Age‐Related Macular Degeneration With Visual Disability Is Associated With Cardiovascular Disease Risk in the Korean Nationwide Cohort

Abstract

Background

Age‐related macular degeneration (AMD) is the leading cause of visual disability. AMD shares some risk factors with the pathogenesis of cardiovascular disease (CVD). However, previous studies examining the association between AMD and the risk of CVD provide conflicting results. Hence, we investigated the association between AMD, visual disability, and the risk of CVD.

Methods and Results

This is a nationwide cohort study using data from the Korean National Health Insurance System database (2009–2019) on subjects who underwent a national health screening program in 2009. A total of 3 789 963 subjects were categorized by the presence of AMD and visual disability. Visual disability was defined as a best‐corrected visual acuity of ≤20/100 by validated documentation from a specialist physician. Cox regression hazard model was used to examine the hazard ratios (HRs) of CVD, including myocardial infarction and ischemic stroke, after adjusting for potential confounders. During a mean 9.77 years of follow‐up, AMD was associated with a 5% higher risk of myocardial infarction (adjusted HR [aHR], 1.05 [95% CI, 1.01–1.10]) but not associated with increased risk of overall CVD (aHR, 1.02 [95% CI, 1.00–1.05]) or ischemic stroke (aHR, 1.02 [95% CI, 0.98–1.06]). However, when AMD was accompanied by visual disability, there was increased risk of CVD (aHR, 1.17 [95% CI, 1.06–1.29]), myocardial infarction (aHR, 1.18 [95% CI, 1.01–1.37]), and ischemic stroke (aHR, 1.20 [95% CI, 1.06–1.35]). These trends were more evident in women and subjects with cardiometabolic comorbidities.

Conclusions

AMD with visual disability, but not all AMD, was associated with an increased risk of CVD. Patients with AMD who have visual disability should be targeted for CVD prevention.

Nonstandard Abbreviations and Acronyms

AMD

age‐related macular degeneration

NHIS

National Health Insurance System

Clinical Perspective.

What Is New?

• In this nationwide cohort study of 3 789 963 subjects with information on lifestyle behaviors, age‐related macular degeneration with visual disability, but not all age‐related macular degeneration, was associated with an increased risk of cardiovascular disease.

What Are the Clinical Implications?

• Patients with age‐related macular degeneration who have visual disability should be targeted for risk factor management and disease prevention for cardiovascular disease.

Age‐related macular degeneration (AMD) is a progressive and degenerative disease of the area of the retina that is responsible for central vision. ¹ It is the leading cause of visual disability in developed countries, particularly in individuals aged >60 years. ² The number of people with AMD worldwide is expected to increase to 288 million by 2040. ² Therefore, it is of clinical significance to address comorbid conditions and health problems in people with AMD.

Cardiovascular disease (CVD) continues to be the leading cause of death in the United States, with 199 of 100 000 people dying of heart disease and stroke in 2019. ³ Because traditional risk factors for CVD, such as aging, cigarette smoking, obesity, elevated cholesterol levels, and hypertension, have been regarded to be risk factors for AMD, ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ AMD and CVD are now considered to share some risk factors and pathogenesis.

However, previous studies examining the association between AMD and the risk of CVD, including myocardial infarction (MI) and ischemic stroke, provide conflicting results (Tables S1 and S2). ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ Previous studies using the US Centers for Medicare & Medicaid Services database reported that AMD was associated with a 19% higher risk of incident MI ¹⁷ and a 20% higher risk of ischemic stroke. ¹⁴ Findings from the ARIC (Atherosclerosis Risk in Communities) study showed that AMD was associated with a 42% higher risk of cerebral infarction in the US population. ¹⁶ On the other hand, a study in the United States showed a 42% decrease in the risk of hospitalized MI (adjusted hazard ratio [aHR], 0.58 [95% CI, 0.48–0.72]) or a 44% decrease in the risk of cerebrovascular accident (aHR, 0.56 [95% CI, 0.45–0.70]). ¹⁹

Several studies have investigated the association of AMD with CVD by stage and have shown no association between late AMD and the risk of CVD, ¹⁵ , ²² but a Taiwanese study using nationwide claims data showed that patients with neovascular AMD were nearly twice as likely to be diagnosed with stroke (aHR, 2.01 [95% CI, 1.34–3.02]). ²⁰ These studies, however, were limited by suboptimal adjustment of potential confounders and a short period of follow‐up time. In addition, although visual disability is associated with cardiometabolic conditions, ²⁶ , ²⁷ to our knowledge, the presence of visual disability was not considered in previous studies on the risk of CVD in patients with AMD. Hence, we performed a retrospective cohort study to investigate the association between AMD, which was further categorized by visual disability, and the risk of CVD using data from the Korean National Health Insurance System (NHIS) database.

