Diabetes-Related Risk Factors for Exudative Age-Related Macular Degeneration: A Nationwide Cohort Study of a Diabetic Population

Abstract

Purpose: The purpose of this study was to identify diabetes-related risk factors for exudative age-related macular degeneration (AMD).

Methods: This was a nationwide population-based cohort study using authorized clinical data provided by the Korean National Health Insurance Service. A total of 1,768,018 participants with diabetes over 50 years of age participated in the Korean National Health Screening Program between 2009 and 2012. Data on covariates, including age, sex, income level, systemic comorbidities, behavioral factors, and diabetes-related parameters, including duration of diabetes, use of insulin for diabetes control, number of oral hypoglycemic agents used, and accompanying vision-threatening diabetic retinopathy, were collected from health screening results and claims data. Patients were followed up until December 2018. Incident cases of exudative AMD were identified using registered diagnostic codes from the claims data. The prospective association of diabetes-related parameters with incident exudative AMD was investigated using the multivariable-adjusted Cox proportional hazard model.

Results: During an average follow-up period of 5.93 years, 7331 patients were newly diagnosed with exudative AMD. Compared to those who had diabetes for less than 5 years, individuals with diabetes for 5 years or more had a greater risk of future exudative AMD development, with a hazard ratio (95% confidence interval) of 1.13 (1.07–1.18) in the fully adjusted model. Use of insulin for diabetes control and the presence of vision-threatening diabetic retinopathy were also associated with an increased risk of exudative AMD with a hazard ratio (95% confidence interval) of 1.16 (1.07–1.25) and 1.40 (1.23–1.61), respectively.

Conclusions: A longer duration of diabetes, administration of insulin for diabetes control, and comorbid vision-threatening diabetic retinopathy were associated with an increased risk of developing exudative AMD.

Age-related macular degeneration (AMD) is a major cause of visual impairment in the elderly and is the most common cause of blindness in developed countries.¹ Aging population worldwide is leading to a steep increase in individuals affected with AMD. The estimated worldwide prevalence of AMD is 196 million in 2020, increasing to 288 million in 2040.² AMD increases health expenditure, accounting for an estimated $255 billion in 2010 worldwide,³ and the socioeconomic burden associated with the disease will expand exponentially in the future. 

Therefore, identifying modifiable risk factors for AMD is of clinical significance. The pathogenesis of AMD is complex and involves multiple risk factors, including genetic variations, ethnicity, hypertension, dyslipidemia, cardiovascular diseases, obesity, and cigarette smoking.⁴ Diabetes mellitus (DM) has also been suggested to be associated with an increased risk of AMD and regarded as a modifiable risk factor, because many studies have reported its impact on the development and progression of AMD.^5–9 Clinicians are mainly concerned about diabetic retinopathy when dealing with patients with diabetes. However, accumulated epidemiologic evidence demonstrates that there is also a positive association between DM and the risk of AMD.^5–12 This underscores the possible beneficial influence of controlling DM on the development of AMD, in addition to preventing the progression of diabetic retinopathy and other systemic comorbidities. 

However, previous studies have mostly focused on how the presence of DM influences the risk of AMD and the impact of severity, duration, and control of DM on development of AMD have not been studied in earlier research. Furthermore, the influence of diabetic retinopathy on the risk of AMD development is poorly understood. To address these questions, we investigated the association of unique traits of diabetes, including duration, treatment modality, and comorbid diabetic retinopathy with the risk of exudative AMD development in patients with diabetes using a nationwide population-based cohort in South Korea. 

Methods

Setting

This nationwide, population-based, retrospective cohort study used data from the Korea National Health Insurance Service (NHIS). The NHIS is a mandatory universal medical care system administered by the South Korean government and covers all registered citizens in South Korea. The NHIS holds medical information for the entire population, including demographic, mortality, and health claims data. Demographic data include age, sex, and income. The claims data include the date of clinical visits, prescription records, and diagnostic codes defined by the Korean Classification of Diseases seventh revision (KCD-7), based on the International Classification of Diseases, 10th revision, with a few changes specific to Korea. The NHIS also provides biennial national health checkup services to the general Korean population for disease prevention and early detection. The data generated from national health screening included patients’ responses to structured questionnaires, anthropometric measurements, and laboratory test results. Demographic, mortality, health claims, and health-screening data can be linked through de-identified key numbers assigned to each individual. This database has been widely used in previous studies to identify associations between various diseases and risk factors.^13,14 Detailed information on the database profile is provided elsewhere.¹⁵ 

