Vision,Medical Cost,and Risk of Falls and Fractures in Geographic Atrophy: An American Academy of Ophthalmology IRIS Registry Analysis

Abstract

Purpose: To evaluate vision loss progression and the association of vision loss with medical care cost, falls, and fractures in patients with geographic atrophy (GA), using the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight). Methods: Patients aged 50 years or older who were diagnosed with GA between April 2016 and December 2021 and had a valid visual acuity (VA) measurement within 90 days before and 1 year after the date of GA diagnosis were evaluated. This post hoc analysis included a homogeneous cohort, a subset of patients for whom the incidence of eye diseases/procedures beyond GA that may affect vision and healthcare resource utilization was controlled. Further evaluation of this cohort described patients with and without subfoveal involvement. Homogeneous claims-linked cohorts were assessed using the IRIS Registry linked to the Komodo Health Research Dataset. Multivariable logistic regression was used to assess the associations of VA with total medical care cost and with incident falls and fractures. Results: A total of 63 616 patients were included. VA was less than 20/100 for 12.8% of patients at diagnosis and 24.3% by year 2. Median time to VA less than 20/40 was 3 years and was faster for those with subfoveal involvement (2.2 years) vs those without (3.8 years). A total of 5827 patients were included in the closed-claims cohorts. By year 2, VA impairment was associated with an increase of more than 26% in total cost of care. Moderate to severe VA impairment was suggestive of an increased risk of falls/fractures. Conclusions: This retrospective study demonstrates an association between vision loss from GA and increased healthcare costs and also suggests a potentially elevated risk of falls and fractures.

Keywords

geographic atrophy visual acuity cost falls

Introduction

Age-related macular degeneration (AMD) is the leading cause of severe vision loss in the developed world, and higher prevalence is observed with advancing age.^1–3 Geographic atrophy (GA) and neovascular AMD (nAMD) are the 2 forms of advanced-stage AMD. GA is characterized by atrophic lesions that enlarge over time, leading to vision loss, reduced quality of life, and decreased independence.^1,4–8

Before 2023, there were no approved treatments for GA, and management consisted largely of supportive care.^4,5 In 2023, pegcetacoplan and avacincaptad pegol were approved by the United States Food and Drug Administration for the treatment of GA based on clinical trials demonstrating significantly decreased lesion growth rate with treatment vs sham.^9,10

GA progression leads to declines in visual function, including loss of visual acuity (VA) and scotoma expansion.⁶ Subfoveal GA is associated with markedly worse VA. While approximately two-thirds of individuals have GA without subfoveal involvement at initial presentation, lesions expand to involve the foveal center point by a median timeframe of 2.5 years from diagnosis.^11–13 Importantly, although VA in individuals without subfoveal GA may be relatively spared, these patients still lose an average of 3.93 Early Treatment Diabetic Retinopathy Study (ETDRS) letters per year and may exhibit deficits in other measures of visual function, including low-luminance VA, dark adaptation, and contrast sensitivity.^14,15

Additionally, because subfoveal involvement of GA is associated with central scotomas, it may cause significant reductions in the size of the usable visual field, even in individuals with preserved VA. This results in difficulty with central vision–dependent tasks, including reading or face recognition.^6,15,16

Visual impairment, with or without subfoveal involvement, affects both vision-related daily functioning and quality of life.⁸ Such outcomes include reductions in quality of life scales compared with individuals without GA; limitations in daily modes of transportation, whether driving, on public transportation, or on foot; difficulty with daily activities such as reading, housework, hobbies, and face recognition; and deleterious effects on emotional well-being, including frustration, loss of independence, and fear of further vision loss.^4,8,17 Individuals with GA are at greater risk of falls and fractures, an often-overlooked element of GA burden that can greatly increase overall morbidity and mortality in older populations.^17,18

Despite the widespread prevalence of GA and the well-studied impact of disease burden on patients’ lives, large studies on the healthcare resource utilization of individuals with GA are limited, including in the US.^8,17,19,20 Understanding healthcare resource utilization in patients with GA is foundational to understanding the true cost of vision loss due to GA, a task that will become more imperative as GA prevalence is expected to increase as the global population ages.³ Using data from the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight) , this study examined the progression of vision loss and the association of vision loss with the cost of medical care and incident falls and fractures in individuals with GA.

Methods

Study Design

This retrospective cohort study utilized the IRIS Registry, the largest electronic health record–based comprehensive eye disease registry in the US. As of July 2024, it includes data from more than 15 000 ophthalmologists and optometrists at more than 2800 practices, and more than 660 million patient visits and 78.9 million patients. These data account for the majority of all practicing US ophthalmologists.