Methods

Data Availability

The data sets used and/or analyzed during this study are available from the corresponding author on reasonable request.

Data Source and Study Setting

The NHIS provides mandatory universal coverage to 97% of the Korean population, and an additional medical aid program provides coverage to 3% of the population in the lowest income bracket, which is funded by general taxation. The NHIS also operates national health check‐up programs, which include a cardiovascular health screening test for every individual aged ≥40 years and all employees regardless of age. Therefore, the NHIS database comprises an eligibility database (age, sex, disability type and severity, socioeconomic variables, income level, and type of eligibility), a medical treatment database (based on medical bills claimed by medical service providers for their medical expense claims), and a health examination database. This nationwide database is widely used in epidemiologic studies in Korea. ²⁸

Study Population

Among 4 470 729 participants (aged ≥50 years) who underwent general health screenings in 2009, we first excluded individuals with missing data on at least one variable (n=250 392). Participants who had a history or previous diagnosis of MI (n=121 690) and ischemic stroke (n=308 684) were also excluded. Finally, 3 789 963 participants were included in the analyses.

Definition of AMD

Subjects with AMD were identified using the International Classification of Diseases, Tenth Revision (ICD‐10), code for AMD (H353) by an ophthalmologist within a year before the health screening examination. The definition of AMD involving a claim with the ICD‐10 code for AMD (H353) has also been used in a previous epidemiological study on AMD. ²⁹

Definition of Visual Disability

We used information from the National Disability Registration System to define visual disability for our study. In Korea, the government provides welfare benefits (eg, subsidy and parking permit for people with disability), based on the type and severity of disability. ³⁰ And most people who can meet these criteria register to this system for this benefit. Registration to this system requires a medical certificate by a specialist physician in Korea, ³¹ and therefore, the information is sufficiently reliable. Visual disability refers to a visual loss or visual field defect, and for registration of visual disability, medical examination and certificate by an ophthalmologist are required. Disability status is defined by the severity index, from 1 (most severe) to 6 (least severe). For instance, a best‐corrected visual acuity of ≤20/100 using the Snellen visual acuity chart is a minimal requirement to register for visual disability (Table S3).

Study Outcomes and Follow‐Up

The end points of the study were incident MI and ischemic stroke. Newly diagnosed MI and ischemic stroke were identified using the ICD‐10 code for MI (ICD‐10 code I21 or I22 during hospitalization, or these codes having been recorded at least 2 times in outpatients) and ischemic stroke (ICD‐10 code I63 or I64 during hospitalization, with claims for brain magnetic resonance imaging or brain computed tomography), respectively. ³² We also defined CVD as a composite of MI and ischemic stroke. The subjects were followed up from the date of the health screening examination in 2009 to the date of incident MI, ischemic stroke, death, censor date (eg, outmigration to another country), or until the end of the study period (December 31, 2019), whichever came first.

Covariates

The subjects provided information on lifestyle behaviors using standardized questionnaires. Smoking status was categorized as none, ex‐smoker, and current smoker. Alcohol consumption was categorized as none, mild, and heavy. Heavy drinking was defined as ≥30 g of alcohol consumption per day. Individuals were said to have regular physical activity if they exercised strenuously ≥1 time/week for at least 20 minutes per session.

Household income was dichotomized by the lowest 20th percentile, according to the health insurance premium. Body mass index was calculated as weight in kilograms divided by height in meters squared. The presence of arterial hypertension, diabetes, and dyslipidemia was assessed using comprehensive information on medical history and clinical and pharmacy ICD‐10 codes. Arterial hypertension was defined as any of the following: systolic blood pressure ≥140 mm Hg; diastolic blood pressure ≥90 mm Hg; or treatment with an antihypertensive medication that was linked to the arterial hypertension ICD‐10 codes (I10–I13 and I15) and resulted in at least one claim in a year. Diabetes was defined as a blood glucose level ≥126 mg/dL or a history of a hypoglycemic medication prescription that was linked to ICD‐10 codes of diabetes (E11–E14) and resulted in at least one claim in a year. Dyslipidemia was defined as total cholesterol ≥240 mg/dL or a history of a lipid‐lowering medication that was associated with an ICD‐10 code (E78).