The study adhered to the Declaration of Helsinki and was approved by the Institutional Review Board of the Samsung Medical Center, Seoul, Republic of Korea (IRB File Number 2020-04-146). The board waived the requirement for informed consent based on the use of de-identified public data and retrospective study design. 

Study Population

Of the 23,452,862 individuals who underwent national health screening between 2009 and 2012, 2,746,079 patients with diabetes were identified based on health screening results and diagnostic code/prescription codes from claims data. Those with a fasting glucose level ≥126 mg/dL at the first health screening examination or having diagnostic codes for diabetes (KCD-7 code: E11-E14) with medication codes for diabetes in the same claim within a year before the health screening examination were considered to have diabetes. Among them, 708,166 individuals aged <50 years at baseline were excluded. Further, we excluded 115,635 subjects diagnosed with macular diseases (KCD-7 code: H35.3) before baseline health screening. Two years of the lag period was applied to reduce the bias of reverse causality. Participants who died within 2 years after the examination (n = 47,267) and who were diagnosed with macular diseases within a 2-year lag period from the examination date (n = 37,142) were excluded. Those with missing data on key variables required to define covariates (n = 69,851) were also excluded. Finally, 1,768,018 patients with diabetes over 50 years of age were included in the study (Fig. 1). 

Figure 1.

 

Flow chart of the cohort study design.

Definition of Exudative Age-Related Macular Degeneration and Follow-Up

According to the Korean government's policy on enhanced support for rare intractable diseases, patients diagnosed with exudative AMD pay 10% of the healthcare cost, including the costs for imaging and intraocular anti-VEGF injections. The government supports the remaining 90% of healthcare costs under the condition that a board-certified ophthalmologist provides medical records, fundus photography, optical coherence tomography, and fluorescein angiography to support the diagnosis of exudative AMD in the patient, and other ophthalmologists confirm the diagnosis in the Health Insurance Review and Assessment Service. These multimodal imaging-confirmed exudative AMD cases were assigned a special registration code, V201, in the NHIS database. 

We regarded individuals with diagnostic code H35.31 and special registration code V201 in the same claim as exudative AMD cases. For each case, we defined the first date of the claim with codes H35.31 and V201 as the time of exudative AMD incidence. Patients were followed starting 2 years (lag period) from the baseline health checkup date to the date of incident exudative AMD (diagnostic code H35.31, registration code V201), death, or the end of the study period (December 31, 2018), whichever came first. 

We did not include dry AMD in our outcomes because it is usually asymptomatic and patients do not receive care, which would lead to bias in outcome ascertainment. 

Exposures and Covariates

Diabetes-related parameters included baseline fasting glucose levels, duration of DM at baseline, insulin usage, number of oral hypoglycemic agent classes used, and comorbid vision-threatening diabetic retinopathy (VTDR). Diabetes medications prescribed before the baseline examination were reviewed. Oral hypoglycemic agents were subdivided into 7 classes: (1) biguanides, (2) sulfonylureas, (3) meglitinides, (4) thiazolidinediones, (5) alpha-glucosidase inhibitors, and (6) dipeptidyl peptidase 4 inhibitors. The DM duration was categorized as <5 years and ≥5 years based on the period prescribed with diabetic medication before the baseline examination. Those who had been prescribed insulin before the baseline examination were classified as having insulin-using DM. The number of oral hypoglycemic agent classes used was categorized into less than three and greater than or equal to three classes. VTDR, including proliferative diabetic retinopathy and clinically significant diabetic macular edema, was defined as having a diagnostic code for diabetic retinopathy (H36.0) with procedure codes for pan-retinal photocoagulation (S6160 and S5161), focal laser for macular edema (S51.64), vitrectomy (S5121 and S5122), intravitreal injection (S5070), and posterior sub-tenon injection (KK130) before baseline health screening. 

