Association Between Blood Cadmium Level and Age-Related Macular Degeneration in a Representative Korean Population

Abstract

Purpose.: Toinvestigate the association between blood cadmium level and AMD.

Methods.: This population-based, cross-sectional study using a nationwide, systemically stratified, multistage, clustered sampling method included 4933 subjects older than 40 years who participated in the Korean National Health and Nutrition Examination Survey (KNHANES) between 2008 and 2012 and had fundus photographs taken. All participants underwent a standardized interview, evaluation of blood cadmium concentrations, and comprehensive ophthalmic examination. A 45° digital fundus photograph was taken from both eyes under physiologic mydriasis, and were graded using the international classification and grading system for AMD.

Results.: Mean blood cadmium levels were 1.47 μg/L in women and 1.19 μg/L in men. After adjusting for potential confounders, including age, sex, and smoking status, the odds ratio (OR) for AMD was significantly increased in the highest quintile blood cadmium group (OR, 1.96; 95% confidence interval [CI], 1.17–3.29; P for trend = 0.017). This association between blood cadmium level and AMD was significant in men (OR, 2.11; 95% CI, 1.11–4.02; P for trend = 0.024), but not in women (OR, 1.29; 95% CI, 0.70–2.52; P for trend = 0.158).

Conclusions.: This study provides the first epidemiologic evidence that higher blood cadmium level is associated with AMD. Results of the present study indicate that an elevated cadmium burden may increase the risk of AMD development.

Introduction

Age-related macular degeneration (AMD) is the leading cause of blindness in the elderly. ¹ Although the exact etiology of this disease remains to be elucidated, it is known to be a multifactorial condition influenced by the interaction of genetic and environmental factors. ² We previously reported associated risk factors for AMD in a representative Korean population. ³ However, established risk factors for AMD, including age and smoking, ^1,4 cannot completely explain the distribution of AMD in the general population. Thus, there is need to identify novel risk factors. 

Cadmium is a naturally occurring, toxic metabolic element, the prevalence of which is increasing in the environment, as it is a product of industrial activities and a contaminant in fertilizers. ⁵ Cadmium is also concentrated in tobacco and food plants. ⁶ Cadmium uptake usually occurs by dietary intake or inhalation of polluted air. The half-life of cadmium is very long (10–30 years), and currently there are no specific therapies to treat cadmium exposure and toxicity. ⁵ Thus, cadmium accumulates with age in the retina and RPE. ⁷  

Cadmium is a potent inflammatory agent that increases oxidative stress. Furthermore, cadmium induces damage in most cells of the body, including neurons, retinal cells, and renal tubules, and this is caused in part by increased levels of reactive oxygen species. ⁸ In an experimental study, low levels of cadmium exposure were toxic to cultured HRPE cells, and it disrupted membrane integrity and increased oxidative stress. ⁹ Retinal cells and the RPE routinely adequately cope with reactive oxygen species generated by metabolic processes and light. However, the levels of specific antioxidant enzymes in the retina decrease with age. Thus, the survival of cells following oxidative stress decreases in older persons. ¹⁰ Furthermore, there is increasing evidence that AMD may be induced by oxidative stress, decreased antioxidant capacity, and inflammation in the choroid and retina. ¹¹  

Recent research has indicated that cadmium is a possible contributor to AMD development. ^7,9,12–15 Erie et al. ¹⁵ reported that cadmium accumulates in human ocular tissue, particularly in the RPE and choroid. In addition, higher urinary cadmium levels, indicating a higher total cadmium body burden, were found to be associated with AMD in smokers. ¹⁴ In another study, higher cadmium levels were found in the neural retinal and RPE of eyes afflicted with AMD compared with controls in male subjects, although differences were not statistically significant in females. ¹³ Moreover, sex differences in cadmium levels were also observed; significantly higher levels were found in the RPE and choroid of retinas from female donors. ⁷ Despite evidence of an association between cadmium and AMD, most previous studies are case control studies using human donor eyes. Furthermore, no epidemiologic studies in the general population have been conducted. Therefore, in this study, we evaluated the relationship between cadmium blood levels and AMD in representative 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, population-based, cross-sectional study that was conducted by the Korean Ministry of Health and Welfare. Details regarding the study design and methods are provided elsewhere. ^16,17 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 noninstitutional civilian population of Korea. Data for the present study were obtained from the fourth (2008–2009), and fifth (2010–2012) KNHANES, to evaluate the association between blood cadmium levels and AMD. For the current study, 11,159 individuals who underwent evaluation of blood cadmium levels were selected. Of these, 5718 subjects under 40 years of age and 508 subjects who did not undergo fundus examination were excluded. Finally, 4933 participants aged 40 years or older were included in the analysis (Fig. 1). 

Figure 1

 

Flow diagram presenting the selection of study participants.