The IRIS Registry is a centralized data repository and reporting tool available for research purposes. Data in the registry are de-identified, and the investigator does not have access to study identifiers; hence, this study does not constitute human subject research, and patient informed consent was not required. This study adheres to the Declaration of Helsinki. This study was reviewed and approved by the Western Institutional Review Board-Copernicus Group Institutional Review Board. IRIS Registry data were linked to claims data from Komodo Health to analyze healthcare resource utilization and associated costs, with data from payers (including medical claims, pharmacy claims, and enrollment data) comprising the primary data source for the Komodo Health database.

The study period was January 1, 2016, to December 31, 2022. The study design is shown in Figure 1. The index date was defined as the date of the first GA diagnosis. The index period was April 1, 2016, to December 31, 2021, and was defined as the period when patients received their GA diagnosis. The pre-index period (look-back period) was at least 90 days before the index date and was used to assess patient demographics, history, and study eligibility based on inclusion and exclusion criteria; this period also provided baseline patient characteristics. The post-index period (follow-up period) was used to assess outcomes of interest and comprised 12 to 24 months of potential follow-up. Follow-up continued from the index date through loss to follow-up (exceeding 1 year without a VA measurement) or end of study.

Figure 1.

Schematic of the study period and description of index periods.

Identification of Patients With GA

Patients aged 50 years or older diagnosed with GA (with or without subfoveal involvement) as documented by an International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) code (H35.3113, H35.3114, H35.3123, H35.3124, H35.3133, H35.3134, H35.3193, H35.3194) during the index period were included. Patients were required to have 1 or more valid VA measurements at the index date (±90 days) and 1 or more valid VA measurements 1 year (±90 days) from the index date for study inclusion. Continued study enrollment required consecutive VA measurements at 365-day (±90 days) intervals for each subsequent study year. Individuals with missing sex information or unknown laterality of GA diagnosis were excluded as a data-quality measure.

The structure and attrition criteria for cohorts are shown in Figure 2. The focus of this study is the homogeneous cohort, defined in a post hoc analysis as a patient subset controlled for the incidence of eye conditions and procedures beyond GA, with the intent to remove vision and healthcare resource utilization changes associated with non-GA vision care. Censoring logic was designed to replicate prospective clinical trial criteria describing the natural history of GA (Proxima A [NCT02479386]/Proxima B [NCT02399072]) using real-world data in a routine clinical practice setting. Patients in this cohort were required to have a baseline VA better than 20/400.

Figure 2.

Patient cohorts and attrition criteria. Cohorts are not mutually exclusive (patients can convert from having no subfoveal involvement to having subfoveal involvement during the study period).

Individuals had no previous diagnosis of nAMD or treatment with antivascular endothelial growth factor (anti-VEGF) during the 90-day baseline period; no diagnosis for other causes of GA (eg, H35.53 unspecified hereditary retinal dystrophy) in the history of observable data; no previous procedures (eg, cataract surgery, laser photocoagulation, vitrectomy surgery, submacular surgery, glaucoma-filtering surgery, corneal transplant) in the history of observable data; no previous intravitreal drug delivery during the 90-day baseline period (eg, corticosteroid injection, antiangiogenic drugs, anticomplement agents, or device implantation); and no other previous exclusionary diagnosis during the 90-day baseline period (eg, clinical trial participants, retinal tear or detachment, aphakia, vitreous hemorrhage, uveitis or vitritis, infectious conjunctivitis, keratitis, scleritis, endophthalmitis, proliferative diabetic retinopathy, uncontrolled glaucoma [defined as intraocular pressure ≥30 mm Hg]).

The homogeneous cohort was split into subcohorts, allowing for the separate evaluation of individuals with and without subfoveal GA at diagnosis. Cohorts were not mutually exclusive, as patients could transition from having no subfoveal involvement to having subfoveal involvement during the study period.

Continued study enrollment criteria for the homogeneous cohort included the addition of censoring events to control for cost and vision changes related to conditions and treatments beyond GA. These included a new diagnosis of nAMD or anti-VEGF treatment, laser surgery or other surgical intervention (eg, cataract surgery, laser photocoagulation, vitrectomy surgery, submacular surgery, glaucoma-filtering surgery, corneal transplantation), other intravitreal drug delivery, enrollment in a clinical trial, or new diagnosis of a serious eye condition (eg, retinal tear or detachment, aphakia, vitreous hemorrhage, uveitis or vitritis, infectious conjunctivitis, keratitis, scleritis, endophthalmitis, proliferative diabetic retinopathy, uncontrolled glaucoma).

