcohol consumption was assessed with five categories (never, <1 drink/mo, 1 drink/mo, 2–4 drink/mo or ⩾4 drink/mo within the last year). Dietary supplement use was assessed with two categories (yes or no).

Assessment of AMD

The presence of early and late onset AMD was determined on the basis of the fundus photograph.5, 33 All fundus photographs were graded twice using the International Age-related Maculopathy Epidemiological Study Group grading system.34 Preliminary grading was performed on the site by trained ophthalmologist. Detailed grading was later done by nine retina specialists having experience in grading early and late AMD. Any inconsistencies between the preliminary and detailed grading was decided by an independent reading specialist. When the fundus photograph for a participant’s eyes were different in severity, the grade was defined based on the more advanced grade. When the fundus photograph for only one eye was assessed, the grade was evaluated by that eye. Participants were diagnosed with early AMD if they met one of the following criteria: the presence of soft, indistinct drusen or reticular drusen; the presence of soft or hard distinct drusen with pigmentary abnormalities in the absence of signs of late AMD. Late AMD was defined as the presence of neovascular AMD or geographic atrophy. Neovascular AMD was identified by the detachment of RPE or neurosensory retina, or the presence of hemorrhages in sub-RPE or subretinal spaces. Geographic atrophy was defined as the presence of a circular discrete depigmented area ⩾175 μm in choroidal vessel diameter. The presence of any AMD was defined as either early (n=180) or late (n=10) AMD in this study.

Dietary assessment and estimation of dietary intake of carotenoids and flavonoids

We used the dietary intake data obtained from a single 24 h recall in the KNHANES data. Participants reported all foods and drinks consumed for the previous day during face-to-face interviews.

To estimate the intake of carotenoid and flavonoid, we merged individual food files from KNHANES with carotenoid and flavonoid databases. The carotenoid database included 2191 food items and was constructed using the food items of KNHANES35 and the Carotenoid Content Database from the United States Department of Agriculture.36 The flavonoid database was developed based on a systematic review and comprised 649 food items based on their analytic values and 900 items replaced with adaptations or calculations from similar items,37 where we used data for the closest food item based on species and nutrients. The carotenoid and flavonoid databases included 68.7 and 68.9% of all plant foods reported in the 24 h dietary recall method, respectively.

Statistical analysis

Differences in the distribution of general characteristics between the group without AMD and the group with AMD were investigated using the SURVEY FREQ procedure. We calculated the crude weighted mean and s.e. of continuous variables using the SURVEY MEAN procedure. Significant difference according to the presence of AMD was analyzed using the SURVEY REG procedure. SURVEY LOGISTIC analysis was performed to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for AMD across quartiles of fruit, vegetable, vitamin A, β-carotene and flavonols intake, where the lowest quartile was set as the reference. The objectives of the fourth Health Plan for Korea include increasing the percentage of Koreans who eat ⩾500 g of fruits and vegetables daily.38 Therefore, we estimated the ORs and CIs by comparing participants who consumed <500 g of fruits and vegetables with those that consumed 500 g or more. For dietary intake, we conducted analyses adjusted for age and then further controlled for potential confounding factors: BMI, postmenopausal period, duration of hormone replacement therapy,39 residential area, education level, family income, smoking status, alcohol consumption, dietary supplement use and total energy intake. All statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC, USA).

Results

Of the 1008 eligible subjects, the prevalence of AMD was 18.8% (data not shown). The mean age was not statistically significantly different between subjects without and with AMD (without 71.6 years, with 72.0 years). The proportion of subjects with AMD was higher in those with education level of high school and above. Except for the educational level, the general characteristics were not significantly different between subjects without and with AMD (Table 1). General demographic and socioeconomic factors were not significantly different between included and excluded subjects (data not shown).

Table 1 General characteristics of participants with or without AMD
Full size table

The subjects with AMD consumed less vegetable (274.1 g/day in subjects without AMD vs 237.0 g/day in subjects with AMD, P-value=0.006), and fruit and vegetable (424.3 g/day in subjects without AMD vs 367.0 g/day in subjects with AMD, P-value=0.009) than those without AMD after adjusting for age, BMI, postmenopausal period, duration of hormone replacement therapy, residential area, education level, family income, smoking status, alcohol consumption, dietary supplement use and total energy intake (Table 2). Sugar and oil/fat consumption of subjects without AMD are higher than those with AMD in our subjects. However, the mean percentages of daily calories from sugar (5.1 g) and oil/fat (3.8 g) were 1.3 and 1.9%, respectively. Therefore, it is difficult to estimate AMD prevalence according to sugar and oil/fat intake.

Table 2 Daily food intake of participants with or without AMD
Full size table

We found that subjects with AMD consumed fewer antioxidant nutrients including vitamin A (P-value=0.019), β-carotene (P-value=0.006) and flavonols (P-value=0.031) than those without AMD after adjusting for confounding factors. Energy intake and other nutrients were not significantly different between the groups after adjusting for confounding factors (Table 3).

