Inverse Relationship Between High Blood 25-Hydroxyvitamin D and Late Stage of Age-Related Macular Degeneration in a Representative Korean Population

Abstract

Purpose.: To investigate the association of 25-hydroxyvitamin D with AMD.

Methods.: A population-based, cross-sectional study using a nationwide, systemic-stratified, multistage-clustered sampling method involved a total of 17,045 subjects older than 40 years who participated in the Korean National Health and Nutrition Examination Survey 2008 to 2012. All participants underwent standardized interviews, evaluation of blood 25-hydroxyvitamin D levels, and comprehensive ophthalmic examinations. A 45° digital fundus photograph of both eyes was taken under physiologic mydriasis. All fundus photographs were graded using the international classification and grading system.

Results.: Blood 25-hydroxyvitamin D levels were 17.5 ng/mL in women and 20.0 ng/mL in men. After adjusting for potential confounders including age, sex, smoking status, hypertension, heart problems, stroke, and sunlight-exposure time, the odds ratio (OR) for late AMD significantly decreased in the highest blood 25-hydroxyvitamin D quintile (OR, 0.32; 95% confidence interval [CI], 0.12–0.81; P for trend = 0.018) compared with the lowest quintile in men, but not in women. Early AMD was not associated with blood 25-hydroxyvitamin D levels in either sex.

Conclusions.: High level of blood 25-hydroxyvitamin D was inversely associated with late AMD in men but not women. Considering antiangiogenic and antifibrotic action of vitamin D, association between two variables warrants further studies.

Introduction

Age-related macular degeneration (AMD) is a leading cause of blindness in the elderly older than 60 years of age. ^1–3 Although the exact etiology is unknown, it is known to be multifactorial, with an interaction of genetic and environmental factors. ⁴ Aging, smoking history, and family history of AMD are consistently associated with AMD. ^5,6 Other potential risk factors are inconsistently associated with AMD across populations, including cardiovascular disease, sunlight exposure, dietary antioxidant intake, and dietary fat intake. ^7–9 We previously reported that age, male sex, and hypertension were risk factor for AMD in representative Korean population. ¹⁰ Recently, inflammation has been found to play an underlying role in AMD pathogenesis. A polymorphism in the complement factor H gene, which is responsible for downregulating inflammation, was strongly associated with AMD. ⁴ The inflammatory nature of AMD pathogenesis is further supported by recent research demonstrating that immunologic components, including immunoglobulins, complement factors, and fibrinogen, were found within drusen. ^11,12 These results implicate local inflammation and activation of the complement cascade as an important mechanism in AMD development. 

Vitamin D, a circulating steroid hormone, has properties that counteract inflammation, ^13,14 as well as angiogenesis, ^15,16 oxidative stress, ^14,17,18 and fibrosis. ^19,20 A number of studies demonstrate an antiinflammatory function of vitamin D in vitro and in vivo. ^21–23 Several human studies have shown an inverse relationship between vitamin D and several chronic conditions associated with inflammation. ^24–27 In the eye, vitamin D may prevent AMD progression via its antiinflammatory and antiangiogenic properties. ²⁸ However, epidemiologic studies on the relationship between vitamin D and AMD have been limited, and the results are inconsistent. To our knowledge, only three epidemiologic studies have been performed thus far. ^29–31 A study using the Third National Health and Nutrition Examination Survey (NHANES) from 20 years ago reported that serum vitamin D levels were inversely associated with early, but not advanced, AMD. Another study found no association between vitamin D levels and AMD. ³⁰ More recently, an association between vitamin D and early AMD was shown in 1313 women aged 50 to 79 years. ²⁹ However, all of these prior studies had at least 1 of the following limitations. Although late AMD is related to vision loss, the numbers of participants with late AMD were insufficient or absent. ^29–31 In another study, AMD diagnosis was based on the disease code rather than a fundus photograph. ³⁰ In addition, there is a lack of information on sunlight exposure, by which approximately 90% of vitamin D is generated in the skin. ^30,31 Finally, only highly selected groups of subjects (for example, postmenopausal women) were investigated. ²⁹  

To address these issues, this study investigated the relationship between 25-hydroxyvitamin D and AMD, especially late AMD, in a large representative sample of Korean adults. 

Methods

Study Population

This study used data acquired for the Korean National Health and Nutrition Examination Survey (KNHANES). The KNHANES is a nationwide and population-based, cross-sectional study conducted by the Korean Ministry of Health and Welfare and the Division of Chronic Disease Surveillance, Korean Center for Disease Control and Prevention. The KNHANES adopted a rolling sampling design, which is a stratified, complex, multistage, probability cluster survey with proportional allocation based on the National Census Registry for the noninstitutional Korean civilian population. Details regarding the study design and methods are provided elsewhere. ^32,33 Data for the present study were obtained from the fourth (2008–2009) and fifth (2010–2012) KNHANES to estimate the association between blood 25-hydroxyvitamin D levels and AMD. For the current study, 35,056 individuals for whom blood 25-hydroxyvitamin D levels were obtained were selected. Of these, 15,942 subjects aged younger than 40 years and 2069 subjects who did not undergo fundus examination were excluded. Finally, 17,045 participants aged 40 years and older were included in the analysis (Fig. 1). The study design followed the tenets of the Declaration of Helsinki for biomedical research. Protocols for this study were approved by the institutional review board of the Catholic University of Korea (Seoul, Korea). All participants signed and provided written informed consent. 