Statistical Analysis

The comparison of baseline characteristics according to the presence of AMD and visual disability was conducted using a t‐test for continuous variables or the χ² test for categorical variables. The incidence rates of CVD were assessed as the incident cases divided by 100 000 person‐years. The Cox regression hazard model was used to examine the hazard ratios (HRs) of CVD. Multivariable analyses were adjusted for age and sex in model 2 and for age, sex, household income, area of residence, body mass index, smoking, alcohol consumption, and regular exercise in model 3. In our final model 4, hypertension, diabetes, dyslipidemia, and Charlson comorbidity index were included in addition to every adjusting variable used in model 3. Covariates were selected on the basis of previous literature, ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ and absence of multicollinearity was tested by variance inflation factor. P for trend was calculated among the HR of control, AMD without visual disability, and AMD with visual disability. Finally, to evaluate the potential effect modification by age, sex, and comorbidity status, P for interaction was calculated using stratified analysis.

The statistical analyses were performed using SAS version 9.4 (SAS Institute Inc, Cary, NC). P<0.05 was considered statistically significant.

Ethics Statement

This study was approved by the Institutional Review Board of Samsung Medical Center (2022‐03‐060). Anonymized and deidentified information was used for analyses; therefore, informed consent was not required. The database is open to all researchers whose study protocols are approved by the official review committee.

Results

Baseline Characteristics

Of the total 3 789 963 individuals enrolled in the final analysis, 43 753 subjects (1.13%) had AMD at the baseline, of whom 3123 subjects (7.1% of AMD group) had visual disability. The descriptive characteristics of the study population, stratified by the presence of AMD, and further by visual disability, are summarized in Table 1.

Table 1.

Baseline Characteristics of the Study Population, According to the Presence of AMD

Characteristics	AMD			AMD
	Absence (N=3 746 210)	Presence (N=43 753)	P value	Without visual disability (N=40 630)	With visual disability (N=3123)	P value
Age, y	60.1±8.1	66.7±8.5	<0.001	66.7±8.5	67.2±8.4	<0.001
Male sex, n (%)	1 815 661 (48.5)	18 543 (42.4)	<0.001	17 058 (42.0)	1485 (47.6)	<0.001
Smoking, n (%)			<0.001			<0.001
Nonsmoker	2 485 174 (66.3)	32 016 (73.2)		29 800 (73.3)	2216 (71.0)
Ex‐smoker	587 471 (15.7)	6866 (15.7)		6366 (15.7)	500 (16.1)
Current smoker	673 565 (18.0)	4871 (11.1)		4464 (11.0)	407 (13.0)
Alcohol consumption, n (%)			<0.001			<0.001
None	2 399 823 (64.1)	32 611 (74.5)		30 261 (74.5)	2350 (75.3)
Mild	1 098 959 (29.3)	9364 (21.4)		8721 (21.5)	643 (20.6)
Heavy	247 428 (6.6)	1778 (4.1)		1648 (4.1)	130 (4.2)
Regular physical activity, n (%)	798 398 (21.3)	9260 (21.2)	0.453	8587 (21.1)	673 (21.6)	0.650
Anthropometrics
Body mass index, kg/m2	24.1±3.0	24.0±3.0	<0.001	24.0±3.0	23.9±3.1	<0.001
WC, cm	82.0±8.3	82.6±8.3	<0.001	82.6±8.3	82.9±8.4	<0.001
SBP, mm Hg	126.3±15.8	128.0±15.9	<0.001	128.0±15.8	127.7±16.3	<0.001
DBP, mm Hg	77.9±10.2	77.5±10.0	<0.001	77.5±10.0	77.4±10.3	<0.001
Comorbidity, n (%)
Hypertension	1 627 704 (43.5)	24 733 (56.5)	<0.001	23 016 (56.7)	1717 (55.0)	<0.001
Diabetes	532 679 (14.2)	9467 (21.6)	<0.001	8741 (21.5)	726 (23.3)	<0.001
Dyslipidemia	998 592 (26.7)	14 204 (32.5)	<0.001	13 219 (32.5)	985 (31.5)	<0.001
Laboratory findings
Glucose, mg/dL	101.7±26.9	103.9±29.2	<0.001	103.8±29.1	104.8±31.4	<0.001
eGFR, mL/min per 1.73 m2	83.2±33.3	80.4±33.4	<0.001	80.4±33.1	80.5±37.1	<0.001
Cholesterol, mg/dL	201.8±38.0	200.2±38.9	<0.001	200.3±38.8	199.1±39.5	<0.001
HDL‐C, mg/dL	55.7±31.0	55.5±34.5	0.054	55.6±34.4	54.5±34.7	0.058
LDL‐C, mg/dL	119.7±39.2	119.2±39.6	0.006	119.3±39.6	118.3±39.1	0.017
Triglycerides, mg/dL*	120.8 (120.7–120.9)	120.5 (119.9–121.1)	0.775	120.3 (119.7–120.9)	123.5 (121.3–125.8)	0.017
Urban residency, n (%)	1 714 128 (45.8)	18 290 (41.8)	<0.001	16 989 (41.8)	1301 (41.7)	<0.001
Income, n (%)
Lowest 20%	815 269 (21.8)	8032 (18.4)	<0.001	7410 (18.2)	621 (19.9)	<0.001
Charlson comorbidity index	1.4±1.3	2.1±1.5	<0.001	2.1±1.5	2.1±1.5	<0.001