Table 3 Daily nutrients, carotenoids and flavonoids intake of participants with or without AMD
Full size table

After adjustment for confounding factors, women with higher fruit and vegetable intakes had ORs<1.0, but there was no statistically significant decreasing risk across intake quartiles. OR for the highest vegetable intake quartile category compared with the lowest quartile category was 0.44 (95% CI 0.25, 0.77, P for trend=0.002). Subjects in the highest quartiles showed significantly decreased prevalence of AMD by 62% for vitamin A (OR (95% CI) 0.38 (0.21–0.68), P for trend=0.001), 64% for β-carotene (OR (95% CI) 0.36 (0.19–0.67), P for trend<0.001) and 55% for flavonols (OR (95% CI) 0.45 (0.25–0.82), P for trend=0.008) after adjusting for confounding factors (Table 4).

Table 4 ORs and 95% CIs of intakes of fruit and vegetable and the prevalence of AMD
Full size table

Although fish consumption has been associated with AMD,40 there was no a significant association between fish consumption and AMD in this study. The associations between fruit, vegetable, and nutrient intake and AMD were not significantly different after adjustment for fish consumption (data not shown).

Multiple logistic regression analysis showed a significant negative association between fruit and vegetable intake and the prevalence of AMD (OR (95% CI) for subjects consuming ⩾500 g=0.61 (0.38–0.97) compared with those consuming<500 g) after adjusting for age, BMI, postmenopausal period, duration of hormone replacement therapy, residential area, education level, family income, smoking status, alcohol consumption, dietary supplement use and total energy intake (Table 5).

Table 5 ORs and 95% CIs of AMD according to fruit and vegetable intake
Full size table

Discussion

This study investigated the association between dietary intake and the prevalence of AMD. We found a significant negative association between fruit and vegetable intake and AMD. The highest quartiles of vitamin A, β-carotene and flavonols intake had significantly decreased odds ratios for AMD compared with the lowest quartiles. In contrast to previous studies, we did not find a significant relationship between lutein/zeaxanthin and AMD.

This is the first study identifying an inverse relationship between fruit and vegetable consumption and AMD in elderly Korean women. High fruit and vegetable intake was inversely associated with the prevalence of AMD, which is consistent with other (non-Korean) studies. A prospective cohort study of postmenopausal women in the United States showed that the OR of AMD for participants who consumed 14 servings per week of vegetables was 0.49 (95% CI 0.27–0.91) compared with those who consumed 6.2 servings per week.41 In the Nurses’ Health Study and Health Professionals Follow-up Study, the pooled multivariate relative risk of neovascular age-related maculopathy for participants who consumed three or more servings of fruit per day was 0.64 (P for trend=0.004), compared with those who consumed <1.5 servings per day.15 Recently, healthy dietary patterns including high intake of fruits and vegetables have been shown to be associated with lower risk of AMD. In the Age-Related Eye Disease Study, subjects in the fifth quintile group of the oriental pattern score consuming high fruits, tomatoes, vegetables and whole grains had lower risk of early AMD by 26% and advanced AMD by 62% compared with those in the first quintile group.19

Beneficial effects from fruit and vegetable intake may derive from the nutrients and phytochemicals, such as fibers, vitamins, minerals, flavonoids and carotenoids.42 In this study, fruit and vegetable intake were significantly positively correlated with β-carotene, lutein+zeaxanthin and β-cryptoxanthin (data not shown). There was a significant inverse relationship between β-carotene intake and AMD, which is consistent with earlier studies in Western and Asian populations. A case–control study conducted in Japan showed that subjects in the highest quintile of β-carotene exhibited a low prevalence of neovascular AMD (OR (95% CI) 0.2 (0.1–0.4)).31 A hospital-based study in India showed that higher levels of dietary β-carotene consumption were associated with reduced AMD risk (OR (95% CI) 0.65 (0.42–0.86)).11 In prospective cohort studies from Nurses’ Health Study and Health Professionals Follow-up Study, subjects in the highest intake quintile for β-carotene had 32% less risk of advanced AMD including neovascular AMD and central geographic atrophy (pooled relative risk (95% CI)=0.68 (0.55–0.85)).43

Carotenoids, which constitute the main pigments of photoreceptors in the eye, help protect against damage and toxins.44 They act as biological sunglasses,45, 46, 47 absorbing and filtering blue light, which appears to be dangerous for RPE cells.48, 49 They also act as antioxidants and oxidative damage in RPE cells is known to contribute to macular degeneration.50 The RPE accumulates α-tocopherol and carotenoids (lutein, zeaxanthin and β-carotene), which serve as antioxidants.45, 46, 51 Chichili et al.51 showed the uptake and protective effects of β-carotene from tomatoes on oxidative stress in human RPE cell lines.