Figure 1

 

Flow diagram presenting the selection of study participants.

Data Collection

Retinal examination was done under physiological mydriasis and AMD was graded as early or late AMD using the international classification and grading system. ³⁴ Digital fundus images were obtained under physiological mydriasis using a digital fundus camera (TRC-NW6S; Topcon, Tokyo, Japan). For each participant, a 45° digital retinal image, centered on the fovea, was obtained for each eye (2 images per person). Each image was graded twice, one preliminary and one detailed. A preliminary grading was done at the photography by trained dispatched ophthalmologists. Detailed grading was performed later by nine retinal specialists who were masked to the subject's characteristics. Final grading was based on detailed grading, and any discrepancy between preliminary and detailed grading was resolved by one reading specialist. ³⁵  

Early AMD was defined as if fundus photograph has one of two criteria: (1) the presence of soft indistinct drusen or reticular drusen, or (2) the presence of hard or soft with distinct drusen with pigment abnormality without signs of late AMD. ³⁵ Late AMD was defined as the presence of wet or dry (geographic atrophy) AMD. Wet AMD was defined as detachment of the RPE or neurosensory retina, the presence of hemorrhages in the subretinal or sub-RPE space, or a disc-form scar in the macular area. Dry AMD was defined as a circular, discrete depigmented area greater than or equal to 175 μm in diameter with visible choroidal vessels. For subjects with AMD lesions in only one eye or asymmetric AMD lesions in both eyes, AMD was defined according to the more affected eye. 

Demographic information was collected during a health interview. Height and weight measurements were obtained with subjects wearing light clothes without shoes. Body mass index was calculated as follows: weight (kg)/height (m). ² Age was classified in 10-year intervals. Smoking status was self-reported, and subjects were classified as current smoker, past-smoker, or never-smoker. Alcohol use was self-reported as an ever-drinker or never-drinker. Data on current sunlight exposure time were obtained from two possible answers to a single question: less than 5 hours or greater than or equal to 5 hours per day. 

Blood samples were collected after an 8-hour fast. Fasting glucose, hemoglobin A1c, total cholesterol, and triglycerides were measured using a Hitachi automatic analyzer 7600 (Hitachi, Tokyo, Japan). 25-hydroxyvitamin D was measured using a radioimmunoassay kit (DiaSorin Inc., Stillwater, MN, USA) with a gamma counter (1470 WIZARD; Perkin-Elmer, Finland). Details of the 25-hydroxyvitamin D analysis have been reported previously. ^36,37 All blood analyses were properly processed, promptly refrigerated, and transported in cold storage to Neodin Medical Institute (Seoul, Korea), a laboratory certified by the Korean Ministry of Health and Welfare. The interassay coefficients of variation were 2.8% to 6.2% for the 2008 to 2009 samples and 1.9% to 6.1% for the 2010 to 2012 samples. The KNHANES participates in the Vitamin D Standardization Program, thus measurement of 25-hydroxyvitamin D was standardized with the recently developed National Institute of Standards and Technology-Ghent University reference procedure. ³⁸  

Blood pressure was measured with a sphygmomanometer with subjects in a sitting position. Three measurements were taken at 5-minute intervals, and the average of the second and third measurements was used for the analysis. The presence of diabetes mellitus was defined as a fasting glucose of greater than or equal to 126 mg/dL or subjects on antiglycemic medication. The presence of hypertension was defined as systolic blood pressure greater than or equal to 140 mm Hg, diastolic blood pressure greater than or equal to 90 mm Hg, or subjects on antihypertensive medication. Heart problems were defined as a history of myocardial infarction or angina, and stroke problems were self-reported. 

Statistical Analyses

Statistical analyses were performed using SPSS ver. 18.0 (SPSS, Inc., Chicago, IL, USA). To account for the complex sampling design, strata, sampling units, and sampling weights were used to obtain unbiased point estimates and robust linearized standard errors. Participants' characteristics were described using means and standard errors for continuous variables and percentages and standard errors for categorical variables according to the presence of AMD. Analysis of variance or chi-square tests were used to compare demographic characteristics. 

To evaluate the effect of blood 25-hydroxyvitamin D levels on AMD prevalence, blood 25-hydroxyvitamin D levels were categorized into quintiles. ³⁹ Simple and multiple logistic regression analyses examined the association between blood 25-hydroxyvitamin D levels and AMD. After calculation of the crude odds ratio (OR; Model 1), values were adjusted for age and sex (Model 2). They were then adjusted for age, sex, and other confounders including smoking, hypertension, heart problem, stroke, and sunlight exposure time as these have been established as risk factors in previous studies (Model 3). ^10,40,41 All variables for logistic regression analysis were examined for multicollinearity, and only variables with a variance inflation factor less than 10 were used. P values were two-tailed, and P less than 0.05 indicated statistical significance. 