The study design followed the tenets of the Declaration of Helsinki for biomedical research. Protocols for this study was approved by the institutional review board of the Catholic University of Korea in Seoul, Korea. All participants signed and provided written informed consent. 

Assessment of AMD

Retinal examination and AMD grading was described in our previous report. ³ Briefly, 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 from each eye (a total of two images per person). Each image was graded twice, one preliminary and one detailed, using the international classification and grading system. ¹⁸ Two different graders analyzed fundus photograph images, and a trained senior grader, an ophthalmologist, was consulted in case of disagreements. These graders were blinded to the clinical status of the subjects. The quality of the survey was verified by the Epidemiologic Survey Committee of the Korean Ophthalmologic Society. Training of participating resident doctors was periodically performed by acting staff members of the National Epidemiologic Survey Committee of the Korean Ophthalmologic Society. 

Early AMD was defined by the presence of soft, indistinct, or reticular drusen; any type of drusen plus hyper- or hypopigmentary changes of the RPE in the macula; or by the presence of soft drusen over a 500-μm diameter area without late AMD signs in the macula. Late AMD was defined as the presence of wet or dry (geographic atrophy) AMD. Wet AMD was defined as detachment of RPE or neurosensory retina, presence of hemorrhages in subretinal or sub-RPE space, or 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. 

Assessment of Blood Cadmium Levels and Other Blood Markers

Details of the cadmium analysis have been reported previously. ^19,20 Cadmium was quantified by Zeeman effect graphite furnace atomic absorption spectrophotometry (SpectrAA-800; Varian, Belrose, Australia) with a detection limit of approximately 0.30 μg/L. Blood samples were collected by venipuncture after 10 to 12 hours of fasting. Next, fasting glucose, hemoglobin A1c, total cholesterol, and triglyceride levels were measured using a Hitachi automatic analyzer 7600 (Hitachi, Tokyo, Japan). All blood analyses were carried out by the Seoul Medical Science Institute (SMSI), a laboratory certified by the Korean Ministry of Health and Welfare. For internal quality assurance and control, commercial standard reference material was obtained from Bio-Rad (Lyphocheck Whole Blood Metals Control; Hercules, CA, USA), which showed that the coefficient of variation was 14.5% for three blood cadmium samples (reference values: 0.37, 1.11, and 4.30 μg/L). In terms of external quality control, the SMSI passed the German External Quality Assessment Scheme operated by Friedrick Alexander University, which is a standard protocol for measuring chemicals at low concentrations. The SMSI was also certified by the Ministry of Labor as one of the laboratories designated for analysis of specific chemicals. 

Assessment of Other Variables

Demographic and socioeconomic information was obtained from a health interview. Height was measured (in centimeters) using a wall-mounted measuring scale, and weight was measured (in kilograms), with the individual wearing light clothing without shoes using calibrated electronic scales. Body mass index (BMI) was calculated as follows: weight (kg)/height (m) ² . Age was classified in 10-year intervals. Smoking status was determined using self-reporting and subjects were classified as current smoker, past smoker, and never smoker. Alcohol use was evaluated and subjects were grouped as ever drinker or never drinker according to self-report. Menopause and estrogen replacement therapy were determined based on self-reporting. 

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

Statistical Analyses

Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS ver. 18.0; IBM Corp., Armonk, NY, USA) to account for the complex sampling design. Strata, sampling units, and sampling weights were used to obtain point estimates and linearized standard errors. Participant 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. The ANOVA or χ² test was used to compare demographic characteristics. 

To evaluate the effect of blood cadmium levels on prevalence of AMD, blood cadmium levels were categorized into quintiles. ²¹ Simple and multiple logistic regression analyses were used to examine the association between blood cadmium levels and AMD. After calculating crude odds ratio (OR; Model 1), we adjusted for age and sex (Model 2). We then adjusted for age, sex, and other confounders, including smoking, hypertension, heart problem, and stroke, which have been established as risk factors in previous studies (Model 3). ^4,22 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 less than 0.05 was considered statistically significance. 

Results

Of the 5441 eligible subjects over 40 years of age that had blood cadmium levels measured, 4933 subjects (90.6%) underwent a fundus examination. Reasons for not having undergone a fundus examination were as follows: small pupil (42.3%), cataract (26.7%), poor cooperation (7.1%), refusal (4.8%), cornea opacity (4.4%), and miscellaneous reasons (14.3%). Thus, in total, 4933 subjects were included in the analysis in the current study. The demographic characteristics of the 4933 subjects enrolled in the study are summarized by AMD status in Table 1. Subjects with AMD were more likely to be older (P < 0.001), have higher systolic blood pressure (P = 0.031), have lower hemoglobin A1c (P = 0.045), have higher blood cadmium levels (P = 0.008), hypertensive (P < 0.001), and an ever drinker (P < 0.001) compared with those without AMD. 