We did not find a relationship between xanthophyll carotenoids, such as lutein, zeaxanthin and β-cryptoxanthin, and AMD. Similarly, epidemiological studies on the association between lutein and zeaxanthin intake and AMD are inconclusive and differ according to the type of AMD. A cross-sectional population-based study showed no relationship between lutein/zeaxanthin and either early or late AMD.14 A recent prospective cohort study in the United States showed a relationship between lutein/zeaxanthin and advanced AMD (primarily neovascular AMD), but no relationship with intermediate AMD.43 Another cohort study conducted in the United States suggested that diets rich in lutein and zeaxanthin may have protective effects against intermediate AMD in women aged 50–74 years.41 Although lutein and zeaxanthin may be effective in preventing AMD, the effect may differ according to the development and progression of AMD. In this study, subjects with late AMD were 10% of the total AMD. Therefore, we could not confirm associations between lutein/zeaxanthin intake and each different type of AMD.

We observed for the first time that the highest quartile of flavonols intake had significantly decreased ORs for AMD compared with the lowest quartile. Subjects with AMD consumed significantly less quercetin than those without AMD after adjusting for confounding factors (P=0.02) (data not shown). The flavonol quercetin occurs abundantly in apples, onions and berries.52 Several previous studies have identified the protective effects of quercetin on human RPE. Kook et al.53 showed that quercetin protects human RPE cells in vitro from oxidative stress-induced cell death and cellular senescence. Saviranta et al.54 showed that quercetin may have protective effects on cells by inhibiting pro-inflammatory products, such as interleukin-6.

In this study, we determined that subjects with AMD consumed fewer anthocyanidins than those without AMD. However, no significant association between anthocyanidins and AMD was found. Association between dietary anthocyanin and AMD has not been reported in a cross-sectional study design, although a recent in vitro cell culture study demonstrated possible mechanisms related to the protective effects of anthocyanin in the eye.55 The study showed that blueberry anthocyanin extract suppressed senescence and ameliorated cell viability against light-induced cellular stress in RPE cells. Blueberry anthocyanin extract is beneficial to RPE cells by suppressing aging and apoptosis, and downregulating overexpressed vascular endothelial growth factor.

The prevalence of AMD among our participants was 18.8% (17.9% for early AMD and 1.0% for late AMD). The prevalence of AMD differ considerably according to age group. For the elderly women aged ⩾65 years, the prevalence of AMD was comparable to the rate from the Shihpai Eye Study conducted in Taiwan (9.5% for early AMD and 1.0% for late AMD). The prevalence for late AMD was 1.0%, which is similar several population based-studies including The Shihpai Eye Study56 and The Coimbra Eye Study.57

We identified an association between dietary intake and AMD in elderly women. The role of female sex hormones in AMD have been reported58 and menopause, accompanied by a decline in circulating estrogen levels, was associated with increased production of reactive oxygen species and lipid peroxidation products.59 Estrogen also helps induce the expression of antioxidant enzymes, such as manganese superoxide dismutase and glutathione peroxidase.60 Therefore, the anti-inflammatory properties of estrogen have a role in AMD pathogenesis.58

The present study has several limitations. First, as KNHANES is a cross-sectional study, we could not confirm the causal relationship between dietary factors and AMD. Second, we used a single 24 h dietary recall, which may not be sufficient to estimate the usual dietary intake. However, variations between a single day 24 h dietary recall and data obtained over 2–10 days were not much different (0.8% for energy, 8.1% for vitamin A and 3.4% for vitamin C) in the 2009 Korea National Health and Nutrition Examination Survey.61 Furthermore, the within-person variation for vitamin A is much lower rate than the difference (33%) in vitamin A intake between with and without AMD. We examined dietary intake of carotenoids, flavonoids and other nutrients, but not dietary supplement intake. There was limited data on antioxidant nutrient intakes from dietary supplements (19.6%) and these data existed only for 2010 and 2011 years but not 2012. Therefore, we were unable to estimate nutrient intakes from foods and supplements. Third, only 10 subjects had late AMD in this study, 9 with neovascular AMD and 1 with geographic atrophy. Therefore, we have insufficient data to identify if dietary intakes differed according to AMD type. Lastly, the carotenoids and flavonoids databases used in this study did not include all foods consumed.

Nevertheless, this study had several strengths. The trial subjects comprise a nationally representative sample of elderly Korean women and this is the first study showing an inverse relationship between dietary factors and AMD for a Korean population.

In conclusion, we found significant associations between fruit, vegetable, vitamin A, β-carotene and flavonols intake and AMD in elderly Korean women. The current results suggest that higher consumption of fruits, vegetables and foods containing multifunctional phytochemicals may help protect against AMD. This study strengthens existing evidence indicating that high fruit and vegetable consumption is beneficial for health.