Results

Of 19,114 eligible subjects aged older than 40 years for whom blood 25-hydroxyvitamin D were measured, the fundus was examined in 17,045 (90.6%). Reasons that fundus photography was not performed for all subjects include a small pupil (39.4%), cataracts (29.7%), poor cooperation (11.8%), refusal (4.3%), cornea opacity (5.2%), or miscellaneous other reasons (9.6%). Thus, 17,045 subjects were included in the current analysis. The demographic characteristics of 17,045 enrolled subjects are summarized by AMD status in Table 1. Subjects with AMD were more likely to be older (P < 0.001) or smokers (P < 0.001) and have higher systolic blood pressure (P < 0.001), longer sun exposure (P < 0.001), higher blood 25-hydroxyvitamin D levels (P = 0.008), and hypertension (P < 0.001) than those without AMD. 

Table 1

 

Demographic and Clinical Characteristics, According to Early- and Late-AMD Status and Participation Status, as Reported in the KHANES 2008–2012

Characteristics	No AMD n = 15,767	Early AMD n = 1163	Late AMD n = 115	P	Participants n = 17,045
Male, %	48.3 (0.4)	40.8 (1.7)	62.6 (5.4)	<0.001*	47.9 (0.4)
Age, y	54.3 (0.1)	64.9 (0.3)	65.9 (1.3)	<0.001*	61.6 (0.4)
Body mass index, kg/m2	24.1 (0.0)	23.7 (0.1)	23.6 (0.3)	0.018*	23.8 (0.1)
Systolic blood pressure, mm Hg	121.9 (0.2)	127.3 (0.6)	132.3 (2.0)	<0.001*	127.2 (0.7)
Diastolic blood pressure, mm Hg	78.5 (0.1)	77.2 (0.3)	78.3 (1.1)	0.004*	78.0 (0.4)
Fasting glucose, mg/dL	100.9 (0.2)	101.1 (0.7)	96.5 (1.2)	0.002*	99.5 (0.5)
HbA1c, %	6.0 (0.0)	6.0 (0.0)	5.8 (0.1)	0.032*	6.0 (0.0)
Total cholesterol, mg/dL	193.4 (0.3)	192.1 (1.2)	192.9 (4.7)	0.600	192.8 (1.6)
Triglyceride, mg/dL	148.3 (1.4)	137.7 (3.6)	146.1 (11.7)	0.022*	144.1 (4.1)
25-hydroxyvitamin D, ng/mL	18.6 (0.1)	19.6 (0.2)	19.1 (0.8)	0.001*	19.1 (0.3)
Diabetes, %	12.3 (0.4)	15.0 (1.3)	7.2 (2.6)	0.020*	12.4 (0.3)
Hypertension, %	37.6 (0.6)	52.5 (1.8)	56.9 (5.6)	<0.001*	38.6 (0.5)
Heart problems, %	2.9 (0.2)	4.9 (0.7)	8.3 (3.7)	0.001*	3.0 (0.2)
Stroke, %	2.0 (0.1)	3.8 (0.7)	3.2 (1.5)	0.001*	2.1 (0.1)
Sun exposure, ≥5 h/d, %	20.0 (0.7)	26.1 (1.8)	23.7 (5.0)	<0.001*	20.4 (0.7)
Smoking status				<0.001*
Never, %	54.3 (0.4)	60.7 (1.8)	37.9 (5.5)		54.6 (0.4)
Former, %	15.7 (0.4)	15.0 (1.3)	22.6 (5.3)		15.7 (0.4)
Current, %	30.0 (0.5)	24.3 (1.5)	39.5 (5.6)		29.8 (0.4)
Smoking dose, cigarette/d	17.1 (0.2)	15.8 (0.7)	16.3 (2.4)	0.241	16.4 (0.8)
Alcohol consumption, %	85.3 (0.4)	75.2 (1.5)	87.6 (3.7)	<0.001*	84.7 (0.3)

 Data are expressed as weighted means or weighted frequency (%) with standard errors.

*   P < 0.05.

Demographic and clinical characteristics by blood 25-hydroxyvitamin D quintiles are presented in Table 2. As blood 25-hydroxyvitamin D levels increased, participants were more likely to be male (P for trend < 0.001), older (P for trend < 0.001), nonsmokers (P for trend < 0.001), hypertensive (P for trend = 0.015), have high body mass index (P for trend < 0.001), and longer sun exposure (P < 0.001). 