Table 1

 

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

Characteristics	No AMD, n = 4621	Early AMD, n = 295	Late AMD, n = 17	P	Total Subjects, n = 4933	Without Fundus Photographs, n = 508
Male, %	55.8 (0.6)	47.9 (3.6)	63.9 (13.5)	0.095	55.5 (0.5)	59.4 (3.3)
Age, y	51.9 (0.1)	61.2 (0.6)	62.2 (2.0)	<0.001*	52.4 (0.1)	59.4 (0.7)
Body mass index, kg/m2	24.1 (0.1)	23.9 (0.2)	24.9 (0.6)	0.258	24.1 (0.1)	24.1 (0.2)
Systolic blood pressure, mm Hg	121.7 (0.3)	125.4 (1.3)	131.6 (4.5)	<0.031*	121.9 (0.3)	125.4 (1.1)
Diastolic blood pressure, mm Hg	79.6 (0.2)	78.7 (0.7)	83.0 (2.7)	0.218	79.5 (0.2)	78.4 (0.7)
Fasting glucose, mg/dL	100.9 (0.5)	101.6 (1.2)	103.2 (2.8)	0.433	100.9 (0.4)	106.3 (1.9)
HbA1c, %	5.9 (0.0)	5.9 (0.1)	5.7 (0.1)	0.045*	5.9 (0.0)	6.3 (0.1)
Total cholesterol, mg/dL	193.7 (0.6)	193.1 (2.3)	184.4 (9.5)	0.330	193.6 (0.6)	190.9 (2.3)
Triglyceride, mg/dL	155.2 (2.7)	141.6 (6.7)	155.7 (24.5)	0.818	154.5 (2.6)	162.9 (7.8)
Cadmium, μg/L	1.31 (0.01)	1.47 (0.05)	1.64 (0.20)	0.008*	1.32 (0.01)	1.41 (0.0)
Diabetes, %	11.1 (0.6)	12.1 (2.3)	1.5 (1.6)	0.291	11.1 (0.6)	20.3 (2.8)
Hypertension, %	36.6 (0.9)	51.5 (3.4)	49.7 (14.5)	<0.001*	37.3 (0.9)	52.1 (3.2)
Heart problems, %	2.4 (0.3)	4.1 (1.3)	1.3 (1.3)	0.143	2.5 (0.3)	3.1 (0.8)
Stroke, %	1.9 (0.2)	2.4 (1.2)	2.8 (2.8)	0.762	1.9 (0.2)	3.3 (1.0)
Smoking status				0.132
Current, %	33.5 (0.8)	30.3 (3.1)	27.5 (13.0)		33.3 (0.8)	36.5 (3.7)
Former, %	18.3 (0.7)	12.6 (2.2)	26.2 (13.2)		18.0 (0.7)	17.4 (2.9)
Never, %	48.2 (0.7)	57.1 (3.5)	46.3 (14.3)		48.6 (0.7)	46.1 (3.6)
Ever drinker, %	88.6 (0.5)	78.2 (2.9)	69.3 (13.0)	<0.001*	88.0 (0.5)	81.6 (2.7)

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

*   P < 0.05.

Demographic and clinical characteristics by quintile blood cadmium categories are presented in Table 2. Participants with higher blood cadmium levels were more likely to be female (P for trend < 0.001), be older (P for trend < 0.001), have high systolic blood pressure (P for trend < 0.001), be hypertensive (P for trend = 0.004), have a history of stroke (P for trend = 0.040), be a smoker (P for trend < 0.001), and be ever drinker (P for trend = 0.005). 

Table 2

 

Demographic and Clinical Characteristics by Quintile Blood Cadmium Categories Among Representative Korean Adults Aged 40 Years or Older

