Prevalence of Age-Related Macular Degeneration in the United States: A Medicare-Based Analysis from 2014 to 2021

Abstract

Introduction

Age-related macular degeneration (AMD) is a leading cause of vision loss among older adults in the USA. Understanding its prevalence and demographic distribution is critical for developing targeted public health strategies. This study aimed to assess the prevalence of AMD and its clinical stages among US Medicare beneficiaries aged 65 years and older.

Methods

We conducted a retrospective cohort study using the Vision and Eye Health Surveillance System (VEHSS) database of Medicare beneficiaries diagnosed with AMD between 2014 and 2021. Crude prevalence rates for overall AMD, early AMD, intermediate AMD, wet AMD, and geographic atrophy (GA) were calculated at national and state levels. Prevalence was stratified by age, sex, and race/ethnicity. Statistical analyses included the Mann-Whitney U test for age and sex comparisons, the Brown-Forsythe one-way ANOVA for racial/ethnic comparisons, and the Dunnett T3 test for post hoc analyses.

Results

In 2021, the VEHSS-Medicare dataset included 24,129,807 individuals aged 65 and older, among whom the national prevalence of AMD was 10.40%. Prevalence rates for early AMD, intermediate AMD, wet AMD, and GA were 2.87%, 6.91%, 2.14%, and 0.73%, respectively. The number of AMD cases increased from 2.33 million in 2014 to 2.51 million in 2021. Prevalence was significantly higher in individuals aged ≥85 years compared to those aged 65–84 years, and in females compared to males. Post hoc analyses demonstrated that White individuals had a significantly higher prevalence of AMD compared with all other racial/ethnic groups.

Conclusions

The prevalence of AMD among US adults aged 65 years and older was 10.4%, with higher rates observed in the oldest age groups, females, and White individuals. These findings highlight the importance of addressing disparities in AMD prevention and care, particularly in populations at greatest risk.

Introduction

Age-related macular degeneration (AMD) is a progressive, multifactorial disease and a leading cause of vision loss and blindness among the elderly worldwide [1–3]. The pathogenesis of AMD remains incompletely understood; however, the current evidence suggests a complex interplay of aging, genetic predisposition, and environmental factors [4]. A hallmark feature of the disease is the accumulation of extracellular deposits, known as drusen, within and beneath the retinal pigment epithelium (RPE). These deposits contribute to oxidative stress and subsequent photoreceptor degeneration [5]. As the disease progresses through its various stages, it leads to substantial vision impairment, imposing a significant socioeconomic burden [2].

AMD exhibits significant phenotypic variability; however, its clinical diagnosis and staging primarily rely on identifying drusen, pigmentary abnormalities, and reticular pseudodrusen [2, 6, 7]. The widely accepted Beckman classification system stratifies AMD into three stages based on these retinal findings [8]. Early AMD is characterized by the presence of medium-sized drusen (63–125 µm), intermediate AMD by larger drusen (>125 µm) and/or pigmentary changes, and late (advanced) AMD is divided into two distinct forms. The dry form, known as geographic atrophy (GA), involves the progressive atrophy of photoreceptors, RPE, and choriocapillaris. In contrast, the wet (neovascular) form is marked by the development of pathological macular neovascularization [9].

Epidemiological studies have shown considerable variability in AMD prevalence rates across different populations and study methodologies [10]. As of May 2013, the global prevalence of AMD among adults aged 45 and older was estimated at 8.7% [3], with projections indicating this number could reach 288 million by 2040, underscoring the growing public health and economic burden of the disease [3, 11]. In the USA, data from 2019 estimated that the crude prevalence rates of AMD among individuals aged 40 and older were 11.64% for early stage and 0.94% for late-stage disease, affecting 18.34 million and 1.49 million individuals, respectively [12]. Given the ongoing development of novel therapeutic strategies for AMD [13, 14], it is critical to periodically reassess the disease prevalence and its societal impact.