Table 2

 

Demographic and Clinical Characteristics by Quintile Blood 25-Hydroxyvitamin D Categories Among Representative Korean Adults Aged 40 Years or Older

Characteristics	Quintile Blood 25-Hydroxyvitamin D Level, ng/mL					P for Trend
	<13.1 n = 3427	13.1–16.4 n = 3421	16.4–19.7 n = 3405	19.7–24.3 n = 3383	>24.3 n = 3409
Male, %	33.7 (1.0)	42.4 (1.1)	50.2 (1.0)	57.7 (1.1)	61.6 (1.0)	<0.001*
Age, y	53.6 (0.2)	53.4 (0.2)	54.2 (0.2)	55.3 (0.2)	56.7 (0.3)	<0.001*
Body mass index, kg/m2	23.9 (0.1)	24.2 (0.1)	24.1 (0.1)	24.2 (0.1)	23.7 (0.1)	<0.001*
Systolic blood pressure, mm Hg	122.1 (0.4)	121.4 (0.3)	121.8 (0.4)	122.7 (0.4)	123.0 (0.4)	0.021*
Diastolic blood pressure, mm Hg	78.1 (0.2)	78.4 (0.2)	78.6 (0.2)	79.1 (0.2)	78.5 (0.3)	0.125
Fasting glucose, mg/dL	100.7 (0.6)	100.0 (0.4)	102.5 (0.6)	100.2 (0.4)	100.9 (0.5)	0.015*
HbA1c, %	6.05 (0.04)	5.97 (0.02)	6.07 (0.03)	5.95 (0.03)	6.08 (0.04)	0.005*
Total cholesterol, mg/dL	192.7 (0.8)	193.2 (0.7)	194.5 (0.8)	195.0 (0.7)	191.4 (0.7)	0.005*
Triglyceride, mg/dL	153.5 (4.0)	149.3 (2.7)	147.9 (2.5)	147.6 (2.5)	139.7 (2.0)	0.005*
Diabetes, %	12.6 (0.7)	10.9 (0.6)	13.5 (0.8)	11.9 (0.7)	13.2 (0.7)	0.049*
Hypertension, %	37.4 (1.1)	37.4 (1.1)	36.3 (1.0)	39.7 (1.0)	40.7 (1.1)	0.015*
Heart problems, %	2.5 (0.3)	3.1 (0.3)	2.7 (0.3)	2.6 (0.3)	3.8 (0.4)	0.042*
Stroke, %	1.9 (0.3)	1.7 (0.3)	1.7 (0.2)	2.7 (0.3)	2.2 (0.3)	0.044*
Sun exposure, ≥5h/d, %	11.2 (0.7)	14.6 (0.9)	17.5 (1.0)	24.8 (1.3)	35.2 (1.6)	<0.001*
Smoking status						<0.001*
Current, %	64.5 (1.0)	58.0 (1.1)	55.0 (1.0)	47.7 (1.1)	44.7 (1.1)
Former, %	10.5 (0.7)	14.5 (0.8)	17.4 (0.8)	20.4 (1.0)	15.6 (0.9)
Never, %	25.0 (1.0)	27.5 (1.0)	27.6 (1.0)	31.9 (1.1)	39.7 (1.1)
Smoking dose, cigarette/d	16.1 (0.4)	16.2 (0.4)	16.6 (0.4)	17.2 (0.4)	18.9 (0.5)	<0.001*
Alcohol consumption, %	82.0 (0.8)	85.9 (0.7)	86.3 (0.7)	86.4 (0.7)	86.1 (0.7)	<0.001*

*   P < 0.05.

Blood 25-hydroxyvitamin D levels of men and women based on categories such as age, hypertension, smoking status, diabetes, and alcohol consumption are presented in Table 3. Blood 25-hydroxyvitamin D levels were 20.0 ng/mL (95% confidence interval [CI], 19.7–20.3) in men and 17.5 ng/mL (95% CI, 17.2–17.7) in women (P < 0.001). In both sexes, there were significantly higher mean blood levels of 25-hydroxyvitamin D in the older age group (P < 0.001) than the younger age group. However, participants with diabetes showed higher blood 25-hydroxyvitamin D levels in women (P = 0.032) but not in men (P = 0.082). There were significant differences of blood 25-hydroxyvitamin D levels between women with and without hypertension (P = 0.005), but not in men (P = 0.720). Mean blood 25-hydroxyvitamin D levels were significantly higher in those with sunlight exposure greater than 5 hours per day than those with exposure less than 5 hours per day in both sexes (P < 0.001). 

Table 3

 

Sex Difference of Blood Vitamin D Levels (ng/mL) of Men and Women According to Age Group and Other Variables Among Representative Korean Adults Aged 40 Years or Older

