Annual prevalence of geographic atrophy and wet age-related macular degeneration among Medicare Advantage enrollees in a US health plan

Abstract

BACKGROUND:

Geographic atrophy (GA) is an advanced form of dry age-related macular degeneration (AMD) that can lead to visual impairment. Published studies estimate approximately 1 million people in the United States have GA in at least 1 eye. There is a lack of real-world evidence from the US payer perspective on the prevalence of AMD and GA among Medicare Advantage prescription drug (MAPD) plan enrollees.

OBJECTIVE:

To estimate the annual prevalence of GA, wet AMD, and co-occurring GA and wet AMD among MAPD plan enrollees from 2018 through 2021.

METHODS:

This retrospective, cross-sectional study estimated the prevalence of GA and AMD based on Medicare Advantage enrollee claims data. Individuals aged 65 years and older who had continuous enrollment throughout each calendar year constituted the denominator for each annual prevalence calculation. Enrollees with at least 1 medical claim with a diagnosis code for GA or wet AMD during each year were identified to estimate annual prevalence for that respective calendar year.

RESULTS:

The total number of patients in the denominator was 2,175,803 (2018); 2,445,163 (2019); 2,680,322 (2020); and 2,905,366 (2021). The annual prevalence of GA was 0.56% (2018), 0.55% (2019), 0.48% (2020), and 0.51% (2021). The annual prevalence of wet AMD was 1.2% (2018), 1.3% (2019), 1.2% (2020), and 1.3% (2021). The prevalence of GA was highest among individuals classified as White race (annual range 0.61% to 0.71%) and among patients with GA aged 75 years and older (range 0.95% to 1.11%). The proportion of patients with GA with co-occurring wet AMD was 25.6% to 28.0%. The annual prevalence of advanced AMD (GA or wet AMD) was 1.6% to 1.7%.

CONCLUSIONS:

In the Medicare populations, the prevalence of GA was greatest among patients aged 75 years and older and individuals classified as White race. A substantial proportion of individuals with GA had evidence of co-occurring wet AMD. MAPD plans should evaluate how their membership may be impacted by the recently approved medications for the treatment of GA.

Plain language summary

Geographic atrophy (GA) and wet age-related macular degeneration (AMD) are eye disorders. Both GA and wet AMD can lead to vision loss or blindness. These impact a patient’s quality of life. We assessed GA and wet AMD prevalence among Medicare Advantage prescription drug (MAPD) plan members. The prevalence of GA was found to be 0.48% to 0.56%. The prevalence of wet AMD was found to be 1.2% to 1.3%. Approximately 25% of patients with GA also had wet AMD.

Implications for managed care pharmacy

Overall, we found 1.7% of the MAPD plan population had advanced AMD. A quarter of patients with GA in this study had co-occurring wet AMD. Understanding the prevalence of GA will help inform policy to guide prescribing practices and allow for more accurate assessments of real-world clinical and humanistic impacts associated with the recently approved treatments for GA.

Age-related macular degeneration (AMD) is a degenerative disorder of the macula and the most common cause of irreversible vision loss globally. AMD is classified as early, intermediate, and advanced non-neovascular (dry) AMD or advanced neovascular (wet) AMD.^1-2 In wet AMD, vascular endothelial growth factor (VEGF)–driven abnormal angiogenesis results in new blood vessel growth under the macula, which exude fluid or blood, leading to vision loss. The most common form of AMD is the dry, nonexudative type. Dry AMD is characterized by progressive degeneration of the macula and in its advanced form leads to atrophic scars and, ultimately, geographic atrophy (GA).

More than 1 million people are estimated to be diagnosed with GA in at least 1 eye in the United States.³ GA lesions may start outside of the central point of the macula known as the fovea (eg, without subfoveal involvement [SFI]), and as the disease progresses, lesions may spread to the fovea (eg, with SFI), leading to significant and irreversible vision loss.^4-6 GA is responsible for 20% of legal blindness in North America.⁷

Historically, wet AMD has been treated with anti-VEGFs, but no treatments were approved for GA. Two new therapies recently available in the United States necessitate further understanding GA epidemiology among seniors. Pegcetacoplan was approved by the US Food and Drug Administration (FDA) in February 2023.⁸ In August 2023, the FDA approved avacincaptad pegol.⁹ These drugs inhibit the complement cascade and are given via intravitreal injection, which slows progression of GA lesion growth.