Possible confounding factors were comprehensively assessed using the methodology of previously published nationwide epidemiologic studies.^13,14 Comorbid hypertension and dyslipidemia were identified based on the health screening results of blood pressure (hypertension, systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg), total cholesterol (dyslipidemia, ≥240 mg/dL), and the presence of diagnostic codes (KCD-7 code: I10-13, I15 for hypertension and E78 for dyslipidemia) combined with medication prescription codes within a year before the health screening examination. Chronic kidney disease was defined as an estimated glomerular filtration rate of <60 mL/min/1.73 m² calculated from the serum creatinine level. Income level was categorized into quartiles according to the insurance premium level, determined by total household income. We also collected data regarding health-related behaviors based on participants’ responses to the health-screening questionnaire. Smoking status was classified as none, past, or current. Drinking habits were categorized as none, mild (<30 g/day), or heavy (≥30 g/day). Regular exercise was defined as performing a moderate level of physical activity for more than 30 minutes per day for more than 5 days per week. The body mass index (BMI) was calculated as the weight (kg) divided by height squared (m²) and categorized as underweight (BMI < 18.5 kg/m²), normal weight (18.5 ≤ BMI < 23 kg/m²), overweight (23 ≤ BMI < 25 kg/m²), obese I (25 ≤ BMI < 30 kg/m²), and obese II (≥ 30 kg/m²), according to the Korean Society for the Study of Obesity.¹⁶ 

Statistical Analyses

We calculated the incidence rates of exudative AMD by dividing the number of incident cases by the total number of person-years. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using the Cox proportional hazard models. The variables adjusted for each Cox model were as follows: model 1 = crude; model 2 = age and sex; model 3 = demographic factors (age, sex, and income level), systemic comorbidities (hypertension, dyslipidemia, and chronic kidney disease), and behavioral factors (smoking status, drinking habits, regular exercise, and BMI); and model 4 = demographic factors, systemic comorbidities, behavioral factors, and diabetes-related parameters (duration of diabetes, insulin usage, number of oral hypoglycemic agent classes, and comorbid VTDR). Additional subgroup analyses were performed based on age and sex. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 4.0.4 (The R Foundation for Statistical Computing, Vienna, Austria; http://www.r-project.org). The P values were two-sided and considered statistically significant at values less than 0.05. 

Results

Characteristics of the Study Population

Table 1 presents the detailed baseline characteristics of the diabetic individuals in this study. The mean age of the study population was 62.41 years; 985,538 (55.74%) individuals were men. Around 35% of individuals with diabetes had diabetes for more than 5 years; 9% used insulin to control their disease, and 2% had VTDR. The average follow-up period was 5.93 years, and 7331 subjects were diagnosed with and treated for exudative AMD during the follow-up period. Patients with exudative AMD had a greater mean age at the start of the study period than those without exudative AMD (62.39 vs. 66.40 years, respectively). Moreover, they showed a higher proportion of insulin usage, comorbid VTDR, and a greater prevalence of hypertension, dyslipidemia, and chronic kidney disease. 

Table 1.

 

Baseline Characteristics of the Study Population

 

Diabetes-Related Risk Factors for Exudative Age-Related Macular Degeneration

Table 2 shows the HRs and 95% CIs for the association between diabetes-related variables and risk for exudative AMD in patients with diabetes. In model 4, the HR for exudative AMD was 1.11 (95% CI = 1.06–1.17) for those who had diabetes for more than 5 years compared with less than 5 years. Insulin usage for diabetes control was also associated with an increased risk of exudative AMD, with an HR of 1.10 (95% CI = 1.02–1.19). Comorbid VTDR significantly increased the risk of exudative AMD, with an HR of 1.31 (95% CI = 1.14–1.50). The number of oral hypoglycemic agent classes did not significantly affect the risk of exudative AMD development in this study. The HRs and 95% CIs for the associations between other covariates and the risk of exudative AMD in patients with diabetes are provided in Supplementary Table S1. 

Table 2.