Characteristics	Blood 25–Hydroxyvitamin D Level, ng/mL
	Men, n = 7367	P for Trend	Women, n = 9678	P for Trend
All subjects aged 40+ y	20.0 (19.7–20.3, 0.14)		17.5 (17.2–17.7, 0.12)
Age groups		<0.001		<0.001
40–49 y	19.2 (18.9–19.6, 0.19)		16.2 (15.9–16.5, 0.14)
50–59 y	20.3 (19.8–20.7, 0.20)		18.0 (17.6–18.3, 0.17)
60–69 y	20.9 (20.4–21.3, 0.23)		18.4 (18.1–18.8, 0.20)
70+	20.5 (20.0–21.1, 0.28)		18.6 (18.1–19.1, 0.25)
AMD		0.018		<0.001
None	19.9 (19.6–20.2, 0.15)		17.4 (17.1–17.6, 0.12)
AM	20.9 (20.1–21.7, 0.41)		18.6 (18.0–19.1, 0.29)
Diabetes		0.082		0.032
Nondiabetes	20.1 (19.8–20.4, 0.15)		17.4 (17.2–17.6, 0.12)
Diabetes	19.6 (19.1–20.1, 0.27)		18.0 (17.4–18.6, 0.30)
Hypertension		0.720		0.005
Nonhypertension	20.1 (19.7–20.4, 0.17)		17.3 (17.1–17.6, 0.13)
Hypertension	20.0 (19.6–20.3, 0.18)		17.8 (17.5–18.2, 0.17)
Sun exposure		<0.001		<0.001
<5h/d	19.1 (18.8–19.4, 0.15)		17.0 (16.8–17.3, 0.11)
≥5h/d	22.2 (21.6–22.7, 0.27)		20.0 (19.4–20.6, 0.30)
Smoking status		0.058		0.003
Never	20.1 (19.6–20.6, 0.25)		17.5 (17.3–17.8, 0.12)
Former	19.6 (19.2–20.5, 0.19)		16.2 (15.5–17.0, 0.37)
Current	20.2 (19.8–20.5, 0.18)		17.3 (16.7–18.0, 0.33)
Alcohol consumption		0.340		0.052
Never	20.4 (19.5–21.2, 0.42)		17.8 (17.4–18.2, 0.20)
Ever	20.0 (19.7–20.3, 0.15)		17.4 (17.1–17.6, 0.12)

Age-related macular degeneration prevalence based on quintiles of blood 25-hydroxyvitamin D are shown in Figure 2. As blood 25-hydroxyvitamin D levels increase, prevalence of any AMD significantly increased from 5.7% in the first quintile to 7.9% in the fifth quintile (P for trend = 0.001). In addition, the prevalence of early AMD significantly increased with increasing blood 25-hydroxyvitamin D quintiles (P for trend = 0.001). However, there was no significant change in the prevalence of late AMD with blood 25-hydroxyvitamin D quintiles (P for trend = 0.291). 

Figure 2

 

Prevalence of any, early, and late AMD according to quintiles of blood 25-hydroxyvitamin D levels in a representative Korean population.

Table 4 shows odds ratio (ORs) for the association of AMD with blood 25-hydroxyvitamin D levels. There was no significant association of any, early, or late AMD with blood 25-hydroxyvitamin D quintiles after adjusting for potential covariates such as age, sex, smoking, hypertension, heart problems, stroke, and sunlight exposure. The adjusted ORs for any, early, and late AMD were 1.20 (95% CI, 0.95–1.51), 1.26 (95% CI, 0.99–1.61), 0.75 (95% CI, 0.33–1.58) respectively, among those within the highest quintile of blood 25-hydroxyvitamin D levels compared with participants within the lowest quintile of blood 25-hydroxyvitamin D levels. 

Table 4

 

Prevalence and Adjusted OR of Early and Late AMD, Stratified According to Quintile Categories of Blood 25-Hydroxyvitamin D Among Representative Korean Adults Aged 40 Years or Older

	Quintile Blood 25–Hydroxyvitamin D Level, ng/mL					P for Trend
	<13.1 n = 3427	13.1–16.4 n = 3421	16.4–19.7 n = 3405	19.7–24.3 n = 3383	>24.3 n = 3409
Any AMD
Prevalence	5.7 (4.9–6.6, 0.4)	5.2 (4.4–6.1, 0.4)	6.0 (5.2–6.9, 0.4)	6.4 (5.5–7.4, 0.5)	7.9 (6.9–9.0, 0.6)	0.001*
number	214	235	245	265	319
OR
Model 1	1.00 (reference)	0.91 (0.72–1.15)	1.05 (0.84–1.30)	1.13 (0.91–1.40)	1.42 (1.14–1.76)*	<0.001*
Model 2	1.00 (reference)	0.96 (0.75–1.21)	1.06 (0.84–1.32)	1.05 (0.83–1.32)	1.20 (0.96–1.50)	0.080
Model 3	1.00 (reference)	0.94 (0.74–1.20)	1.08 (0.86–1.36)	1.06 (0.84–1.34)	1.20 (0.95–1.51)	0.071
Early AMD
Prevalence	5.0 (4.3–5.9, 0.4)	4.8 (4.0–5.7, 0.4)	5.5 (4.8–6.4, 0.4)	5.6 (4.8–6.5, 0.4)	7.3 (6.3–8.4, 0.5)	0.001*
number	191	214	225	239	294
OR
Model 1	1.00 (reference)	0.94 (0.73–1.22)	1.10 (0.87–1.37)	1.12 (0.88–1.41)	1.47 (1.17–1.85)*	<0.001*
Model 2	1.00 (reference)	1.00 (0.77–1.29)	1.12 (0.88–1.42)	1.05 (0.82–1.34)	1.27 (1.01–1.61)*	0.047*
Model 3	1.00 (reference)	0.97 (0.75–1.26)	1.13 (0.89–1.43)	1.04 (0.81–1.34)	1.26 (0.99–1.61)	0.053
Late AMD
Prevalence	0.6 (0.4–1.1, 0.2)	0.4 (0.2–0.7, 0.1)	0.4 (0.2–0.7, 0.1)	0.8 (0.5–1.3, 0.2)	0.6 (0.4–1.0, 0.1)	0.291
number	23	21	20	26	25
OR
Model 1	1.00 (reference)	0.62 (0.29–1.32)	0.66 (0.30–1.43)	1.22 (0.59–2.51)	0.98 (0.49–1.93)	0.459
Model 2	1.00 (reference)	0.62 (0.29–1.34)	0.58 (0.26–1.32)	0.96 (0.44–2.09)	0.67 (0.32–1.40)	0.684
Model 3	1.00 (reference)	0.69 (0.30–1.57)	0.68 (0.29–1.56)	1.13 (0.51–2.51)	0.75 (0.33–1.58)	0.896