Previous studies of AMD in older US adults have analyzed data from the Eye Diseases Prevalence Research Group, the National Health and Nutrition Examination Survey III, the American Community Survey, and Medicare fee-for-service (FFS).^3,10-12 These studies calculated point estimates of prevalence using systematic reviews and meta-analysis, included patients aged from 40 to 90 years, or included geographic regions outside the United States. Only 2 of these studies used Medicare FFS data.^10,12 Most published studies reported prevalence for any AMD, dry AMD, and wet AMD. None of these studies focused on GA prevalence among Medicare Advantage prescription drug (MAPD) plan enrollees in the United States. To address this gap, our study examined the prevalence of GA, wet AMD, co-occurring GA and wet AMD, and other stages of AMD among enrollees of an MAPD plan.

Methods

DATA SOURCE

This retrospective study used the Humana Research Database, which contains health care administrative claims for Humana MAPD plan enrollees. The database includes records from 2007 to the present. As of 2024, Humana is the second largest MAPD provider in the United States, covering approximately 5.9 million individual and group MAPD enrollees. Member enrollment data and medical claims information for MAPD patients were used to capture GA, wet AMD, and co-occurring GA and wet AMD diagnoses originating from physician visits, other outpatient encounters, emergency department visits, and inpatient hospitalizations.

This study did not constitute human subjects research and did not require institutional review board oversight, as determined by the Humana Healthcare Research Human Subject Protection Office, which uses HHS regulations 45 CFR 46 and the Office for Human Research Protections guidance on Coded Private Information or Specimens Use in Research, Guidance (2008).

STUDY POPULATION AND DESIGN

This study used a cross-sectional design. The annual prevalence of AMD was estimated by identifying patients with at least 1 claim with an AMD diagnosis (International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] codes: early-stage AMD [H35.31x1], intermediate-stage AMD [H35.31x2], GA [H35.31x3 to H35.31x4], and wet AMD [H35.32x0 to H35.32x1, H35.32x3]) in one or both eyes or in an unspecified eye during calendar years 2018-2021. Each calendar year was assessed independently. To be eligible for inclusion, all patients were required to be aged 65 years or older as of January 1 and have continuous enrollment during the calendar year being assessed. Patients with diagnoses for multiple stages of dry AMD were assigned only the most advanced stage during the calendar year. Similarly, bilateral GA diagnosis was prioritized over unilateral GA. Because it is possible that individuals may have a diagnosis for both dry AMD and wet AMD during the assessment year, wet AMD diagnosis was assessed independently. Individuals with both dry AMD and wet AMD were included in prevalence estimates for both conditions.

STUDY OUTCOMES

The main study outcome was annual prevalence of AMD. Stages of AMD included early, intermediate, and advanced (GA and/or wet AMD). Prevalence was also calculated separately for GA, wet AMD, and co-occurring GA and wet AMD.

Denominator for Prevalence Calculation. MAPD enrollees aged 65 years and older as of January 1 with continuous enrollment during each assessment year were identified independently.

Numerator for Prevalence Calculation. Patients with at least 1 medical claim with a diagnosis code for a specific stage of AMD in any diagnosis position were identified independently for each calendar year.

STATISTICAL ANALYSIS

Annual prevalence was calculated by dividing the numerator for each stage of dry and wet AMD by the denominator for the assessment year and reported as percent of all enrollees. Prevalence was calculated separately for each of the AMD stages by calendar year, and subjects with multiple codes in 1 year were classified to the most severe stage. The prevalence of GA by race (White/Black/Other/Unknown) and age as well as proportion of patients with GA with co-occurring wet AMD were calculated for each assessment year.