Data are expressed as mean±SD or number (percentage), except for triglycerides, which are presented as geometric mean with a 95% CI using the Wilcoxon rank‐sum test. AMD indicates age‐related macular degeneration; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; SBP, systolic blood pressure; and WC, waist circumference.

^*Geometric mean (95% CI).

Among subjects with AMD (AMD group), the average age at baseline was 66.7±8.5 years, 42.4% were men, and 11.1% were current smokers. At baseline, the AMD group was older, with a higher proportion of women and nonsmokers, and a higher prevalence of hypertension, diabetes, and dyslipidemia than the non‐AMD group, or subjects without AMD (all P<0.001). For subjects who had AMD with visual disability at baseline (AMD with visual disability group), the average age at baseline was 67.2±8.4 years, with 47.6% men, and 13.0% current smokers. At baseline, the AMD with visual disability group was older, with a higher proportion of men and current smokers, and had a lower prevalence of hypertension and dyslipidemia than the AMD without visual disability group (all P<0.001). Detailed information is shown in Table 1.

AMD and the Risk of CVD, According to Visual Disability Status

During a mean 9.77 years of follow‐up, there were 254 670 overall CVD events (6.72%): 112 174 cases of MI and 156 675 cases of ischemic stroke. The association between AMD and the risk of CVD (MI and ischemic stroke) is shown in Table 2.

Table 2.

Association Between AMD and the Risk of Cardiovascular Outcome, According to Visual Disability

Variable	Subjects, N	Cases, n	IR per 1000 person‐years	Model 1 (crude)	Model 2	Model 3	Model 4
				HR (95% CI)	aHR (95% CI)	aHR (95% CI)	aHR (95% CI)
Cardiovascular disease, composite of myocardial infraction and ischemic stroke
Non‐AMD	3 746 210	249 914	6.8	1 (Reference)	1 (Reference)	1 (Reference)	1 (Reference)
AMD	43 753	4756	11.6	1.72 (1.67–1.77)*	1.09 (1.06–1.12)*	1.11 (1.07–1.14)*	1.02 (1.00–1.05)
P value				<0.001	<0.001	<0.001	0.101
AMD without visual disability	40 630	4365	11.5	1.69 (1.64–1.74)*	1.08 (1.05–1.11)*	1.09 (1.06–1.13)*	1.01 (0.98–1.04)
AMD with visual disability	3123	391	13.9	2.06 (1.86–2.27)*	1.24 (1.13–1.37)*	1.25 (1.14–1.39)*	1.17 (1.06–1.29)*
P for trend†				<0.001	<0.001	<0.001	0.025
Myocardial infarction
Non‐AMD	3 746 210	110 119	3.0	1 (Reference)	1 (Reference)	1 (Reference)	1 (Reference)
AMD	43 753	2055	4.9	1.67 (1.60–1.74)*	1.14 (1.09–1.19)*	1.15 (1.10–1.20)*	1.05 (1.01–1.10)*
P value				<0.001	<0.001	<0.001	0.020
AMD without visual disability	40 630	1888	4.8	1.65 (1.57–1.72)*	1.13 (1.08–1.18)*	1.14 (1.09–1.19)*	1.04 (1.00–1.09)*
AMD with visual disability	3123	167	5.8	1.97 (1.69–2.29)*	1.28 (1.10–1.49)*	1.29 (1.11–1.50)*	1.18 (1.01–1.37)*
P for trend†				<0.001	<0.001	<0.001	0.009
Ischemic stroke
Non‐AMD	3 746 210	153 614	4.2	1 (Reference)	1 (Reference)	1 (Reference)	1 (Reference)
AMD	43 753	3061	7.4	1.78 (1.72–1.85)*	1.07 (1.04–1.11)*	1.09 (1.05–1.13)*	1.02 (0.98–1.06)
P value				<0.001	<0.001	<0.001	0.332
AMD without visual disability	40 630	2800	7.3	1.75 (1.69–1.82)*	1.06 (1.02–1.10)*	1.08 (1.04–1.12)*	1.00 (0.97–1.04)
AMD with visual disability	3123	261	9.1	2.21 (1.96–2.50)*	1.26 (1.12–1.42)*	1.28 (1.13–1.45)*	1.20 (1.06–1.35)*
P for trend†				<0.001	<0.001	<0.001	0.112