In the USA, Medicare serves as the national health insurance program for individuals aged 65 and older, as well as those with certain disabilities and conditions [15]. As of March 2023, nearly 65 million individuals over the age of 65 were enrolled in Medicare, representing a significant proportion of this age group [16]. De-identified, person-level data for these individuals are reported through the Vision and Eye Health Surveillance System (VEHSS), facilitating population-level insights into eye health [17, 18].

This study examines the crude prevalence rates of AMD among USA residents aged 65 and older from 2014 to 2021, using data from the VEHSS database. Furthermore, it examines variations in prevalence across different stages of AMD and among demographic subgroups, stratified by age, sex, and racial/ethnic groups.

Methods

Data Source and Study Design

VEHSS was developed at the University of Chicago in collaboration with the Centers for Disease Control and Prevention (CDC), to aggregate data from a variety of public and private health sources, including Medicare, Medicaid, employer-sponsored insurance, managed vision care providers, and the American Academy of Ophthalmology’s Intelligent Research in Sight (IRIS) registry [18]. VEHSS-CDC plays a crucial role in monitoring epidemiologic trends related to vision loss and ophthalmic conditions in the USA, and disseminates findings to healthcare professional, researchers, and policymakers [17].

In this retrospective cohort analysis, we used Medicare data provided to VEHSS as the primary source to determine the prevalence, demographics, and health service utilization associated with AMD in the USA. This data, derived from reimbursed claims within traditional fee-for-service (FFS) Medicare plans, includes individual-level information on eligibility, enrollment, service usage, diagnoses, and payments [18].

Study Population

The study cohort included all patients aged 65 years and older who were enrolled in Medicare and diagnosed with AMD between 2014 and 2021. AMD diagnoses and stages were identified using the International Classification of Diseases (ICD)-9 (362.51-52) and ICD-10 (H35.31-32) codes, as recorded in Medicare FFS reimbursement claims (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000548724). Specifically, we calculated the national crude prevalence rates across various demographic subgroups, including age (below and above 85 years), sex, and racial/ethnic subgroups (Asian, White, Black, Hispanic, Native American, Other).

Statistical Analysis

Statistical analyses were conducted using R-Studio Version 2023.12.0+369 (R Foundation for Statistical Computing). The total annual burden of AMD was calculated based on the yearly prevalence data and reported, with upper and lower thresholds determined by 95% confidence intervals.

The Mann-Whitney U test was used to assess differences in AMD prevalence between binary groups categorized by age and sex. To compare prevalence across multiple racial subgroups, a one-way Brown-Forsythe analysis of variance (ANOVA) was used. Post hoc comparisons among racial/ethnic groups were performed using the Dunnett T3 multiple comparison test. Statistical significance was defined as a p value <0.05.

Results

This study analyzed data from approximately 25 million individuals aged 65 and older, enrolled in Medicare from 2014 to 2021, using the VEHSS-CDC database. By 2021, the total US population aged 65 and older was approximately 55.89 million, indicating that Medicare data represented nearly half of this demographic during the study period (47.0–53.7%). The annual ratio of Medicare enrollment to the total population aged 65 and older was calculated, and the national case burden was estimated through linear extrapolation based on these ratios.

The mean estimated prevalence of any AMD in this cohort increased progressively from 9.40% in 2014 (estimated case burden: 4,180,838) to 10.40% in 2021 (estimated case burden: 5,812,769) (Fig. 1). These findings, including the Medicare case burden and the annual estimated AMD case burden and prevalence from 2014 to 2021, are detailed in Table 1.

Fig. 1.

Prevalence of AMD and its various stages in Medicare beneficiaries over the age of 65 during the study period.

Table 1.