*   P < 0.05.

Age-related macular degeneration prevalence in men and women by quintiles of blood 25-hydroxyvitamin D is shown in Table 5. The prevalence of early AMD was positively correlated with increased blood 25-hydroxyvitamin D in both sexes (P for trend = 0.026 for men and 0.013 for women). However, the prevalence of late AMD was negatively correlated with increased blood 25-hydroxyvitamin D in men (P for trend = 0.035) not in women (P for trend = 0.051). 

Table 5

 

Sex Difference of Prevalence and Adjusted OR of AMD, Stratified According to Quintile Categories of Blood 25-Hydroxyvitamin D Among Representative Korean Adults Aged 40 Years or Older

	Quintile Blood 25–Hydroxyvitamin D Level, ng/mL
	Quintile 1	Quintile 2	Quintile 3	Quintile 4	Quintile 5	P for Trend
Male
Range	<14.5 (n = 1473)	14.5–17.9 (n = 1469)	17.9–21.2 (n = 1478)	21.2–25.7 (n = 1473)	>25.7 (n = 1474)
Any AMD
Prevalence	5.0 (4.0–6.3, 0.6)	4.8 (3.8–6.1, 0.6)	5.2 (4.2–6.5, 0.6)	5.7 (4.4–7.3, 0.7)	7.3 (5.8–9.0, 0.8)	0.071
number	88	105	108	103	132
OR
Model 1	1.00 (reference)	0.95 (0.67–1.35)	1.05 (0.75–1.47)	1.14 (0.80–1.64)	1.48 (1.05–2.09)*	0.014*
Model 2	1.00 (reference)	0.97 (0.68–1.39)	0.96 (0.68–1.36)	1.03 (0.71–1.50)	1.23 (0.86–1.75)	0.239
Model 3	1.00 (reference)	1.01 (0.70–1.46)	1.02 (0.71–1.46)	1.04 (0.70–1.53)	1.21 (0.84–1.75)	0.324
Early AMD
Prevalence	3.9 (3.0–5.1, 0.5)	4.4 (3.4–5.6, 0.6)	4.8 (3.8–6.1, 0.6)	4.7 (3.6–6.2, 0.7)	6.8 (5.4–8.5, 0.8)	0.026
number	73	92	96	91	120
OR
Model 1	1.00 (reference)	1.11 (0.76–1.63)	1.23 (0.85–1.77)	1.21 (0.81–1.80)	1.77 (1.23–2.57)*	0.003
Model 2	1.00 (reference)	1.13 (0.76–1.67)	1.12 (0.77–1.65)	1.08 (0.72–1.64)	1.46 (1.00–2.15)*	0.083
Model 3	1.00 (reference)	1.16 (0.78–1.73)	1.18 (0.79–1.75)	1.07 (0.70–1.64)	1.45 (0.97–2.16)	0.126
Late AMD
Prevalence	1.1 (0.6–2.0, 0.3)	0.5 (0.2–1.0, 0.2)	0.5 (0.3–0.9, 0.2)	0.5 (0.2–1.0, 0.3)	0.5 (0.3–1.0, 0.2)	0.035*
number	15	13	12	12	12
OR
Model 1	1.00 (reference)	0.40 (0.15–1.06)	0.43 (0.18–1.77)	0.45 (0.19–1.11)	0.46 (0.19–1.12)	0.145
Model 2	1.00 (reference)	0.41 (0.15–1.10)	0.39 (0.16–0.93)	0.38 (0.16–0.93)	0.37 (0.15–0.91)*	0.022*
Model 3	1.00 (reference)	0.46 (0.16–1.28)	0.44 (0.18–1.06)	0.40 (0.16–1.01)	0.32 (0.12–0.81)*	0.018*
Female
Range	<12.3 (n = 932)	12.3–15.3 (n = 1938)	15.3–18.3 (n = 1929)	18.3–22.7 (n = 1940)	>22.7 (n = 1939)
Any AMD
Prevalence	5.6 (4.6–6.8, 0.6)	6.0 (4.9–6.8, 0.6)	6.7 (5.5–8.2, 0.7)	7.0 (5.8–8.4, 0.7)	9.0 (7.6–10.7, 0.8)	0.004*
number	126	130	137	162	187
OR
Model 1	1.00 (reference)	1.20 (0.92–1.55)	1.11 (0.85–1.46)	1.49 (1.11–2.00)*	1.66 (1.27–2.17)*	<0.001*
Model 2	1.00 (reference)	1.20 (0.91–1.57)	1.01 (0.77–1.34)	1.20 (0.88–1.64)	1.18 (0.90–1.56)	0.112
Model 3	1.00 (reference)	1.10 (0.79–1.53)	1.25 (0.91–1.73)	1.16 (0.84–1.59)	1.29 (0.96–1.74)	0.087
Early AMD
Prevalence	5.3 (4.3–6.5, 0.6)	5.7 (4.6–7.0, 0.6)	6.5 (5.3–8.0, 0.7)	6.3 (5.1–7.6, 0.6)	8.4 (7.1–10.0, 0.8)	0.013
number	118	122	129	148	174
OR
Model 1	1.00 (reference)	1.09 (0.79–1.50)	1.26 (0.91–1.73)	1.20 (0.88–1.63)	1.66 (1.24–2.22)*	0.001
Model 2	1.00 (reference)	1.14 (0.82–1.59)	1.29 (0.93–1.79)	1.08 (0.78–1.49)	1.27 (0.93–1.73)	0.201
Model 3	1.00 (reference)	1.11 (0.79–1.56)	1.28 (0.92–1.78)	1.07 (0.77–1.48)	1.26 (0.92–1.71)	0.202
Late AMD
Prevalence	0.3 (0.2–0.8, 0.1)	0.3 (0.1–0.6, 0.1)	0.2 (0.1–0.4, 0.1)	0.8 (0.4–1.6, 0.3)	0.7 (0.4–1.2, 0.2)	0.051
number	8	8	8	14	13
OR
Model 1	1.00 (reference)	0.76 (0.25–2.32)	0.61 (0.20–1.84)	2.25 (0.75–6.77)	1.93 (0.69–5.34)	0.037*
Model 2	1.00 (reference)	0.82 (0.27–2.46)	0.63 (0.21–1.91)	2.09 (0.70–6.23)	1.41 (0.52–3.85)	0.146
Model 3	1.00 (reference)	0.91 (0.29–2.87)	0.72 (0.23–2.23)	2.60 (0.86–7.85)	1.90 (0.66–5.44)	0.059