For the healthcare resource utilization analysis, claims-linked cohorts were created from the homogeneous cohort using data from Komodo Health. Claims-linked cohorts required patients to have continuity of medical claims enrollment coverage for 90 days or more before and at least 12 months after the index date. Patients were censored at the earliest of previously described end dates of follow-up or discontinued medical claims enrollment.

Study Measures and Outcomes

The study had 4 primary analytical objectives. The first was to describe the baseline demographic and clinical characteristics of the homogeneous cohort and claims-linked cohorts. Clinical characteristics evaluated included GA laterality, ocular comorbidities, and other comorbidities.

The second objective was to describe disease progression for the homogeneous cohort and claims-linked cohorts by measuring change in VA over the follow-up period relative to the index date. VA readings were selected as the value nearest each index and annual outcome date (±90 days). When more than 1 reading was available, the best-documented VA for the patient was selected. For the homogeneous cohort and claims-linked cohorts, a time to vision loss analysis was performed. Eligible patients were required to have a VA better than 20/40 at index; vision loss was defined as a best-documented VA reading of 20/40 or worse in at least 1 eye. The Centers for Disease Control and Prevention Vision and Eye Health Surveillance System case definition for vision loss is the best-corrected VA (BCVA) reading of 20/40 or worse in the better-seeing eye.

The third objective was to characterize healthcare resource utilization and associated costs by year of follow-up and visual impairment categories in the claims-linked cohorts. For these analyses, cost is reported as US dollars per patient per month, while care days (service dates) are reported as the number of care days per 1000 patient-years. Healthcare resource utilization was evaluated for all settings (cost of care and care days): inpatient (cost of care, admissions, readmissions, and length of stay), emergency (cost of care and care days), and outpatient (cost of care and care days).

The final objective was to describe the total cost of care and fall/fracture risk by employing multivariable regression modeling performed on claims-linked cohorts.

Statistical Analysis

For all cohorts, the study size was based on a convenience sampling method to select patients in the IRIS Registry meeting inclusion and exclusion criteria. Objectives and baseline characteristics were analyzed using descriptive statistics. Counts and percentages described categorical variables, while mean, median, and quartiles described continuous variables.

Time to vision loss was analyzed using Kaplan-Meier curves, and reported statistics include survival probability and median survival time. The relationship between VA impairment and cost of care was characterized using a ɣ-distributed log-linked generalized linear model. The dependent variable was defined as the annualized total cost of medical care from index through year 2. Independent variables included VA, mean annual VA change, and the presence of subfoveal involvement. Demographic variables were also included to control for potential confounding.

The relationship between VA impairment and incident fall or fracture was characterized by a multivariable logistic regression model. Incident fall or fracture during the 2-year study period was described among patients without prior fall or fracture during the look-back period. Independent variables included VA, mean annual VA change, and the presence of subfoveal involvement. Demographic variables and conditions that increase fall risk were included to control for potential confounding. Because incident fall and fracture events are not discretely coded in claims, falls and fractures were described by a binary variable defined as 1 or more fall or fracture claim (ie, ≥1 medical claim with a primary diagnosis of fall or fracture) from index through year 2. Data were analyzed by Verana Health using PySpark Version 3.4 (Apache Software Foundation).

Results

Homogeneous Cohorts: Baseline Demographics and Clinical Characteristics

Table 1 shows baseline demographics and clinical characteristics of the homogeneous cohort and subcohorts. Most patients were aged 70 years and older, White, and not Hispanic. Medicare Fee-for-Service was the insurance category for most patients (62.6%). Of patients in the homogeneous cohort, 73.4% had a bilateral diagnosis of GA, and 38.2% had cataract as an ocular comorbidity, while 21.2% had glaucoma or ocular hypertension. Many of the patients (65.4%) were under the care of a retina specialist. Similar trends were seen in patients with and without subfoveal involvement. The median follow-up time for patients was 2 years.

Table 1.

Demographics and Clinical Characteristics at Baseline.