Prevalence and case burden of AMD among Medicare beneficiaries during the study period

Year	% prevalence of AMD	Total number of individuals enrolled in Medicare	Estimated case burden of AMD in Medicare patients
2014	9.43	24,799,576	4,180,838
2015	9.93	24,948,640	4,553,406
2016	10.69	25,427,502	5,087,529
2017	10.60	25,448,113	5,223,044
2018	10.65	25,388,732	5,414,480
2019	10.74	25,430,167	5,669,431
2020	9.38	24,952,025	5,235,208
2021	10.40	24,129,807	5,812,769

Prevalence by Age and Gender

Throughout the study period, the estimated total case burden of AMD among individuals enrolled in Medicare was consistently higher in the 65–84 years group and showed a progressive increase in both analyzed age groups (online suppl. Table 2). The estimated prevalence of any type of AMD was significantly higher in Medicare-enrolled individuals aged 85 and older compared to the 65-84 age group (20.78–23.55% vs. 7.28–8.36%; p < 0.0001) (Fig. 2). Notably, prevalence was estimated at 9.26–10.77% in females and 6.32–7.44% in males, with females consistently exhibiting a higher prevalence in Medicare-enrolled individuals throughout the study period (p < 0.001) (Fig. 3). The comparative annual case burden in males and females is detailed in online supplementary Table 3.

Fig. 2.

Prevalence of AMD and its various stages in Medicare beneficiaries between the age of 65–84 years and those older than 85 years, during the study period.

Fig. 3.

Prevalence of AMD and its various stages in male and female Medicare beneficiaries over the age of 65, during the study period.

Prevalence by Race and Ethnicity

From 2014 to 2021, AMD prevalence in Medicare-enrolled individuals and case burden increased across all racial groups, with the highest prevalence and cases observed among White individuals (Table 2). The post hoc analysis showed significant disparities, with White individuals showed significantly higher prevalence rates compared to all other racial groups (Table 3). Asians had significantly higher prevalence compared to Native Americans, Hispanics, and Black individuals. Native Americans and Hispanics individuals had higher rates than Black individuals.

Table 2.

Prevalence rates and case burden comparing AMD cases in different racial groups

Year	Asian, n (%)	White, n (%)	Black, n (%)	Hispanic, n (%)	Native American, n (%)	Other, n (%)
2014	49,400 (7.07)	2,172,900 (8.96)	61,600 (2.13)	78,700 (4.55)	7,600 (4.61)	13,200 (5.78)
2015	54,100 (7.58)	2,297,500 (9.45)	66,500 (2.32)	83,900 (4.86)	8,200 (4.84)	14,700 (6.44)
2016	65,600 (8.68)	2,506,900 (10.19)	75,700 (2.62)	95,900 (5.38)	9,500 (5.52)	17,200 (7.34)
2017	69,100 (8.86)	2,480,700 (10.15)	73,500 (2.60)	96,000 (5.36)	9,900 (5.58)	18,000 (7.73)
2018	72,100 (9.08)	2,481,800 (10.24)	72,200 (2.66)	96,200 (5.48)	10,100 (5.73)	19,000 (8.04)
2019	74,800 (9.29)	2,503,200 (10.40)	71,100 (2.73)	96,000 (5.55)	10,400 (6.05)	19,900 (8.38)
2020	58,500 (7.25)	2,156,600 (9.20)	57,400 (2.37)	76,400 (4.59)	8,400 (5.19)	16,900 (7.17)
2021	67,500 (8.49)	2,304,400 (10.23)	60,200 (2.81)	81,400 (5.20)	9,000 (6.27)	19,300 (8.40)

Table 3.

Post hoc analysis comparing AMD cases in different racial groups*

​	Asian	White	Black	Hispanic	Native
Asian	1	​	​	​	​
White	<0.0001	1	​	​	​
Black	<0.0001	<0.0001	1	​	​
Hispanic	<0.0001	<0.0001	<0.0001	1	​
Native	<0.0001	<0.0001	<0.0001	0.1956	1

*Adjusted p values calculated with Dunnett’s T3 multiple comparison test.

Prevalence across Different Stages of the Disease

Our analysis across AMD stages in Medicare-enrolled individuals showed a progressive increase in early AMD (2.23–2.87%), intermediate AMD (6.79–6.91%), and geographic atrophy (GA) (0.34–0.73%) during the study period. In contrast, the prevalence of wet AMD remained relatively stable, fluctuating from 2.17% in 2014 to 2.14% in 2021.