 Prevalence was expressed as weighted estimates (%; 95% CIs, standard errors [%]). Model 1: Crude OR. Model 2: adjusted for age. Model 3: adjusted for age, smoking, hypertension, heart problem, stroke, and sun exposure.

*   P < 0.05.

In men, the adjusted OR for early AMD after adjusting for potential confounders was 1.45 (95% CI, 0.97–2.16) in those within the highest 25-hydroxyvitamin D quintile than those within the lowest quintile. However, the adjusted OR for late AMD in men was 0.32 (95% CI, 0.12–0.81, P for trend = 0.018) in highest 25-hydroxyvitamin D quintile than the lowest one. In women, there was no significant association between blood 25-hydroxyvitamin D levels and any, early, or late AMD. 

Discussion

The current study showed that prevalence of late AMD decreased in individuals with the highest blood 25-hydroxyvitamin D levels compared with those with the lowest levels. In addition, this association of blood 25-hydroxyvitamin D levels occurred in men, but not in women. Finally, prevalence of early AMD was not inversely associated with blood 25-hydroxyvitamin D levels. 

The risk of late AMD in men was 68% less in subjects within the highest quintile of blood 25-hydroxyvitamin D levels than in those within the lowest quintile after adjusting for age, smoking status, hypertension, heart problems, stroke, and sun exposure time. These results are contrary to those of a previous epidemiologic study, which showed no significant association between blood 25-hydroxyvitamin D levels and advanced AMD. However, this previous study was limited by a small number of patients in the advanced AMD group (n = 54). ³¹ One possible reason for the discrepancy is the difference in average blood 25-hydroxyvitamin D levels between the two studies. Mean (standard error) blood 25-hydroxyvitamin D level in the present study was 19.1 (0.3) ng/mL, which is lower than those from the US NHANES (23.0 [0.1] ng/mL). Ethnic differences may be one cause for this difference. It is reported that serum vitamin D levels are lower in Asians than Caucasians. ⁴² Another potential explanation for this discrepancy is that vitamin D deficiency may have increased in the 20 years between the US NHANES III and the present study. ⁴³ Industrialization has reduced exposure to sunlight, by which approximately 90% of vitamin D is generated in the skin. 