Characteristics	Quintile Blood Cadmium Level, μg/L					P for Trend
	0.79<, n = 1011	0.79–1.07, n = 989	1.07–1.36, n = 964	1.36–1.80, n = 993	1.80>, n = 976
Male, %	73.2 (1.5)	61.4 (1.7)	52.9 (1.8)	43.8 (1.9)	44.0 (1.9)	<0.001*
Age, y	50.8 (0.2)	51.8 (0.3)	52.8 (0.3)	53.1 (0.3)	53.4 (0.3)	<0.001*
BMI, kg/m2	24.3 (0.1)	24.1 (0.1)	24.1 (0.1)	23.8 (0.1)	24.0 (0.1)	0.059
Systolic blood pressure, mm Hg	119.8 (0.6)	120.7 (0.5)	122.5 (0.6)	122.9 (0.7)	124.1 (0.7)	<0.001*
Diastolic blood pressure, mm Hg	79.0 (0.4)	79.3 (0.3)	79.7 (0.4)	79.4 (0.4)	80.2 (0.4)	0.430
Fasting glucose, mg/dL	100.7 (0.9)	101.3 (1.1)	100.8 (0.8)	99.5 (0.7)	102.3 (1.3)	0.385
HbA1c, %	5.97 (0.05)	5.97 (0.05)	5.95 (0.05)	5.85 (0.03)	6.03 (0.08)	0.111
Total cholesterol, mg/dL	192.0 (1.4)	194.0 (1.4)	191.0 (1.5)	196.2 (1.2)	194.8 (1.7)	0.077
Triglyceride, mg/dL	150.4 (7.0)	146.6 (3.4)	156.6 (7.0)	158.2 (5.1)	161.6 (6.4)	0.184
Diabetes, %	12.3 (1.2)	10.5 (1.2)	11.4 (1.2)	10.0 (1.2)	11.5 (1.3)	0.677
Hypertension, %	32.9 (1.8)	34.2 (1.8)	40.1 (1.8)	38.6 (1.9)	41.2 (2.0)	0.004*
Heart problems, %	2.3 (0.5)	2.9 (0.6)	2.2 (0.6)	2.6 (0.7)	2.2 (0.5)	0.892
Stroke, %	2.4 (0.6)	1.0 (0.3)	2.1 (0.5)	1.3 (0.4)	2.8 (0.7)	0.040*
Smoking status						<0.001*
Current, %	26.5 (1.6)	28.4 (1.7)	32.3 (1.8)	35.2 (1.9)	43.9 (1.9)
Former, %	29.6 (1.8)	22.7 (1.6)	18.5 (1.5)	11.5 (1.3)	6.0 (1.0)
Never, %	43.9 (1.9)	47.8 (1.9)	48.4 (1.8)	52.6 (2.0)	49.5 (2.0)
Ever drinker, %	91.3 (1.0)	88.8 (1.1)	87.2 (1.2)	87.2 (1.2)	85.4 (1.2)	0.005*

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

*   P < 0.05.

Blood cadmium levels of men and women by category, such as age, hypertension, smoking status, diabetes, and alcohol consumption, are presented in Table 3. Blood cadmium levels were 1.19 μg/L (95% confidence interval [CI], 1.16–1.22) in men and 1.47 μg/L (95% CI, 1.44–1.51) in women (P < 0.001). In both sexes, there were significantly higher mean blood cadmium levels in the oldest age group (P = 0.008), the hypertension group (P = 0.031), and smokers (P < 0.001) compared with the youngest age group, nonhypertension group, and nonsmokers. However, male participants with any type of AMD showed higher blood cadmium levels (P = 0.003), but this was not the case for females (P = 0.208). 

Table 3

 

Sex Difference of Blood Cadmium Levels (mg/L) of Men and Women According to Age Group and Other Variables Among Representative Korean Adults Aged 40 Years or Older

Characteristics	Blood Cadmium Levels, μg/L			P for Trend
	Men, n = 2446	P for Trend	Women, n = 2487
All subjects aged 40+ y	1.19 (1.16–1.22, 0.016)		1.47 (1.44–1.51, 0.018)
Age groups		0.008*		<0.001*
40–49 y	1.14 (1.09–1.19, 0.026)		1.39 (1.33–1.45, 0.029)
50–59 y	1.21 (1.17–1.26, 0.025)		1.55 (1.50–1.61, 0.029)
60–69 y	1.23 (1.18–1.29, 0.028)		1.50 (1.43–1.57, 0.034)
70+	1.33 (1.22–1.43, 0.054)		1.57 (1.44–1.70, 0.067)
AMD		0.003*		0.208
None	1.18 (1.15–1.21, 0.016)		1.47 (1.43–1.51, 0.018)
AMD	1.41 (1.26–1.55, 0.074)		1.55 (1.43–1.68, 0.065)
Diabetes		0.226		0.030*
Nondiabetes	1.18 (1.15–1.22, 0.018)		1.49 (1.45–1.52, 0.018)
Diabetes	1.23 (1.16–1.31, 0.038)		1.37 (1.27–1.47, 0.052)
Hypertension		0.031*		<0.001*
Nonhypertension	1.16 (1.12–1.21, 0.021)		1.43 (1.38–1.47, 0.021)
Hypertension	1.23 (1.19–1.28, 0.024)		1.58 (1.52–1.64, 0.030)
Smoking status		<0.001*		<0.001*
Never	0.88 (0.81–0.95, 0.036)		1.45 (1.41–1.49, 0.019)
Former	0.99 (0.95–1.03, 0.019)		1.51 (1.34–1.67, 0.084)
Current	1.39 (1.34–1.43, 0.023)		1.81 (1.64–1.99, 0.089)
Ever drinker		0.002*		0.572
Never	1.04 (0.94–1.13, 0.049)		1.49 (1.43–1.56, 0.034)
Ever	1.20 (1.17–1.23, 0.017)		1.47 (1.43–1.51, 0.021)

 Blood cadmium levels were expressed as weighted estimates (%; 95% CIs, standard errors [%]).