On analyzing the demographic groups affect by different stages of the disease, we observed that early AMD in Medicare-enrolled individuals was significantly more prevalent in individuals aged 85 and older (3.22–4.02% vs. 2.05–2.69%; p < 0.0001) (online suppl. Table 2), in females (2.57–3.3% vs. 1.79–2.31%; p < 0.0001) (online suppl. Table 3). Post hoc analysis, showed that early AMD was more prevalence in White and Asian populations compared to Black, Hispanic, and Native American groups (all p < 0.0001) (online suppl. Table 4). Intermediate AMD in Medicare-enrolled individuals also showed higher prevalence among those aged 85 and older (15.84–16.75% vs. 5.08–5.38%; p < 0.001) and in females (7.7–8.3% vs. 5.6–6.0%; p < 0.001) (online suppl. Tables 2, 3), with decreasing prevalence across White (7.5–8.0%), Asian (5.5–6.4%), Native American (4.9–5.4%), Hispanic (4.0–4.6%), and Black (1.8–2.0%) populations (online suppl. Table 5). Both GA and wet AMD in Medicare-enrolled individuals had as significantly higher prevalence in those aged 85 and older and in females (online suppl. Tables 2, 3). Among in Medicare-enrolled individuals, GA was particularly prevalent in White (0.4–0.9%) compared to Black and Hispanic groups (both p < 0.01) (online suppl. Table 6), while the prevalence of wet AMD showed a progressive decrease across White, Native American, Asian, Hispanic, and Black populations (p < 0.0001) (online suppl. Table 7). The state wise distribution of annual prevalence rates of various AMD stages in 2021 are highlighted in Figure 4.

Fig. 4.

a–e Geographic distribution of AMD cases in Medicare beneficiaries over the age of 65 in 2019.

Discussion

In this epidemiological claims-based study, we analyzed the estimated prevalence of AMD in the USA population aged 65 and older from 2014 to 2021, using data from 24,129,807 FFS Medicare beneficiaries reported to the VEHSS-CDC. We observed a steady increase in the prevalence of any AMD, rising from 9.43% in 2014 to 10.40% to 2021, with the highest rates among individuals aged 85 and older (20.78–23.55%), females (10.7–12.2%), and the White group (8.96–10.23%). Intermediate AMD emerged as the most prevalent stage (6.79–6.91%). Within each stage, individuals aged 85 and older, females, and the White population were the most affected subgroups, with no significant differences between White individuals and Asians in the prevalence of early AMD and GA.

AMD ranks as the third leading cause of blindness worldwide [19], and its prevalence is anticipated to rise due to the aging baby boomer generation, improved access to healthcare, and advancements in ophthalmic imaging techniques [3, 20]. The growing need for updated data on AMD burden is further emphasized by the recent introduction of new treatment options for advanced stages of the disease and the rapidly aging population [14, 21, 22]. With nearly the entire USA population over 65 being covered by Medicare [17], and the CDC-VEHSS system representing about half of this demographic, this data provides an crucial source for epidemiological studies on AMD [19].

Our findings confirm the trend of a progressive increase in AMD prevalence, as evidenced by the rise from 9.43% (estimated case burden: 4,180,838) in 2014 to 10.40% (estimated case burden: 5,812,769) in 2021. This supports the projections of a seminal systematic review, which forecasted a global increase in the total number of AMD cases [3]. Although reported prevalence rates largely vary [10, 12], our results are consistent with those of Rein et al. [12], who estimated a prevalence of 11.64% among Americans over 40 in 2019, using meta-regression on VEHSS data from multiple sources. Furthermore, Klein et al. [23] reported a prevalence of 6.5% among individuals over 40 based on 2005–2008 NHANES data, which increased to 13.4% (11.5–15.5) among those over 60.