To assess the impact of a 2-year ascertainment period compared with the 1-year period used for the primary analysis, we performed a sensitivity analysis that included an additional 12 months before the calendar year of assessment. All patients in the numerator and denominator were required to have 24 months of continuous enrollment (eg, for annual prevalence calculation for year 2018, continuous enrollment during 2017-2018 was required and then claims with AMD diagnosis were identified during 2017-2018).

Results

The number of MAPD enrollees identified for the denominator in each assessment year was as follows: 2,175,803 (2018); 2,445,163 (2019); 2,680,322 (2020); and 2,905,366 (2021) (Table 1). The annual prevalence of early-stage, dry AMD was 2.6% (2018), 2.7% (2019), 2.4% (2020), and 2.6% (2021). The annual prevalence of advanced AMD was 1.7% (2018), 1.7% (2019), 1.6% (2020), and 1.6% (2021).

TABLE 1.

Annual Prevalence of Dry AMD, Wet AMD, and GA

Annual prevalence measure	Year 2018	Year 2019	Year 2020	Year 2021
Patients with continuous enrollment during the calendar year, N	2,175,803	2,445,163	2,680,322	2,905,366
Overall AMD (early AMD/intermediate AMD/GA/wet AMD) (H35.31x1-H35.31x4, H35.32x0-H35.32x1), %	5.32	5.48	4.90	5.24
Early-stage AMD (H35.31x1), %	2.60	2.71	2.37	2.55
Intermediate-stage AMD (H35.31x2), %	1.57	1.63	1.45	1.60
Advanced AMD (GA [H35.31x3-H35.31x4] and/or wet AMD [H35.32x0-H35.32x1, H35.32x3]), %	1.66	1.67	1.58	1.64
Overall GA (H35.31x3-H35.31x4), %	0.56	0.55	0.48	0.51
Overall wet AMD (H35.32x0-H35.32x1, H35.32x3), %	1.24	1.26	1.23	1.27
Co-occurring GA (H35.31x3-H35.31x4) and wet AMD (H35.32x0-H35.32x1, H35.32x3), %	0.14	0.15	0.13	0.14
Proportion of patients with GA (H35.31x3-H35.31x4) with wet AMD (H35.32x0-H35.32x1, H35.32x3), %	25.6	26.6	28.0	28.0

Each patient with dry AMD in a given year was counted in only 1 dry AMD category corresponding to the most severe diagnosis identified for that patient in that year.

Each patient with wet AMD in a given year was counted in only 1 wet AMD category corresponding to the most severe diagnosis identified for that patient in that year.

Patients with both dry and wet AMD were counted in both a dry AMD category and a wet AMD category according to the logic above.

Note that within the International Classification of Diseases, Tenth Revision, Clinical Modification codes for AMD, sixth position indicates laterality and seventh position indicates staging.

AMD = age-related macular degeneration; GA = geographic atrophy.

The annual prevalence for overall GA was 0.56% (2018), 0.55% (2019), 0.48% (2020), and 0.51% (2021). From 2018 to 2021, the annual prevalence of co-occurring GA and wet AMD ranged from 0.13% to 0.15%. The proportion of patients with GA with co-occurring wet AMD ranged from 25.6% to 28.0% from 2018 to 2021 (Table 1). The annual prevalence for GA without SFI was 0.21% to 0.25% across the assessment years (Supplementary Table 1 ^{(168.2KB, pdf)} , available in online article). The annual prevalence for GA with SFI was 0.27% to 0.31% during the same period. The annual prevalence of GA also was estimated at the level of eyes, with bilateral GA prevalence between 0.24% and 0.29% and unilateral GA between 0.23% and 0.27% from 2018 to 2021.

The annual prevalence of GA was highest among individuals classified as White race: 0.67% (2018), 0.66% (2019), 0.61% (2020), and 0.71% (2021) (Supplementary Table 1 ^{(168.2KB, pdf)} ). Among the individuals classified as Black race, annual prevalence ranged from 0.09% to 0.11%. The annual prevalence of GA was calculated for 2 age categories; it was higher among individuals aged 75 years and older (0.95% to 1.11%) compared with people aged younger than 75 years (0.15% to 0.18%).