*   P < 0.05.

Prevalence of AMD according to blood cadmium quintiles is shown in Figure 2. As blood cadmium increases, the prevalence of any type of AMD significantly increased from 3.1% in the first quintile to 7.2% in the fifth quintile (P for trend = 0.003). In addition, the prevalence of early and late AMD significantly increased according to the increase in blood cadmium quintiles (P for trend = 0.011 and = 0.020, respectively). 

Figure 2

 

Prevalence of any, early, and late AMD according to blood cadmium quintile in a representative Korean population.

Table 4 shows ORs for the association between AMD and blood cadmium levels. The adjusted OR for any AMD after adjusting for age and sex (Model 2) was 1.94 (95% CI, 1.18–3.17; P for trend = 0.014) among those with the highest blood cadmium level quintile compared with participants in the lowest blood cadmium level quintile. In addition, the adjusted OR after adjusting for all potential covariates, such as age, sex, smoking, and hypertension, was 1.96 (95% CI, 1.17–3.29; P for trend = 0.017) in those in the highest quintile cadmium category than those in the lowest quintile category. This association between blood cadmium and AMD was supported statistically in early AMD (OR, 1.78; 95% CI, 1.05–3.02; P for trend = 0.046), and cadmium was associated with increased, but not statistically significant, odds of late AMD (OR, 6.42; 95% CI, 0.93–44.23; P for trend = 0.058). 

Table 4

 

Prevalence and Adjusted OR of Early and Late AMD, Stratified According to Quintile Categories of Blood Cadmium (mg/L) Among Representative Korean Adults Aged 40 Years or Older

	Blood Cadmium Quintiles, μg/L
	0.79<, n = 1011	0.79–1.07, n = 989	1.07–1.36, n = 964	1.36–1.80, n = 993	1.80>, n = 976	P for Trend
Any AMD
prevalence	3.1 (2.2–4.5, 0.6)	4.5 (3.2–6.1, 0.7)	4.9 (3.7–6.6, 0.7)	5.1 (3.8–6.7, 0.7)	7.2 (5.6–9.3, 0.9)	0.003*
OR
Model 1	1.00 (reference)	1.43 (0.87–2.36)	1.59 (0.99–2.57)	1.63 (1.30–2.59)*	2.38 (1.50–3.80)*	<0.001*
Model 2	1.00 (reference)	1.27 (0.75–2.15)	1.34 (0.81–2.21)	1.26 (0.78–2.05)	1.94 (1.18–3.17)*	0.014*
Model 3	1.00 (reference)	1.29 (0.76–2.19)	1.29 (0.77–2.14)	1.27 (0.77–2.10)	1.96 (1.17–3.29)*	0.017*
Early AMD
prevalence	3.0 (2.4–4.3, 0.5)	4.2 (3.0–5.8, 0.7)	4.8 (3.5–6.4, 0.7)	4.7 (3.5–6.3, 0.7)	6.5 (5.0–8.5, 0.9)	0.011*
OR
Model 1	1.00 (reference)	1.41 (0.84–2.35)	1.62 (1.00–2.63)	1.60 (1.01–2.54)*	2.26 (1.41–3.62)*	0.001*
Model 2	1.00 (reference)	1.24 (0.73–2.12)	1.35 (0.81–2.23)	1.22 (0.75–2.00)	1.80 (1.10–2.96)*	0.031*
Model 3	1.00 (reference)	1.25 (0.72–2.15)	1.28 (0.76–2.14)	1.22 (0.73–2.02)	1.78 (1.05–3.02)*	0.046*
Late AMD
prevalence	0.1 (0.0–0.3, 0.1)	0.1 (0.0–0.5, 0.1)	0.1 (0.0–0.5, 0.1)	0.3 (0.1–0.9, 0.2)	0.6 (0.2–1.4, 0.3)	0.020*
OR
Model 1	1.00 (reference)	1.81 (0.18–17.61)	1.05 (0.08–13.27)	2.24 (0.21–23.88)	4.51 (0.51–39.66)	0.130
Model 2	1.00 (reference)	1.75 (0.18–16.97)	1.01 (0.08–12.92)	2.13 (0.19–22.95)	4.42 (0.51–38.21)	0.135
Model 3	1.00 (reference)	1.99 (0.23–16.91)	1.16 (0.11–12.34)	2.62 (0.28–24.05)	6.42 (0.93–44.23)	0.058

 Prevalence was expressed as weighted estimates (%; 95% CIs, standard errors [%]). Model 1: crude OR. Model 2: adjusted for sex and age. Model 3: adjusted for sex, age, smoking, and hypertension.

*   P < 0.05.