Our study also marks the first systematic analysis of AMD prevalence across all stages using ICD-9 and ICD-10 diagnostic codes. Throughout the study period, intermediate AMD emerged as the most frequently diagnosed stage (6.79–6.91%), followed by early AMD (2.23–2.87%), wet AMD (2.17–2.14%), and GA (0.34–0.73%). Although not directly comparable to previous studies that typically categorize AMD as early- and late-stage, our findings confirm a higher prevalence of earlier stages of the disease compared to late or advanced AMD [12, 24]. Notably, the prevalence of late AMD observed in our study was slightly higher than those reported in previous research. This difference may be attributed to our focus on Medicare beneficiaries aged 65 and older, in contrast to other studies that included younger populations, typically 40 and above (Friedman et al. [25], 1.47%; Klein et al. [23] 0.8%; Rein et al. [12], 0.94%). Additionally, we observed a higher prevalence of wet AMD compared to GA, which contrasts with findings from prior meta-analyses by Wong and Jonas [3, 24]. This could be explained by the more symptomatic nature of wet AMD, which leads to more frequent clinical visits and a broader range of treatment options, resulting in greater data availability for these patients.

Our study also examined AMD prevalence across various age, sex, and racial and ethnic groups. We found a significantly higher prevalence of AMD among Medicare beneficiaries aged 85 and older compared to those aged 65–84, both for any AMD and across all stages. This is consistent with earlier studies indicating a greater disease burden with advancing age [24, 26]. Discussions surrounding sex differences in AMD prevalence also warrant attention. While some earlier studies suggest a possible association between female sex and the development of late AMD [4, 25, 27, 28], more recent meta-analyses have not identified significant gender differences across AMD stages [3, 24]. In contrast, our findings indicate a consistently higher prevalence of AMD among female Medicare beneficiaries throughout the study period (10.7–12.2% vs. 7.7–8.8%, p < 0.001), which aligns with data reported by Wittenborn et al. [26] Although some ascertainment bias might have affected our analysis – potentially due to higher rates of AMD-related blindness occurring in women [24] – we cannot entirely rule out the influence of factors such as postmenopausal estrogen level changes, genetic predisposition, and sex-specific differences in lipid metabolism on AMD pathogenesis and progression [27, 28]. Additionally, the distribution of male and female beneficiaries aged 65 and older in Medicare during 2014–2021 is consistent with the USA male-to-female ratio reported in the Census.

Lastly, our findings confirm that Whites exhibit a higher prevalence of any AMD (8.96–10.23%) compared to Asians and other ethnic groups [23, 29, 30], corroborating previous studies [3]. This may be linked to a higher genetic susceptibility, particularly related to variants in the CFH gene [31, 32]. Notably, Wittenborn and colleagues recently estimated AMD prevalence rates of 10.24% for Whites and 9.08% for Asians in 2018 among Medicare FFS beneficiaries. Stage-wise analysis revealed that both early and intermediate AMD are particularly prevalent among White and Asian individuals, with no significant differences in early AMD prevalence between these two groups. Despite the scarcity of data on early AMD in the literature, our results align with previous meta-analyses suggesting a slightly higher prevalence of intermediate AMD (often classified as early stage AMD) among individuals of European descent [3]. GA was more prevalent among Whites compared to Black and Hispanic groups, further consolidating findings from earlier meta-analyses [24]. In contrast, while wet AMD has historically been considered more common among Asians, our study found a higher prevalence among Whites in the USA. This difference could be attributed not only to various epidemiological factors but also to advances in retinal imaging that have improved the differentiation between wet AMD and neovascularization secondary to other causes, like pathological myopia or pachychoroid disease [33].

This study offers a comprehensive estimate of the AMD case burden among Medicare beneficiaries aged 65 and older in the USA from 2014 to 2021. We observed a steady increase in AMD prevalence, rising from 9.4% to 10.40% during this period, with most cases classified as intermediate AMD. Subgroup analyses revealed a significantly higher prevalence of AMD and its stages, particularly among individuals over the age of 85. Disparities in the diagnosis and treatment of AMD are influenced by several factors, including socioeconomic status, geographic location, comorbidities, and racial/ethnic differences. Patients from lower socioeconomic backgrounds are more likely to present with more severe visual acuity reduction at baseline and experience delays in treatment completion for wet AMD [34–36]. This association persists even after adjusting for age, gender, and distance from the hospital. Additionally, geographic disparities exist, with patients in remote areas and regions with fewer ophthalmologists and optometrists being less likely to receive timely anti-VEGF treatment for neovascular AMD [37]. Furthermore, there is a significant delay in the care pathway from symptom onset to treatment initiation, exacerbated by suboptimal awareness of AMD among patients [38]. Expanding access to advanced diagnostic tools in underserved areas and integrating routine eye exams into Medicare-covered services could enhance early detection and treatment access. Comorbid conditions such as dementia also contribute to disparities. Patients with dementia have reduced rates of eye care visits and receive fewer intravitreal injections for neovascular AMD compared to those without dementia [39]. This suggests that cognitive impairment can significantly impact the management and outcomes of AMD. Therefore, age-specific screening programs focusing on older individuals, who exhibit significantly higher prevalence rates, are essential.