	Number of Patients (%)
Characteristic	Homogeneous Cohort(n = 63 616, 100%)	Without Subfoveal Involvement(n = 28 906, 45.7%)	With Subfoveal Involvement(n = 40 784, 64.1%)
Age ≥70 years^a	56 676 (89.1)	25 424 (88.0)	36 800 (90.2)
Female	40 842 (64.2)	18 631 (64.5)	26 214 (64.3)
White	46 006 (72.3)	20 831 (72.1)	29 810 (73.1)
Insurance category
Medicare Fee-for-Service	39 795 (62.6)	17 905 (61.9)	25 819 (63.3)
Medicare Advantage	6119 (9.6)	2800 (9.7)	4054 (9.9)
Medicaid	771 (1.2)	345 (1.2)	474 (1.2)
Commercial	7525 (11.8)	3483 (12.1)	4711 (11.6)
Other/unknown	9366 (14.7)	4359 (15.1)	5696 (14.0)
No insurance	40 (0.06)	14 (0.05)	30 (0.07)
Bilateral diagnosis	46 684 (73.4)	20 753 (71.8)	30 377 (74.5)
Ocular comorbidities^b
Diabetic retinopathy without diabetic macular edema	393 (0.62)	205 (0.71)	214 (0.52)
Diabetic retinopathy with diabetic macular edema	2237 (3.5)	1064 (3.7)	1344 (3.3)
Central retinal vein occlusion with macular edema	156 (0.25)	74 (0.26)	96 (0.24)
Branch retinal vein occlusion with macular edema	279 (0.44)	133 (0.46)	162 (0.40)
Glaucoma or ocular hypertension	13 508 (21.2)	6291 (21.8)	8413 (20.6)
Cataract	24 306 (38.2)	11 542 (39.9)	14 977 (36.7)
Myopia	820 (1.3)	356 (1.2)	510 (1.3)
Provider subspecialty: retina specialist	41 625 (65.4)	18 104 (62.6)	27 529 (67.5)
Censored year
1 year	12 279 (19.3)	5640 (19.5)	7775 (19.1)
2 years	25 978 (40.8)	11 601 (40.1)	16 411 (40.2)
3 years	13 905 (21.9)	6366 (22.0)	8897 (21.8)
4 years	6566 (10.3)	3016 (10.4)	4378 (10.7)
5 years	3524 (5.5)	1686 (5.8)	2351 (5.8)
6 years	2254 (3.5)	1045 (3.6)	1536 (3.8)

Cohorts are not mutually exclusive (patients can convert from having no subfoveal involvement to having subfoveal involvement during the study period).

To reduce re-identification risk, patients of advanced age were censored to 90 years old as of the run date for the study data source. Patients with censored birth years are aged 83 to 89 in this study.

Ophthalmic comorbidities removed during inclusion or exclusion criteria application: neovascular age-related macular degeneration, retinal tear, vitreous hemorrhage, anterior uveitis, vitritis, posterior uveitis, panuveitis, endophthalmitis, and proliferative diabetic retinopathy.

Homogeneous Cohorts: VA

In the homogeneous cohort, VA was worse than 20/100 (moderately severe or severely impaired vision) for 12.8% of patients at diagnosis, which increased to 24.3% by year 2 (Figure 3A). Furthermore, 25.3% of patients in the homogeneous cohort experienced a 15 or greater ETDRS letter loss in VA over 2 years. VA impairment was worst among cohorts with subfoveal involvement, with 16.4% of patients diagnosed with moderately severe or severely impaired vision, compared with 6.9% of patients without subfoveal involvement. By year 2, these percentages increased to 31.0% and 14.6%, respectively (Figure 3, B and C). Consistent with this finding, 30.1% of patients with subfoveal involvement experienced a 15 or greater ETDRS letter loss over 2 years, compared with 19.9% of patients without subfoveal involvement.

Figure 3.

Visual acuity in the (A) homogeneous, (B) subfoveal involvement, and (C) no subfoveal involvement cohorts.

Decline in VA was also evident in patients with slight or no impairment at baseline. For patients in the homogeneous cohort with baseline VA of 20/40 or better, the probability of maintaining that VA at year 2 was 60.5%, and the median time to VA worse than 20/40 was 3 years. Patients with subfoveal GA lesions and with baseline 20/40 VA or better fared worse over 2 years than those without subfoveal involvement. The probability of maintaining a VA of 20/40 or better at year 2 was 52.8% and 66.5% in patients with and without subfoveal involvement, respectively (Figure 4). Additionally, patients with subfoveal involvement had a shorter median time to VA worse than 20/40 than patients without subfoveal involvement (2.2 vs 3.8 years).

Figure 4.

Probability of maintaining visual acuity of 20/40 or better in the homogeneous cohort and in subcohorts with and without subfoveal involvement.