The prevalence of AMD in males and females by blood cadmium quintile is shown in Figure 3. The prevalence of any AMD in men positively correlated with increased blood cadmium (P for trend = 0.005), whereas in women it did not (P for trend = 0.531). In men, the adjusted OR for any AMD type after adjusting for potential confounders, including age, smoking, and hypertension, was 2.11 (95% CI, 1.11–4.02; P for trend = 0.024) in those in the highest quintile cadmium category than in those in the lowest quintile category. However, we did not find any association between blood cadmium level and any AMD type in women after adjusting for covariates (OR, 1.29; 95% CI, 0.70–2.52; P for trend = 0.158; Table 5). Similarly, in the subtype analysis, subjects in the highest quintile cadmium level had 1.82 times more risk (95% CI, 1.04–3.50) for early AMD after adjusting for covariates than those in the lowest quintile in men, but not in women. Late AMD showed no association with blood cadmium level in either sex. 

Figure 3

 

Prevalence of AMD in males and females by blood cadmium quintile in a representative Korean population.

Table 5

 

Sex Difference of Prevalence and Adjusted OR of Any AMD, Stratified According to Quintile Categories of Blood Cadmium (mg/L) Among Representative Korean Adults Aged 40 Years or Older.

	Blood Cadmium Quintiles, μg/L
	Quintile 1	Quintile 2	Quintile 3	Quintile 4	Quintile 5	P for Trend
Men
Range	0.69<, n = 489	0.69–0.93, n = 493	0.93–1.19, n = 487	1.19–1.60, n = 487	1.60>, n = 490
Any AMD
prevalence	2.7 (1.6–4.3, 0.7)	2.9 (1.8.–4.8, 0.7)	5.7 (3.8–8.5, 1.2)	3.7 (2.3–5.9, 0.9)	6.7 (4.8–9.3, 1.1)	0.005*
OR
Model 1	1.00 (reference)	1.09 (0.53–2.24)	2.20 (1.13–4.27)*	1.40 (0.71–2.76)	2.62 (1.45–4.76)*	0.001*
Model 2	1.00 (reference)	1.00 (0.47–2.10)	1.93 (0.98–3.78)	1.22 (0.61–2.42)	2.15 (1.16–3.97)*	0.011*
Model 3	1.00 (reference)	1.04 (0.49–2.19)	1.91 (0.94–3.86)	1.22 (0.60–2.44)	2.11 (1.11–4.02)*	0.024*
Early AMD
prevalence	2.7 (1.6–4.3, 0.7)	2.6 (1.6–4.3, 0.7)	5.4 (3.5–8.2, 1.2)	3.1 (1.9–5.1, 0.8)	6.1 (4.3–8.5, 1.1)	0.008*
OR
Model 1	1.00 (reference)	0.98 (0.47–2.02)	2.07 (1.06–4.06)	1.17 (0.59–2.30)	2.35 (1.29–4.28)*	0.004*
Model 2	1.00 (reference)	0.89 (0.42–1.88)	1.82 (0.92–3.61)	1.03 (0.51–2.04)	1.93 (1.04–3.57)*	0.029*
Model 3	1.00 (reference)	0.92 (0.44–1.95)	1.79 (0.87–3.65)	1.00 (0.49–2.01)	1.82 (1.04–3.50)*	0.077
Late AMD
prevalence	0.0 (0.0–0.0, 0.0)	0.3 (0.0–2.1, 0.3)	0.3 (0.1–1.8, 0.3)	0.6 (0.1–2.5, 0.4)	0.7 (0.2–2.1, 0.4)	0.407
OR
Model 1	Non applicable	1.00 (reference)	1.04 (0.76–14.27)	1.99 (0.17–22.89)	2.30 (0.23–22.30)	0.072
Model 2	Non applicable	1.00 (reference)	0.96 (0.07–13.29)	1.89 (0.16–22.27)	1.97 (0.19–19.68)	0.115
Model 3	Non applicable	1.00 (reference)	0.92 (0.10–11.57)	1.91 (0.22–18.87)	2.27 (0.41–15.53)	0.025*
Women
Range	0.93<, n = 500	0.93–1.21, n = 497	1.21–1.50, n = 486	1.50–1.94, n = 511	1.94>, n = 493
Any AMD
prevalence	4.7 (3.0–7.5, 1.1)	5.1 (3.5–7.5, 1.0)	4.9 (3.1–7.5, 1.1)	7.2 (5.0–10.3, 1.3)	6.3 (4.1–9.4, 1.3)	0.531
OR
Model 1	1.00 (reference)	1.08 (0.57–2.07)	1.03 (0.51–2.06)	1.55 (0.83–2.88)	1.34 (0.70–2.56)	0.171
Model 2	1.00 (reference)	1.03 (0.50–2.10)	0.88 (0.42–1.85)	1.37 (0.69–2.70)	1.18 (0.57–2.46)	0.401
Model 3	1.00 (reference)	0.91 (0.47–1.79)	0.90 (0.47–1.77)	1.31 (0.71–2.57)	1.29 (0.70–2.52)	0.158
Early AMD
prevalence	4.7 (3.0–7.5, 1.1)	4.9 (3.3–7.1, 1.0)	4.8 (3.1–7.5, 1.1)	6.7 (4.6–9.8, 1.3)	5.8 (3.8–8.8, 1.2)	0.701
OR
Model 1	1.00 (reference)	1.02 (0.53–1.96)	1.01 (0.50–2.04)	1.44 (0.77–2.71)	1.23 (0.64–2.37)	0.265
Model 2	1.00 (reference)	0.97 (0.47–1.99)	1.86 (0.41–1.83)	1.27 (0.64–2.54)	1.08 (0.52–2.25)	0.550
Model 3	1.00 (reference)	0.85 (0.43–1.71)	0.87 (0.44–1.75)	1.29 (0.60–2.46)	1.15 (0.52–2.25)	0.328
Late AMD
prevalence	0.0 (0.0–0.0, 0.0)	0.3 (0.1–1.2, 0.2)	0.1 (0.0–0.6, 0.1)	0.5 (0.2–1.6, 0.3)	0.5 (0.1–3.4, 0.5)	0.434
OR
Model 1	Non applicable	1.00 (reference)	0.27 (0.02–3.04)	1.72 (0.27–10.97)	1.64 (0.14–18.61)	0.137
Model 2	Non applicable	1.00 (reference)	0.24 (0.02–2.79)	1.57 (0.24–10.07)	1.51 (0.13–17.57)	0.165
Model 3	Non applicable	1.00 (reference)	0.26 (0.12–2.82)	1.71 (0.24–10.35)	1.74 (0.21–21.44)	0.203