The annual prevalence of wet AMD was 1.2% (2018), 1.3% (2019), 1.2% (2020), and 1.3% (2021) (Figure 1). Approximately 46% to 48% of patients with wet AMD were diagnosed with unspecified dry AMD (coded as ICD-10-CM: [H35.31x0], Supplementary Table 1 ^{(168.2KB, pdf)} ), and approximately 41% to 43% of patients with wet AMD had a diagnosis of early/intermediate-stage dry AMD.

FIGURE 1.

Annual Prevalence of AMD

In the sensitivity analyses, including an additional 12 months for measuring prevalence, the annual prevalence of GA increased: 0.85% (2018), 0.83% (2019), 0.77% (2020), and 0.74% (2021) (Supplementary Table 2 ^{(168.2KB, pdf)} ). The proportion of patients with GA with co-occurring wet AMD ranged from 27.7% to 29.2% across the 4 years. The annual prevalence of advanced AMD increased relative to the 12-month prevalence: 2.1% (2018), 2.2% (2019), 2.1% (2020), and 2.1% (2021).

Discussion

This is the first study that has calculated the prevalence of GA and other AMD stages using MAPD plan enrollee data. In our study, the annual prevalence of GA ranged between 0.48% and 0.56% and was relatively stable from 2018 to 2021. Sensitivity analyses that increased the ascertainment window to 24 months resulted in higher prevalence estimates (0.74%-0.85%), suggesting longer time frames may be required to identify cases of GA in claims data. We observed numerically lower prevalence of GA during 2020 relative to the other calendar years assessed, which may reflect lower utilization of certain health care services and therefore lower coding intensity during the early phases of the COVID-19 pandemic.

Similar to other published studies, we found the prevalence of GA was higher among individuals aged 75 years and older relative to younger individuals. For example, Friedman et al estimated prevalence of GA for various patient age groups and found prevalence ranged from 0.48% to 0.90% for those aged 65 to 74 years, whereas for patients aged 75 years and older, the observed range was 1.78% to 6.89%.¹³ Their estimates were based on a meta-analysis of population-based studies in the United States, Australia, and Europe. In the Baltimore Eye Survey included in that analysis, the authors reported a GA prevalence of 0.5% overall, similar to our overall prevalence estimates.¹⁴ Similarly, a systematic review and Bayesian meta-regression of population-based studies estimated the prevalence of GA to range from 0.16% at age 60 years to 2.91% at age 80 years.¹⁵ Additionally, similar to our results, Klein et al reported the prevalence of GA to be 1.4% among people aged 60 years and older.¹⁶

We observed a greater prevalence of GA among individuals classified as White compared with those classified as Black. Similarly, Rein et al estimated the prevalence of advanced AMD among non-Hispanic White race (1.03%) to be higher than that among Black race (0.65%), and Wong et al also reported that the prevalence of advanced AMD was higher in Europeans than in Africans.^3,17 Further supporting our findings, Vanderbeek et al observed significantly lower rates of dry and wet AMD at age 60 relative to age 80 and in Black individuals relative to White patients.^3,18

We found the overall prevalence of GA and co-occurring wet AMD ranged between 0.13% and 0.15%. Similar to prior reports in the literature, we found the proportion of patients with GA with co-occurring wet AMD ranged from 25.6% to 28.0% during 2018-2021.¹⁹

In our study, the prevalence of overall wet AMD ranged between 1.2% and 1.3%. A recently published study by Ehrlich et al reported wet AMD prevalence during 2018-2020 to be between 2.4% and 2.6%.¹² In contrast to our study, this study was conducted using Medicare FFS data and calculated the prevalence of overall AMD, dry AMD, and wet AMD. Additionally, the patient population in the study by Ehrlich et al included patients aged 68 years or older, whereas our study included patients aged 65 years or older. Furthermore, Ehrlich et al required patients to have a total of 3 years of continuous enrollment compared with 1 year in our study. Ehrlich et al did not publish diagnosis codes used for identifying wet AMD, making direct comparisons to our study difficult. All these factors could have resulted in the difference in our findings. In our sensitivity analyses, by requiring continuous enrollment for 2 years for inclusion in the numerator and denominator, the annual prevalence of wet AMD slightly increased and ranged from 1.5% to 1.6%, which was still lower than the findings of Ehrlich et al.