Model 2, adjusted for age and sex; model 3, model 2+adjusted for income, area of residence, body mass index, smoking, alcohol consumption, and regular physical activity; and model 4, model 3+adjusted for hypertension, diabetes, dyslipidemia, and Charlson comorbidity index. aHR indicates adjusted HR; AMD, age‐related macular degeneration; HR, hazard ratio; and IR, incidence rate.

^*Statistically significant results.

^† P for trend was calculated by the (adjusted) HRs among control, AMD without visual disability, and AMD with visual disability groups.

AMD was not associated with the risk of overall CVD (aHR, 1.02 [95% CI, 1.00–1.05]; P =0.101). Further analysis that specified AMD by visual disability status demonstrated no increased risk in AMD without visual disability (aHR, 1.01 [95% CI, 0.98–1.04]) and a 17% increased risk for CVD in AMD with visual disability (aHR, 1.17 [95% CI, 1.06–1.29]). In addition, there was a significant trend in the risk of CVD among the non‐AMD, AMD without visual disability, and AMD with visual disability groups ( P for trend=0.025).

AMD was associated with a 5% increased risk of MI (aHR, 1.05 [95% CI, 1.01–1.10]). Further analysis that evaluated AMD by visual disability demonstrated that there were 4% and 18% increased risks for MI in AMD without and with visual disability, respectively (aHR, 1.04 [95% CI, 1.00–1.09]; and aHR, 1.18 [95% CI, 1.01–1.37], respectively). In addition, a significant trend in the risk of MI was found among the control, AMD without visual disability, and AMD with visual disability groups (P for trend=0.009).

AMD was not associated with the risk of ischemic stroke (aHR, 1.02 [95% CI, 0.98–1.06]). However, the association was significant between AMD with visual disability and the incidence of ischemic stroke (aHR, 1.20 [95% CI, 1.06–1.35]). Contrary to the findings in MI, no significant trend of ischemic stroke risk was found among the control, AMD without visual disability, and AMD with visual disability groups ( P for trend=0.112).

Stratified Analysis

We performed stratified analysis according to age, sex, and cardiometabolic comorbidities (hypertension, diabetes, or dyslipidemia) (Table 3). There was no significant interaction according to age groups. The relationship between AMD and CVD was more evident in women, with a significant interaction between AMD and sex (P‐interaction=0.006). Trends in the risk of CVD were higher with AMD and visual disability in subjects who had cardiometabolic comorbidities than in those who did not: aHR (95% CI) values for CVD, MI, and ischemic stroke were 1.20 (1.07–1.34) versus 1.26 (1.07–1.48); 1.21 (1.05–1.38) versus 1.04 (0.83–1.31); and 0.88 (0.60–1.28) versus 1.16 (0.88–1.54), respectively, and P‐interaction was significant for MI (P‐interaction=0.014).

Table 3.