Claims-Linked Cohorts: Baseline Demographics, Clinical Characteristics, and VA

The claims-linked cohorts (with and without subfoveal involvement) were slightly younger than the homogeneous cohorts (79.2% vs 89.1% of patients aged ≥70 years), and approximately twice as many were commercially insured compared with the homogeneous cohorts (20.9% vs 11.8%). Compared with the homogeneous cohorts, similar proportions of patients in the claims-linked cohorts had bilateral GA at diagnosis as well as ocular comorbidities, and the majority were taken care of by retina specialists (Supplemental Table S1).

Similar trends in VA at 2 years were observed in the homogeneous cohorts and the claims-linked cohorts. VA was worse than 20/100 (moderately severe or severely impaired vision) in 12.8% of patients at diagnosis, which increased to 23.4% by year 2. A 20 or greater ETDRS letter loss in VA over 2 years was observed in 16.6% of patients in the claims-linked cohort. Consistent with the findings in the homogeneous cohort with subfoveal involvement, 20.8% of patients with subfoveal involvement experienced a 20 or more letter loss over 2 years compared with 13.3% of patients without subfoveal involvement. The probability of having a VA of 20/40 or better at year 2 was 64.5%, 70.2%, and 57.0%, while the median time to VA worse than 20/40 was 3.7, 4.7, and 2.5 years for the homogeneous, without subfoveal involvement, and with subfoveal involvement claims-linked cohorts, respectively.

Claims-Linked Cohorts: Healthcare Resource Utilization

Mean total cost at year 1 of follow-up was generally higher in patients with moderate (20/50 to 20/100) to severe (worse than 20/200) VA impairment compared with those with no or mild VA impairment. Similar observations were seen in patients with subfoveal involvement. No clear trend was observed in those without subfoveal involvement. At year 2, the mean total cost was similar across all VA impairment categories. When the total cost based on VA change was evaluated, patients in the homogeneous claims-linked cohort who had a decline in VA (≥20 ETDRS letter loss) had a higher total cost ($1,058) than those with improvement in VA (≥15 ETDRS letter gain; $645) at year 2. A similar trend was observed for those with and without subfoveal involvement; however, this was not observed at year 1. No clear trend was observed between the VA categories or VA change in cost or length of stay for acute inpatient, cost of emergency department, or outpatient healthcare resource utilization cost at year 1 or year 2.

Multivariable log-link generalized linear model with ɣ distribution analyses identified variables that were significantly associated with high healthcare costs in year 2. In models through year 2, mild (20/25 to 20/40) and moderate (20/50 to 20/100) VA impairment were associated with a more than 30% increase in total cost of care; moderately severe VA impairment (20/125 to 20/200) was associated with a 26% increase in cost; and severe VA impairment (worse than 20/200) was associated with a 43% increase in cost (mild: adjusted cost ratio, 1.31 [95% CI, 0.98-1.74, P = .061); moderate: adjusted cost ratio, 1.37 [95% CI, 1.00-1.86], P = .045; moderately severe: adjusted cost ratio, 1.26 [95% CI, 0.86-1.83], P = .237; severe: adjusted cost ratio, 1.43 [95% CI, 0.88-2.34], P = .127).

Subfoveal involvement did not affect cost after controlling for VA, age, insurance category, race, ethnicity, and sex (Figure 5A). Hispanic ethnicity was also correlated with increased cost of care (adjusted cost ratio, 1.79 [95% CI, 1.24-2.68], P = .003), as was Black or African American race (adjusted cost ratio, 1.76 [95% CI, 1.01-3.45], P = .065) after controlling for VA, age, insurance category, lesion location, and sex.

Figure 5.

(A) Annualized cost of care^a and (B) risk of falls and fractures^b 2-year model.

Using adjusted models, moderate to severe VA impairment (20/50 to worse than 20/200) was suggestive of an increased risk of falls and fractures compared with no impairment (20/20 or better); however, this did not reach statistical significance (moderate: odds ratio [OR], 1.33 [95% CI, 0.625-3.185], P = .484; moderately severe: OR, 1.39 [95% CI, 0.627-3.445], P = .440; severe: OR, 1.19 [95% CI, 0.365, 3.912], P = .772; Figure 5B). Using the unadjusted models, mild (20/25 to 20/40), moderate (20/50 to 20/100), and severe (worse than 20/200) VA impairment were each associated with an increased risk of falls and fractures (mild: OR, 2.1 [95% CI, 1.18-4.20], P = .019; moderate: OR, 2.4 [95% CI, 1.31-4.78], P = .007; moderately severe: OR, 1.24 [95% CI, 0.56-2.79], P = .604; severe: OR, 2.4 [95% CI, 1.22-4.99], P = .014; Figure 5B). Subfoveal involvement did not independently affect fall risk after controlling for age, sex, race, ethnicity, and fall risk factors.