We also observed that the annual prevalence of advanced AMD, including GA and wet AMD, among patients enrolled in MAPD ranged between 1.6% and 1.7% across 2018 to 2021. Rein et al reported the prevalence of advanced AMD for the year 2019 to be 0.94% among patients aged 40 years and older using 4 different data sources and Bayesian meta-regression methods.³ Also similar to our findings, Klein et al reported the prevalence of advanced AMD to be 2.2% among patients aged 60 years and older.¹⁶

In contrast to our results, Rein et al reported that the prevalence of advanced AMD increased with age, from 0.02% among those aged 40 to 44 years to 11.39% among people aged 85 years or older.³ Additionally, Friedman et al had estimated the prevalence of advanced AMD among people aged 65 years or older to be 0.9% to 11.8%. However, these rates were estimated based on meta-analysis of results of population-based studies and applied to 2000 US Census data for projecting to year 2020.¹³

Although various sources state that GA and wet AMD occur at similar rates,²⁰ our study found the proportion of patients with GA to be less than half of those with a diagnosis of wet AMD. We observed that 46% to 48% of patients with wet AMD had a diagnosis of unspecified dry AMD (ICD-10-CM: H35.3110-H35.3190). Undercoding may have contributed to this result; during the study time period, there were no approved treatment options for GA, potentially influencing providers’ coding practices.

Patients with coexisting wet AMD and GA are less likely to have a primary diagnosis of GA. A potential explanation is that prior to the approval of complement inhibitors for GA treatment, there were no therapeutic options, and clinicians may not have accurately documented the code for GA in the diagnosis portion of the medical record.^21-23 With the availability of 2 FDA-approved treatments for GA, the accuracy of GA coding within claims data may improve in the near future.

LIMITATIONS

A key limitation of this study is the use of administrative claims data to estimate the prevalence of GA. Only conditions coded on medical claims associated with health care encounters could be identified, and coding practices may have been influenced by a variety of factors including reimbursement incentives and treatment options. Also, we required only 1 claim with diagnosis of GA instead of at least 2 claims separated by 30 days. Because of the lack of GA treatments during the study period, we believe most patients with dry AMD would visit a provider once per year for eye monitoring. Unspecified dry AMD (H35.31x0) was not counted because severity of disease is unknown and unspecified coding may be used by optometry before referral to a specialist. Wet (exudative) AMD with inactive choroidal neovascularization (H35.32x2) is not considered active wet AMD by specialists, so this was excluded to avoid overestimation of active wet AMD prevalence. Additionally, this study estimated prevalence during a period that included 2020 and 2021, and these estimates may be influenced by broader trends in health care utilization and potentially deferred care during the early phases of the COVID-19 pandemic.

Conclusions

GA previously was not a treatable condition. FDA approval of therapies to treat this disease necessitates accurate epidemiology data to allow MAPD plans to evaluate the potential impact of these therapies on enrollees. The lower prevalence of GA observed in claims data relative to other studies warrants careful consideration; coding practices likely will evolve with new GA therapies on the market. Consistent with other studies, GA was more prevalent among individuals aged 75 years and older and among individuals classified as White race. Additionally, more than 25% of patients with GA had a co-occurring diagnosis of wet AMD. These data will enable MAPD plans to better understand how this devastating disease is impacting their patients and evaluate trends as coding practices improve with new therapies on the market.

Funding Statement

This study was funded by Apellis Pharmaceuticals (Waltham, MA, USA).