 

Hazard Ratios and 95% Confidence Intervals for Development of Exudative Age-Related Macular Degeneration According to Baseline Diabetes-Related Variables

 

Subgroup Analyses

Figure 2 shows the stratified HRs and 95% CIs for the associations between diabetes-related variables and the risk of exudative AMD in the subgroups by age and sex. The influence of diabetes on insulin and comorbid VTDR on exudative AMD resulted in significant interactions in subgroup analysis by age. The younger subgroup (<65 years) had an HR of 1.32 in the insulin-using group, whereas the older subgroup (≥65 years) had an HR of 1.07 (P for interaction = 0.007). The younger subgroup also had a greater HR of 1.71 for comorbid VTDR compared with an HR of 1.22 in the older subgroup (P for interaction = 0.013). Supplementary Table S2 presents the outcomes of stratified Cox analyses conducted within subgroups based on the duration of diabetes, insulin usage, and comorbid VTDR, revealing no significant interaction effects observed. 

Figure 2.

 

Association of diabetes-related variables with incident exudative age-related macular degeneration in subgroups by age and sex. Exudative age-related macular degeneration risks were estimated using the Cox proportional hazards model, adjusted for demographic factors, systemic comorbidities, and behavioral factors. (A) Significant interaction existed between insulin usage and age (P = 0.007) and vision-threatening diabetic retinopathy and age (P = 0.013) in development of exudative age-related macular degeneration. (B) No significant interactions were observed, and the associations were consistent in subgroup analyses by sex.

Discussion

This nationwide population-based cohort study revealed the influence of various diabetes-related parameters on the incidence of exudative AMD. After adjustment for relevant covariates, longer duration of DM, DM on insulin, and VTDR were associated with an increased risk of exudative AMD. The detrimental influence of insulin usage and comorbid VTDR was greater in younger individuals (<65 years). 

Previous studies from various data sources, designs, and ethnic groups demonstrated an increased risk of AMD in patients with diabetes.^5–8 Some prospective cohort studies showed a significant association of diabetes with advanced AMD in particular, but not with early AMD. The Barbados Eye Study¹⁰ and Women's Health Initiative Study¹¹ reported that DM is associated with an increased risk of late AMD. The Age-Related Eye Disease Study (AREDS) also concluded that DM increased the risk of neovascular AMD by 1.88 fold.¹² Additionally, both the Beaver Dam Eye Study and the European Eye Study (EUREYE), cross-sectional population-based studies from the United States and Europe, presented that patients with DM have an increased prevalence of exudative AMD compared with individuals without diabetes.^17,18 Some other contradictory evidence also exists, which reports that DM is not related to the development of AMD.^19,20 However, systematic reviews and meta-analyses concluded that DM is a risk factor for AMD, and the association is especially greater for advanced AMD.^8,9 Taken together, previous literature suggests that patients with diabetes carry a higher risk of AMD compared with patients without diabetes. 

Although the underlying molecular mechanisms are not well understood, the increased risk of AMD in individuals with DM is thought to involve functional and structural changes in the RPE, Bruch's membrane, and choroidal circulation. Hyperglycemia-induced chronic inflammatory responses and the accumulation of advanced glycation end products (AGE) in RPE and photoreceptor cells may contribute to cellular damage and apoptosis.^21,22 Histopathological studies also have shown alterations in the choriocapillaris basement membrane, Bruch's membrane, and choroidal thickness in patients with diabetes.^23,24 These changes can lead to RPE dysfunction, outer retinal hypoxia, and a higher likelihood of developing drusen and AMD. Additionally, AGE accumulation may upregulate VEGF mRNA expression in RPE cells, potentially playing a role in choroidal neovascularization.^25,26 

In the present study, individuals using insulin to control their diabetes had a higher risk of exudative AMD. Insulin treatment is typically reserved for cases of severe or poorly controlled diabetes, suggesting a potential association between disease severity and AMD risk. Additionally, longer diabetes duration was associated with an increased risk of exudative AMD. These findings align with the suggested mechanism of diabetes impacting AMD development. Prolonged diabetes duration and more severe disease may lead to greater damage to the outer retina due to inflammation and hypoxia, as well as increased VEGF production, thereby increasing susceptibility to exudative AMD. 