Risk of Cardiovascular Outcome by AMD and Visual Disability, According to Age, Sex, and Comorbidities

Variable		Subjects, N	Cardiovascular disease		Myocardial infarction		Ischemic stroke
			Cases, n	aHR (95% CI)*	Cases, n	aHR (95% CI)*	Cases, n	aHR (95% CI)*
Age group, y
50–64	Non‐AMD	2 699 819	118 815	1 (Reference)	57 207	1 (Reference)	66 324	1 (Reference)
	AMD without visual disability	15 931	867	1.01 (0.95–1.08)	422	1.03 (0.93–1.13)	499	1.03 (0.94–1.12)
	AMD with visual disability	1146	77	1.22 (0.98–1.53)	42	1.40 (1.03–1.89)†	44	1.24 (0.92–1.66)
65–74	Non‐AMD	844 539	95 275	1 (Reference)	38 668	1 (Reference)	62 845	1 (Reference)
	AMD without visual disability	17 705	2150	1.00 (0.96–1.04)	941	1.07 (1.00–1.14)†	1379	0.97 (0.92–1.02)
	AMD with visual disability	1383	209	1.26 (1.10–1.45)†	77	1.10 (0.88–1.38)	147	1.35 (1.15–1.58)†
≥75	Non‐AMD	201 852	35 824	1 (Reference)	14 244	1 (Reference)	24 445	1 (Reference)
	AMD without visual disability	6994	1348	1.02 (0.97–1.08)	525	0.98 (0.90–1.07)	922	1.03 (0.97–1.10)
	AMD with visual disability	594	105	0.96 (0.79–1.17)	48	1.10 (0.83–1.46)	70	0.94 (0.75–1.19)
P interaction			0.215		0.363		0.072
Sex
Men	Non‐AMD	1 815 661	138 864	1 (Reference)	61 421	1 (Reference)	85 132	1 (Reference)
	AMD without visual disability	17 058	1950	0.96 (0.92–1.00)	876	1.02 (0.95–1.09)	1222	0.93 (0.88–0.99)
	AMD with visual disability	1485	200	1.11 (0.97–1.27)	84	1.10 (0.88–1.36)	137	1.17 (0.99–1.38)
Women	Non‐AMD	1 930 549	111 050	1 (Reference)	48 698	1 (Reference)	68 482	1 (Reference)
	AMD without visual disability	23 572	2415	1.05 (1.02–1.10)†	1012	1.05 (0.99–1.12)	1578	1.06 (1.01–1.12)†
	AMD with visual disability	1638	191	1.23 (1.06–1.41)†	83	1.26 (1.02–1.57)†	124	1.23 (1.03–1.46)†
P interaction			0.006†		0.523		0.003†
Cardiometabolic comorbidities
No	Non‐AMD	1 556 949	69 242	1 (Reference)	31 860	1 (Reference)	40 285	1 (Reference)
	AMD without visual disability	11 862	791	0.96 (0.90–1.03)	329	0.92 (0.83–1.03)	498	0.99 (0.90–1.08)
	AMD with visual disability	959	73	1.04 (0.83–1.31)	27	0.88 (0.60–1.28)	50	1.16 (0.88–1.54)
Yes	Non‐AMD	2 189 261	180 672	1 (Reference)	78 259	1 (Reference)	113 329	1 (Reference)
	AMD without visual disability	28 768	3574	1.02 (0.99–1.06)	1559	1.07 (1.01–1.12)†	2302	1.00 (0.96–1.05)
	AMD with visual disability	2164	318	1.20 (1.07–1.34)†	140	1.26 (1.07–1.48)†	211	1.21 (1.05–1.38)†
P interaction			0.201		0.014†		0.927

aHR indicates adjusted hazard ratio; and AMD, age‐related macular degeneration.

^*Adjusted for age, sex, income, area of residence, body mass index, smoking, alcohol consumption, regular physical activity, hypertension, diabetes, dyslipidemia, and Charlson comorbidity index, but excluding the variable used in stratification.

^† Statistically significant results.

Discussion

In this population‐based longitudinal study of the Korean population, we found that patients with AMD overall had no increased risk of CVD than non‐AMD groups. However, the risk was higher when they have visual disability (17%).