Conclusions

This retrospective analysis using the IRIS Registry investigated GA progression and associated vision loss, as well as the impact on total cost of care and fall and fracture risk. Many patients were aged 70 years or older and had subfoveal involvement of the GA lesion. Multiple analyses revealed loss of visual function over time, which was consistent with the progressive nature of GA. Approximately 25% of patients had a visual decline (≥15 ETDRS letter equivalents) through year 2. In patients with slight or no impairment at diagnosis, the median time to vision loss (defined as VA worse than 20/40) was 3 years, and the decline was faster in patients with subfoveal involvement than in those without subfoveal involvement.

Such VA declines in patients with GA are consistent with previously published data. Early studies by Sunness et al²¹ demonstrated that in patients with GA, eyes with a baseline VA of 20/50 or better had a high risk of losing 3 lines or more (≥15 ETDRS letter loss) at 2 years. Additionally, in a retrospective cohort study (N=1901) of a multicenter electronic health record database, mean VA in the worse-seeing eye of patients with bilateral GA decreased over 2 years and continued to decline over 60 months, with a 10.9-letter loss. Over this same timeframe, the better-seeing eye exhibited a steeper trajectory of VA loss, with a 22.6-letter loss at month 60.²²

Other studies have noted differences in visual impairment between individuals with and without subfoveal involvement. Colijn et al¹¹ reported that eyes with subfoveal involvement had severe visual impairment more often and were more likely to have worse BCVA than eyes without subfoveal involvement. Leng et al modeled the annualized loss of ETDRS letters in GA patients with and without subfoveal involvement. They found that although the rates of decline were comparable (3.93 vs 2.50 letters for without vs with subfoveal involvement, respectively), 57.1% of patients without subfoveal involvement at baseline had a VA better than 20/63, compared with 28% of patients with subfoveal involvement.¹⁴

Although a decline in VA in patients with GA was observed, other studies found no significant difference in VA decline between subfoveal and extrafoveal GA lesions.^23,24 Importantly, other visual function measures, including low-luminance VA, microperimetry, and reading speed, together with BCVA, may provide a more complete picture of visual decline than BCVA or VA alone.⁵

Using the claims-linked cohorts, the multivariable modeling suggests that moderate to severe VA impairment may be associated with increased cost of care after controlling for race, ethnicity, age, sex, and insurance. Statistical significance was observed using the adjusted cost ratio for moderate visual impairment; however, a trend was observed for severe visual impairment. The median time to VA worse than 20/40 was 2.5 years for the claims-linked cohort with subfoveal involvement, which was faster than that of patients without subfoveal involvement; however, the 2-year model did not suggest that subfoveal involvement affected cost after controlling for VA, age, insurance category, race, ethnicity, and sex. This may be attributable to the model’s homogeneous cohort.

While data on visual function were not reported by Kim et al, they found that US patients with GA with or without subfoveal involvement required more outpatient visits and inpatient admissions. They also incurred greater patient- and payer-related costs compared with patients with early or intermediate AMD.²⁰ Furthermore, Kim et al²⁰ did not specify a homogeneous cohort that removed healthcare resource utilization changes that may be associated with non-GA vision care. In the United Kingdom, Chakravarthy et al¹⁹ demonstrated higher ophthalmic healthcare resource utilization over 2 years for patients with GA compared with those with early or intermediate AMD, although they noted lower healthcare resource utilization compared with patients with choroidal neovascularization in 1 eye and GA in the other. The impact of VA impairment, however, was not evaluated.

Similar to total cost, data from the 2-year model suggest that worse visual impairment may be associated with an increased fall risk after controlling for leading risk factors such as race, ethnicity, age, sex, and insurance. Both the adjusted and raw ORs were greater than 1 for moderate to severe VA; however, the adjusted model did not reach statistical significance. This finding adds further insight to the limited published data. In a study of the Medicare population, Anastasopoulos et al²⁵ found that patients with a code for atrophic AMD were 11% more likely to experience an incident hip fracture than patients without a code for AMD (OR, 1.11 [95% CI, 1.06-1.16], P < .001). In a cross-sectional study in which historical healthcare resource utilization data were collected from patients with GA (n = 137), falls were cited as the main reason for inpatient or outpatient admission.⁸ Caregivers have reported injuries in patients with GA, and eye care professionals suspect that some injuries may be associated with GA; however, it is difficult to confirm the causality because patients with GA tend to be older and have comorbidities.²⁶ Because literature examining healthcare resource utilization and fall risk in GA patients is scarce, additional studies would further elucidate the direct and indirect costs incurred by this patient population.