 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, and hypertension.

*   P < 0.05.

The effect of menopause and hormone replacement therapy on blood cadmium level is shown in Table 6. There is no significant difference in blood cadmium levels between premenopausal women (1.44 μg/L) and postmenopausal women (1.49 μg/L, P = 0.177), while blood cadmium levels in postmenopausal women taking hormone replacement therapy (1.37 μg/L) was significantly higher than those not taking hormone replacement therapy (1.52 μg/L, P < 0.001). Compared with males (1.19 μg/L), premenopausal women (1.44 μg/L) had significantly higher blood cadmium levels (P < 0.001). 

Table 6

 

Effect of Menopause and Hormone Replacement Therapy on the Blood Cadmium Levels Among Representative Korean Women Aged 40 Years or Older

Characteristics	Blood Cadmium Levels, μg/L	P Value
Total, n = 4933	1.32 (1.29–1.34, 0.012)
Male, n = 2446	1.19 (1.16–1.22, 0.016)	<0.001*
Female, n = 2487	1.47 (1.44–1.51, 0.018)
Premenopause, n = 807	1.44 (1.38–1.50, 0.030)	0.177
Postmenopause, n = 1579	1.49 (1.45–1.54, 0.021)
With HRT, n = 338	1.37 (1.31–1.43, 0.031)	<0.001*
Without HRT, n = 1237	1.52 (1.47–1.57, 0.026)

*   P < 0.05.

Discussion

The present study is the first epidemiologic study to evaluate the association between blood cadmium levels and AMD. Our study demonstrated that the risk of any AMD type and early AMD significantly increased in those with high blood cadmium levels, compared with those in the lowest blood cadmium quintile. However, this association was only evident in men, and not in women. Finally, blood cadmium levels were higher in women (1.47 μg/L) than in men (1.19 μg/L, P < 0.001). 

The risk of AMD was 2.38 times higher in subjects with high blood cadmium relative to those in the lowest blood cadmium quintile. This risk was 1.96 times higher even after adjusting for potential confounders, such as sex, age, smoking status, and hypertension. This result is supported by previous findings. For instance, a recent case control study using donor eyes demonstrated that cadmium levels in the retina and RPE is higher in those with AMD (11 eyes) than in age-matched controls (28 eyes). ¹³ In another study examining aqueous humor trace elements, 12 patients with dry AMD had significantly higher cadmium levels (median, 0.70 μM) compared with 11 patients without AMD (median, 0.06 μM). ¹² However, these studies were case control studies and were limited by the small number of subjects. Our findings confirmed cadmium as an etiological factor for AMD using a population-based epidemiologic approach with a relatively large number of participants (n = 4933). In the AMD subtype analysis, late AMD was not significantly associated with blood cadmium level, although early AMD was associated. The OR for late AMD is higher than for early AMD, and a wide CI included a null point. This seems to be due to an insufficient number of subjects with late AMD (17 subjects). Further studies including sufficient number of subjects with late AMD are needed to elucidate the association between cadmium and late AMD. 