Our findings showed that AMD was associated with a 5% increased risk of MI. Previous investigations have mostly suggested no significant association between AMD and MI. Prospective studies from US populations (MESA [Multi‐Ethnic Study of Atherosclerosis] ²⁵ and ARIC study ²⁴ ) and retrospective studies using Israeli claims data ¹⁸ demonstrated no significant association. On the other hand, one US study using the US Centers for Medicare & Medicaid Services database showed a 19% higher risk of MI (aHR, 1.19 [95% CI, 1.16–1.22]), ¹⁷ whereas the findings from the study using the US Ingenix LabRx database concluded that the rates of MI were even lower in patients with neovascular AMD than in the control (aHR, 0.58 [95% CI, 0.48–0.72]). ¹⁹

We found that the risk of ischemic stroke was not significantly associated with AMD. Previous studies have demonstrated conflicting results: no association (the Rotterdam study ¹⁵ ) or a higher risk of ischemic stroke (40% in ARIC study ¹⁶ and 20% in the US Medicare study ¹⁴ ). Three meta‐analyses ¹¹ , ¹² , ³³ also found a nonsignificant association between AMD and stroke: Wu et al, 1.13 (95% CI, 0.93–1.36) ³³ ; Fernandez et al, 1.08 (95% CI, 0.81–1.44) ¹² ; and Wang et al, 1.13 (95% CI, 0.93–1.44). ¹¹

Several hypotheses have been proposed for the mechanisms of the association between AMD and the risk of CVD. Chronic inflammation and oxidative stress are plausible biologic mechanisms for both AMD and CVD. ³⁴ , ³⁵ , ³⁶ Current knowledge of the pathophysiology underlying AMD has highlighted the major role of oxidative stress in triggering outer blood–retinal barrier degeneration. ³⁷ , ³⁸ Therefore, current therapeutic options for neovascular AMD are targeting vascular endothelial growth factor in the retina. The vascular model proposes that AMD is a vascular disease characterized by the impairment of choroidal perfusion. ³⁹ , ⁴⁰ Hemodynamic changes in the choroidal vasculature are closely associated with oxidative stress, atherosclerotic process, and blood pressure, ³⁹ , ⁴⁰ suggesting a link between AMD and CVD. However, contrary to this hypothesis, our results suggest that AMD itself is not strongly associated with CVD risk and has only a weak association with MI if any, but no significant association with ischemic stroke.

Notably, in further analysis that subdivides AMD by visual disability, we found a significantly increased risk of CVD in patients with AMD who have visual disability. The reasons for this are not clear, but potential explanations include the following.

First, the differential association between AMD without visual disability and with visual disability may partially reflect the difference in severity and pathophysiology of AMD types. Visual disability in patients with AMD occurs when early AMD progresses to late AMD. In neovascular AMD, a late AMD type, the formation of choroidal neovascularization caused by vascular endothelial growth factor secreted as a result of choroidal ischemia, is a key to pathogenesis. ⁴¹ Choroidal vascular change, which can induce choroidal ischemia, is associated with systemic vascular abnormalities. ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ Therefore, neovascular AMD may have systemic vascular abnormalities at the baseline, which could lead to CVD during the follow‐up period. However, this hypothesis has a limitation in that it cannot explain the association between CVD and individuals with geographic atrophy, another type of late AMD. Findings from the CHS (Cardiovascular Health Study) showed that late AMD was not associated with incident coronary heart disease (aHR, 0.78 [95% CI, 0.25–2.48]) in a US population, ²² and one Dutch study (Rotterdam study) reported that late stages of AMD were not associated with cerebral infarction (aHR, 0.89 [95% CI, 0.43–1.82]). ¹⁵ In addition, a meta‐analysis showed that late AMD is not associated with total CVD (relative risk [RR], 1.17 [95% CI, 0.98–1.40]), coronary heart disease (RR, 1.12 [95% CI, 0.64–1.96]), or stroke (RR, 1.15 [95% CI, 0.74–1.78]). A positive association between late AMD and stroke in a Taiwanese population, but not in a Dutch population, might suggest a potential ethnic difference between late AMD and stroke. Further studies are needed to investigate whether different severity of AMD is differentially associated with CVD risk.