Limitations of this study included the fact that claims-linked cohorts were slightly younger, more diverse, and more likely to be commercially insured. They also represent a population engaged in routine care, reflecting possible differences in healthcare resource utilization outcomes compared with the general population of patients with GA. Moreover, while the ICD-10-CM codes were used to identify patients with GA from the registry, it is acknowledged that the codes may not always be used in clinical practice. To analyze healthcare resource utilization, claims, and cost, data sourced from Komodo rely on imputation, but outputs derived by Komodo are consistent with those offered to the life science industry for standard use in health economics and outcomes research with real-world data. While this analysis included a large database, the sample size was reduced with follow-up.

Additionally, the design is a retrospective cohort study and is associational and descriptive. This retrospective, observational study relies on electronic health record documentation from multiple health systems and is subject to inherent limitations. These include provider variation in VA measurement approaches and classifications due to a lack of standardized protocols. This analysis is based on real-world data and reflects a real-world presentation of patients with GA from 2016 to 2022. Finally, diagnosis coding determined lesion location. Changes in clinical treatment patterns outside the study period are not reflected here.

In conclusion, the data from this large dataset provide further evidence of the progressive nature of GA and its effect on VA. Patients with subfoveal involvement had a shorter time to vision loss than patients without subfoveal involvement, and the data suggest worse categories of VA impairment after 2 years of follow-up may be associated with increased cost of care and heightened fall risk. These costs and effects on quality of life associated with progressive vision loss in GA should be considered when weighing the benefits and risks of interventions for GA.

Supplemental Material

sj-docx-1-vrd-10.1177_24741264261448421 – Supplemental material for Vision, Medical Cost, and Risk of Falls and Fractures in Geographic Atrophy: An American Academy of Ophthalmology IRIS Registry Analysis

Supplemental material, sj-docx-1-vrd-10.1177_24741264261448421 for Vision, Medical Cost, and Risk of Falls and Fractures in Geographic Atrophy: An American Academy of Ophthalmology IRIS Registry Analysis by Durga Borkar, Kristin Arkin-Leydig, Jennifer Toth, Nikhil Patel, Dina Abulon and Aki Shiozawa in Journal of VitreoRetinal Diseases

Footnotes

Acknowledgements

Medical writing and editorial support were provided by Lesley Wassef-Birosik, PhD, CMPP, of IMPRINT Science, and funded by Astellas Pharma Inc.

Authors’ Note

Portions of these data were presented as a poster at Academy of Managed Care Pharmacy Nexus 2024 (October 14-17, 2024, Las Vegas, NV, USA) and as a presentation at the Asia-Pacific Vitreo-retina Society 2024 Congress (November 22-24, 2024, Singapore). The authors had full control of the content and made the final decision on all aspects of publication.

ORCID iDs

Durga Borkar

Aki Shiozawa

Ethical Considerations

This study adheres to the Declaration of Helsinki. This study was reviewed and approved by the Western Institutional Review Board-Copernicus Group (WCG IRB), Puyallup, WA, USA.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Astellas Pharma US. Verana Health was a contracted consultant to Iveric Bio, an Astellas company.

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Borkar is a consultant for 4DMT, AbbVie/Allergan, Alimera Sciences, Apellis Pharmaceuticals, Astellas Pharma, EyePoint Pharmaceuticals, Glaukos, Genentech, ONL Therapeutics, Regeneron, and Verana Health, and she is a speaker for Astellas. Dr. Toth is an employee of Verana Health. Kristin Arkin-Leydig, at the time of the analyses, was an employee of Verana Health and is currently at Amgen Inc, Thousand Oaks, CA, USA. Dr. Patel, at the time of the analyses, was an employee of Astellas Pharma and is currently at 4DMT, Morris Plains, NJ, USA. Dr. Abulon and Dr. Shiozawa are employees of Astellas Pharma.

Data Availability Statement

Researchers may request access to the data used to support this article by contacting the corresponding author.