Sex differences in the association between blood cadmium and AMD were identified in the present study. Men with high blood cadmium levels have a 2.11 times higher risk for AMD compared with men in the lowest blood cadmium quintile level even after adjusting for potential confounders. However, in women, there was no statistical significant association between blood cadmium and AMD after adjusting for age. A prior study reported a similar association; Wills et al. ¹³ reported that higher cadmium levels were only found in the retina and RPE of donor eyes with AMD from males and not females. It is unclear why men would be more susceptible to cadmium-related AMD. One possible reason for this is the high current smoking rate in men (55.1%) compared with women (6.1%) in the Korean population used in the present study. Smoking plants is a resource for cadmium intake, as well as being a consistent risk factor for AMD. ⁶ The prior study by Wills et al. ¹³ included no information on donor smoking history; therefore, the authors could not determine the exact cause for sex difference in cadmium level. However, in the present study, this sex difference persisted even after adjusting for smoking status. Thus, the results of the present study are suggestive of possible sex differences related to the metabolism of cadmium metals in the human retina. 

Blood cadmium level was significantly higher in women (1.47 μg/L) than in men (1.19 μg/L, P < 0.001). This finding is consistent with previous studies. For instance, Satarug et al. ²³ reported that women who were nonsmokers had nearly the same cadmium body burden as men who smoked an average of nine cigarettes per day for 10 years. Another study using donor eyes showed that cadmium levels in the retina and RPE were significantly higher in females (n = 17) than males (n = 15). ⁷ Although the exact reason for this discrepancy is unknown, higher blood cadmium levels in women can be explained by higher absorption of cadmium due to low iron status. ²⁴ Women have low levels of iron and are susceptible to chronic iron deficiency and iron deficiency may enhance cadmium absorption by up to 10-fold. ²⁵ Unfortunately, we had no information regarding iron status in this study. To identify another possible cause for this sex difference, we examined the effect of menopause and estrogen on cadmium levels; there is no significant difference in blood cadmium levels between pre- (1.44 μg/L) and postmenopausal women (1.49 μg/L). In addition, premenopausal women (1.44 μg/L) have significantly higher blood cadmium levels than men (1.19 μg/L). This finding suggests that sex difference in cadmium levels in Korean populations are not caused by menopause or estrogen hormone levels. However this stratified analysis by menopausal status could be confounded by varying iron supplementation practices. 

The finding that blood cadmium levels are higher in women, and that blood cadmium levels were associated with AMD only in males and not in females, are intriguing and may reflect a sex difference in the uptake and metabolism of this metal ion. Further studies, are needed to identify factors responsible for this difference, and to elucidate the exact biologic mechanisms of cadmium absorption and regulation by sex. 

The major strength of the present study is the relatively large number of participants (n = 4933) and the study design, which implemented systemic stratified, multistage, clustered, random sampling methods. Another strength is the rigorous quality control for ophthalmic fundus examination and blood cadmium measurement in KNHANES. However, our study has several limitations. First, we did not examine occupational exposure, such as smelting, electroplating, pigment manufacture and application, and alkaline battery manufacturing, as KNHANES does not include information about occupation or job. Second, we used blood cadmium levels instead of urinary levels. Urinary cadmium level is known as a better cumulative biomarker of lifetime cadmium exposure, while blood cadmium may reflect only recent exposure. ²⁶ However, blood cadmium is also a good proxy for cadmium body burden in populations with low level environmental exposure. ²⁷ Third, as subjects not included in the analysis were more likely to be older, have higher cadmium levels, be hypertensive, and be a smoker (Table 1), exclusion of nonparticipants might have biased the results. Finally, our study has a cross-sectional design, which makes inferring causality difficult. For example, if a patient receives a diagnosis of AMD, this may alter their future smoking behavior, which affect the blood cadmium levels. However, in this case, we have documented associations between blood cadmium levels and AMD on the basis of existing evidence regarding their effects on AMD development. This differential blood cadmium pattern is unlikely to be due to AMD. 

In conclusion, the present study provides the first population-based epidemiologic evidence of an association between blood cadmium and AMD in a representative Korean population. Blood cadmium level was associated with AMD in men but not in women after adjusting for potential confounders, including age, smoking status, hypertension, heart problems, and stroke. However, overall blood cadmium levels were higher in women than in men. We hypothesize that accumulation of cadmium is detrimental to retinal function and likely contributes to pathophysiological conditions, such as oxidative stress, that underlie AMD development and progression. Consequently, low-dose lifetime environmental exposure to cadmium from dietary resources, polluted air, and tobacco smoking may lead to retinal health problem. Our results have important health policy and disease prevention implications. 

Acknowledgments

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

Supported by funding from St. Vincent's Hospital, Research Institute of Medical Science Foundation 2013 (5-2014-B0001-00101). 

Disclosure: E.C. Kim, None; E. Cho, None; D. Jee, None