The most likely biological explanation for the association in the present study is that vitamin D inhibits angiogenesis and fibrosis, which are the most critical characteristics of late AMD. Vitamin D inhibits angiogenesis by reducing the expression of VEGF, reducing endothelial cell proliferation, and increasing the expression of platelet-derived growth factor. ^15,16,24 Moreover, vitamin D also inhibits matrix metalloproteinase-9, which plays a role in choroidal neovascularization. ⁴⁴ In addition to the antiangiogenic effects of vitamin D, it is also a potent inhibitor of fibrosis. Development of late AMD is closely implicated with growth factors regulating fibrotic changes, including TGF-β. ⁴⁵ Transforming growth factor–β is a potent promoter of fibrogenesis through modulation of fibroblast phenotype and function, myofibroblast transdifferentiation, and matrix preservation. ^46–49 Vitamin D has an inverse relationship with TGF-β. ⁵⁰ The present findings suggest that vitamin D may play a key role in the inhibition of late AMD development. 

Sex differences in the association of blood 25-hydroxyvitamin D with late AMD were found in the present study. Although men with the highest blood 25-hydroxyvitamin D levels had 68% less risk for late AMD than those with the lowest levels after adjusting for potential confounders, this association was not found in women. Moreover, a sex-related effect modification was observed. The direction of the OR for late AMD in women (OR, 1.90) was opposite to that in men (OR, 0.32), which would explain the lack of an overall association (Table 4). It is unclear why 25-hydroxyvitamin D is inversely correlated with prevalence of late AMD in men, but not women. It is possible that the mechanism of action of vitamin D on late AMD is different between sexes. Further study is needed to identify factors responsible for this difference, especially to elucidate the exact sex-specific biologic mechanisms by which 25-hydroxyvitamin D inhibits the development of late AMD. 

For early AMD, association of 25-hydroxyvitamin D was not significant in either sex. One interesting finding is that the prevalence of early AMD increased with higher 25-hydroxyvitamin D levels (P = 0.001), and the crude OR for early AMD in the highest quintile of 25-hydroxyvitamin D was significantly higher than in the lowest quintile (crude OR = 1.77 for men and 1.66 for women). Although adjustment for potential confounders attenuated this relationship to be nonsignificant, there are marginal significance. This raises a possible positive relationship between two variables. One possible explanation is that 25-hydroxyvitamin D is correlated by sunlight exposure, because most 25-hydroxyvitamin D is produced in the skin by UV sunlight. Although sunlight exposure was adjusted in Model 3, classification of sunlight exposure into greater than 5 hours a day or not is crude and insufficient to adjust the effect of sunlight on AMD. Sunlight has been demonstrated to be a risk factor for AMD development in many experimental and epidemiologic studies, ^51–55 although some results are controversial. ⁵⁶ A recent meta-analysis examining 14 studies confirmed that sunlight exposure is indeed a risk factor for AMD. ⁵⁷ Especially, the recent ozone depletion and subsequent increase of solar UV radiation raises the concern about the detrimental effect of sunlight exposure on AMD development. ^58,59 The present findings suggest that early AMD may be influenced by the adverse effects of UV sunlight exposure rather than vitamin D, byproduct of sunlight. 

The major strength of the present study is the relatively large number of participants (n = 17,045) and the study design using systemic stratified, multistage, clustered, random sampling methods. Another strength is the rigorous quality control for ophthalmic examination of the fundus and measurement of blood 25-hydroxyvitamin D in KNHANES. This study also has several limitations. First, it was not possible to adjust for seasonal variation of 25-hydroxyvitamin D, because KNHANES does not have information about the examination date. However, a recent study showed that an Asian population displayed no significant seasonal variation in vitamin D status. ⁴² Second limitation is that the current study has a cross-sectional design, which makes inferring causality difficult. There is possibility that people with late AMD may have low vision, and therefore less mobile, spending more time indoors, thereby reducing their sun exposure time that is positively associated with 25-hydroxyvitamin D levels. The third limitation is that ascertainment of wet AMD may be difficult in such eyes that have been treated with antivascular endothelial growth factor therapies. This would not be evident on fundus photographs and would require the reported history of antivascular endothelial growth factor therapies on which KHANES does not have. Finally, evaluation of sunlight exposure was relatively crude. Effect of sunlight exposure may be variable depending on the months of the year and the geographic location of the residence (latitude). 

In conclusion, the present study provides population-based epidemiologic evidence of a relationship between blood 25-hydroxyvitamin D with AMD in a representative Korean population. Blood 25-hydroxyvitamin D levels were inversely associated with late AMD in men, but not in women. However, there was no such relationship for early AMD. The differential effect of 25-hydroxyvitamin D on early and late AMD implies a potentially different pathogenesis of early and late AMD. 25-hydroxyvitamin D may have a preventive effect on the development of late AMD, the most functionally disabling stage of the condition, through antiangiogenic and antifibrotic actions. Considering that the worldwide prevalence of vitamin D deficiency has increased during the past few decades, association of vitamin D and AMD warrants further studies. 

Acknowledgments

The authors thank the Epidemiologic Survey Committee of the Korean Ophthalmologic Society for conducting examinations in KNHANES and for supplying data for this study. 

No author has a financial or proprietary interest in any material or method mentioned. 

Disclosure: E.C. Kim, None; K. Han, None; D. Jee, None