Second, patients with AMD who have visual disability have a more traditional risk factor of CVD than patients without visual disability. Visual impairment is associated with physical inactivity ⁴⁶ , ⁴⁷ , ⁴⁸ and is a risk factor for falls and hip fractures, ⁴⁹ , ⁵⁰ which may have additional adverse results for people with visual disability. For instance, individuals with visual disability may experience great difficulty in accessing health care services. For individuals with severe central vision loss, which is typical in advanced AMD, their inability to read a text may cause poor compliance (eg, by misunderstanding the prescription). Visual impairment can also have adverse functional consequences, including social isolation, restriction of daily activities, poor quality of life, and frailty, ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ especially in the elderly population. In addition, visual impairment significantly affects nutritional status, with a higher prevalence of obesity and malnutrition. ⁵⁹ For the multiple components that interplay between visual impairment and adverse health outcomes, our finding that individuals with AMD and visual disability have a much higher risk of CVD than the non‐AMD group or even AMD without visual disability may be explained. A further prospective study that specifies a chronological order in AMD and visual disability may enhance our understanding.

Notably, to our best knowledge, this is the first cohort study that demonstrated that AMD is differently associated with the risk of CVD (especially for ischemic stroke) by sex. Multiple mechanisms may contribute to the sex difference in our finding. First, estrogen may play a role in the risk of stroke ⁶⁰ , ⁶¹ , ⁶² and exudative AMD. ⁶³ Second, female predominance (86.2%) of subretinal drusenoid deposit in the Korean population, ⁶⁴ which is associated with the risk of stroke, ⁶⁵ may be another explanation. Last, the interaction of sex with the complement system, a key factor in the pathogenesis of AMD ⁶⁶ , ⁶⁷ and ischemic injuries, ⁶⁸ , ⁶⁹ may explain our findings. However, identifying the exact mechanism underlying these differential associations is beyond the scope of our study.

It is intriguing that associations between AMD, visual disability, and CVD were more prominent in patients with AMD with underlying cardiometabolic comorbidity than in patients with AMD without cardiometabolic comorbidity. A recent study reported that in individuals with visual impairment, the risk of cardiometabolic diseases was 70% higher than in age‐ and sex‐matched controls (HR, 1.7 [95% CI, 1.4–2.0]). ²⁷ However, previous studies have already shown that individuals with a disability were less likely to use appropriate health care services, ⁷⁰ , ⁷¹ and those with visual disability have lower medication adherence than those without visual disability. ⁷² Management of hypertension, diabetes, and dyslipidemia, which were altogether defined as cardiometabolic comorbidity in our study, involves an integrated care approach to lifestyle modifications, education, and medical treatment. Hence, clinical care should focus on screening and management of chronic diseases in patients with AMD for the prevention of CVD.

Our findings showed that AMD without visual disability was not associated with an increased risk of CVD, whereas AMD with visual disability was associated with a 17% increased risk of CVD. Therefore, individuals with both AMD and visual disability should be monitored for CVD risk. The underlying pathophysiology between advanced AMD and CVD needs to be elucidated in future studies.

Several limitations should be considered in interpreting our results. First, this is an observational study that can only suggest associations and cannot establish a causal relationship. Second, as we relied on the diagnostic code for AMD diagnoses, AMD might be underdiagnosed, especially when subjects do not have visual disability. In addition, most patients with AMD in Korea have early AMD without symptoms that cannot be identified in a claims database. Therefore, the association between AMD and CVD could be more attenuated in our findings than in the actual population. Third, only 3‐digit ICD‐10 codes (eg, H353) were available for the study population from the Korean NHIS database in 2009. Therefore, individuals diagnosed with H35.33 (macular hole) were likely included among the study participants but are only estimated to represent ≈5% of our study population (data not shown); thus, the impact of their inclusion is unlikely to be substantial. Fourth, our data from the NHIS database do not have the cause of death information, and therefore, we were unable to consider cardiovascular mortality in this study. Fifth, our results cannot be generalized to other ethnic groups considering the ethnic difference in genetic factors related to AMD. ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ Last, other possible confounders, including nutrition and dietary factors, ⁵⁰ , ⁷⁷ were not controlled.

In summary, AMD with visual disability, but not all AMD, was associated with an increased risk of CVD. Patients with AMD who have visual disability should be targeted for risk factor management and disease prevention for CVD.

Sources of Funding

This research was partially supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute, funded by the Ministry of Health and Welfare, Republic of Korea (grant HI20C1073), which was received by Dr Shin; and by a grant from the National Research Foundation of Korea, funded by Ministry of Education, Republic of Korea (NRF‐2021R1C1C1007795), which was received by Dr Lim.

Disclosures

None.

Supporting information

Tables S1–S3