Racial and ethnic disparities are evident in clinical trial participation and disease prevalence. Black, Hispanic, and other non-White participants are underrepresented in clinical trials leading to FDA drug approvals for AMD, which may affect the generalizability of trial results and subsequent treatment guidelines [40]. Furthermore, the prevalence of AMD varies by race, with higher rates observed in White compared to African Americans and Hispanics [41]. Therefore, public health initiatives should raise awareness about genetic predispositions and modifiable lifestyle risk factors, such as excessive sunlight exposure, alcohol consumption, hypertension, and high body mass index, to help mitigate disease progression in high-risk population. These strategies, combined with ongoing data collection to monitor AMD trends and address disparities, are essential to reducing the public health burden of AMD and improving outcomes for at-risk groups.

Several limitations of our study warrant acknowledgment. First, the data are limited to two age groups (65–84 and >85 years) due to lack of further stratification of age groups within the study cohort. Second, while most Americans over age 65 are covered by Medicare (Nancy Ochieng, 2023) the dataset reported to VEHSS-CDC reflects only about 50% of this demographic. Individuals working beyond age 65 may be underrepresented, as they may still have access to employer-based private insurance plans. Additionally, AMD can also be diagnosed in individuals younger than 65, particularly in the earlier stages, which may lead to an underestimation of the overall prevalence of early and intermediate AMD. Finally, the underestimation of dry AMD based on ICD codes–compared to epidemiological and observational studies – may be due to its relatively mild symptomatology, except in cases of fovea-involving GA, and the asymmetry of macular findings between eyes. In fact, although the presence of late AMD predisposes the fellow eye to progression, only 54% of AMD patients exhibit symmetric findings [42], which could contribute to potential gaps in the data.

Conclusions

In conclusion, these findings provide valuable insights into the current epidemiology of AMD in the USA and offer a critical foundation for developing effective screening strategies and treatment plans, particularly with the advent of new therapeutic options for both wet and dry AMD.

Statement of Ethics

The Medicare data managed by the Centers for Medicare and Medicaid Services is publicly accessible, de-identified, and permissible for use in publications by VEHSS without requiring informed consent. Hence, this study was exempt from ethical review by the Institutional Review Board of Massachusetts General Brigham (ID#: 1614) and complies with the Declaration of Helsinki.

Conflict of Interest Statement

Dr. Rohan B. Singh was a member of the journal’s Editorial Board at the time of submission. The other authors have no conflicts of interest to report.

Funding Sources

This study was not supported by any funding sources.

Author Contributions

Rohan Bir Singh and Isabella Stettler contributed equally to this work and share first authorship and were responsible for study conception and design, data acquisition, analysis, and manuscript drafting. Francesco Romano and Uday Pratap Singh Parmar contributed to data interpretation, literature review, and critical revision of the manuscript. Pier Luigi Surico and Xinyi Ding participated in data collection and figure preparation. Janice J. Kim and Karima K. Rai assisted with statistical analysis and data management. Joan W. Miller and John B. Miller provided senior oversight, project supervision, and critical revision of the manuscript for important intellectual content. All authors reviewed and approved the final manuscript.

Funding Statement

This study was not supported by any funding sources.

Data Availability Statement

The data in this statement are publicly accessible and available through Centers for Disease Control and Prevention (CDC) Vision and Eye Health Surveillance System.