Also of interest, the presence of VTDR was associated with an increased risk of exudative AMD. The association between diabetic retinopathy and AMD remains controversial and poorly understood. Some studies suggest that DR increases the prevalence of AMD in patients with diabetes,^6,8,27 whereas others propose a protective effect.^28,29 Diabetic retinopathy and AMD are distinct clinical entities and involve different anatomic structures. However, both conditions involve a common pathogenic process, potentially driven by inflammation. Pro-inflammatory cytokines play an important role in the pathogenesis of diabetic retinopathy³⁰ by compromising retinal/choroidal vasculature and causing blood-retinal barrier breakdown.³¹ These processes can induce hypoxia in the outer retina and may possibly lead to an increased risk of exudative AMD. In light of our study results, presence of VTDR might have been a marker for increased levels of pro-inflammatory cytokines that are involved in pathogenesis of both diabetic retinopathy and exudative AMD. 

Subgroup analyses suggest insulin usage and presence of VTDR were stronger markers of risk of exudative AMD in younger (<65 years) individuals. The reason for this finding is not clear. Previous studies reported that ocular involvement in diabetes is different in the older than in a younger population, with a greater prevalence of diabetic retinopathy but less frequent proliferative diabetic retinopathy in the older individuals.^32,33 These findings suggest that the influence of DM on the eyes may differ with the age of the affected individuals. In addition, the influence of DM may be relatively less in older age groups because the influence of age on AMD might be greater in older individuals over 65 years of age. Future studies should provide a plausible explanation for this finding. 

The overall results of the present study indicate that more prolonged exposure to severe DM accompanied by VTDR is associated with an increased risk of exudative AMD, especially in the younger age group. This suggests that good control of DM may decrease the frequency of VTDR and reduce exudative AMD in patients with diabetes. Proving causality might be difficult as DM has a complex interplay with other systemic comorbidities. Further, the representative oral hypoglycemic agent, metformin, has also been suggested to influence the progression of AMD,^34,35 making the interpretation of results more complicated. Our study demonstrates the prospective association of duration, therapeutic control of DM, and diabetic retinopathy with the future development of exudative AMD. Whether strict control of DM could reduce not only the risk of diabetic retinopathy, but also the risk of AMD should be elucidated in future research. 

The strengths of the present study include the large number of subjects using population-based claims data and the adoption of a longitudinal study design for the evaluation of prospective associations. In addition, the topic and question of the study have not been investigated in previous literature, making the results worth reporting. However, this study has some limitations that need to be recognized. First, we could not include dry AMD as an outcome because it is usually asymptomatic, and only patients who visited the ophthalmology clinic by chance would have been recorded in the claims data, leading to bias in outcome ascertainment. Future cohort research with serial fundus examinations is required to evaluate the association between diabetes-related variables and dry AMD. Second, the study might have missed those who were unable to access the health care system. This explains our finding that a higher income level is associated with a greater risk of exudative AMD. However, we adjusted for income levels in the multivariable analysis. This would have decreased the bias that could have been caused by those who had exudative AMD but did not receive proper treatment. Last, some misclassification in outcomes and covariates may inevitably have occurred. Although we have utilized well-established definitions in order to mitigate the impact of misclassification on our analyses and results, complete elimination is unattainable. Further studies addressing these limitations are required. 

In conclusion, this nationwide population-based cohort study of patients with diabetes demonstrated that a longer duration of DM, insulin administration for DM control, and accompanying VTDR are independent risk factors for the development of exudative AMD. The results suggest that improving the control of DM and reducing the occurrence of VTDR may decrease the development of exudative AMD in patients with diabetes. However, future studies are needed to elucidate the exact relationship between diabetes control and AMD. 

Acknowledgments

Supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant numbers HI19C0481 and HC19C0142), and a National Research Foundation of Korea grant funded by the Korean government's Ministry of Education (NRF-2021R1C1C1007795; Seoul, Korea), which D.H.L. received. The funder had no role in the design or conduct of the study. 

Author Contributions: All authors attest that they meet the current ICMJE criteria for authorship. 

Disclosure: S. Hwang, None; S.W. Kang, None; S.J. Kim, None; K.N. Lee, None; K. Han, None; D.H. Lim, None