Consent to Participate

Data in the IRIS Registry are de-identified, and the investigator does not have access to study identifiers; hence, this does not constitute human subject research, and patient informed consent was not required.

Supplemental Material

Supplemental material is available online with this article.

References

Fleckenstein

Keenan

TDL

Guymer

, et al. Age-related macular degeneration. Nat Rev Dis Primers. 2021;7:31.

Rudnicka

Jarrar

Wormald

Cook

Fletcher

Owen

CG.

Age and gender variations in age-related macular degeneration prevalence in populations of European ancestry: a meta-analysis. Ophthalmology. 2012;119:571-580.

Wong

, et al. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. Lancet Glob Health. 2014;2:e106-e116.

Carlton

Barnes

Haywood

Patient perspectives in geographic atrophy (GA): exploratory qualitative research to understand the impact of GA for patients and their families. Br Ir Orthopt J. 2019;15:133-141.

Fleckenstein

Mitchell

Freund

, et al. The progression of geographic atrophy secondary to age-related macular degeneration. Ophthalmology. 2018;125:369-390.

Heier

Pieramici

Chakravarthy

, et al. Visual function decline resulting from geographic atrophy: results from the Chroma and Spectri phase 3 trials. Ophthalmol Retina. 2020;4:673-688.

Kaszubski

Ben Ami

Saade

Smith

RT.

Geographic atrophy and choroidal neovascularization in the same eye: a review. Ophthalmic Res. 2016;55:185-193.

Patel

Ziemssen

, et al. Burden of illness in geographic atrophy: a study of vision-related quality of life and health care resource use. Clin Ophthalmol. 2020;14:15-28.

Syfovre. Prescribing Information. Apellis Pharmaceuticals Inc; 2023.

10.

Izervay. Prescribing Information. Astellas Pharma US, Inc; 2025.

11.

Colijn

Liefers

Joachim

, et al. Enlargement of geographic atrophy from first diagnosis to end of life. JAMA Ophthalmol. 2021;139:743-750.

12.

Keenan

Agrón

Domalpally

, et al. Progression of geographic atrophy in age-related macular degeneration: AREDS2 report number 16. Ophthalmology. 2018;125:1913-1928.

13.

Lindblad

Lloyd

Clemons

, et al. Change in area of geographic atrophy in the Age-Related Eye Disease Study: AREDS report number 26. Arch Ophthalmol. 2009;127:1168-1174.

14.

Leng

Schwartz

Nimke

, et al. Dry age-related macular degeneration: distribution of visual acuity and progression risk in a large registry. Ophthalmol Ther. 2023;12:325-340.

15.

Sunness

Rubin

Applegate

, et al. Visual function abnormalities and prognosis in eyes with age-related geographic atrophy of the macula and good visual acuity. Ophthalmology. 1997;104:1677-1691.

16.

Sunness

Rubin

Zuckerbrod

Applegate

CA.

Foveal-sparing scotomas in advanced dry age-related macular degeneration. J Vis Impair Blind. 2008;102:600-610.

17.

Sarda

Heyes

Bektas

, et al. Humanistic and economic burden of geographic atrophy: a systematic literature review. Clin Ophthalmol. 2021;15:4629-4644.

18.

Mattig

Falls in the elderly: a major public health challenge with some encouraging developments. A mini review. J Gerontology Geriatrics. 2020;68:180-183.

19.

Chakravarthy

Bailey

Scanlon

, et al. Direct ophthalmic healthcare resource use among patients with geographic atrophy in a large cohort from the United Kingdom. Ophthalmol Retina. 2019;3:920-926.

20.

Kim

Devine

Campbell

, et al. Healthcare resource utilization and costs in patients with geographic atrophy secondary to age-related macular degeneration. Clin Ophthalmol. 2021;15:2643-2651.

21.

Sunness

Gonzalez-Baron

Applegate

, et al. Enlargement of atrophy and visual acuity loss in the geographic atrophy form of age-related macular degeneration. Ophthalmology. 1999;106:1768-1779.

22.

Chakravarthy

Bailey

Johnston

, et al. Characterizing disease burden and progression of geographic atrophy secondary to age-related macular degeneration. Ophthalmology. 2018;125:842-849.

23.

Schmitz-Valckenberg

Sahel

Danis

, et al. Natural history of geographic atrophy progression secondary to age-related macular degeneration (geographic atrophy progression study). Ophthalmology. 2016;123:361-368.

24.

Shen

Sun

Ahluwalia

, et al. Relationship of topographic distribution of geographic atrophy to visual acuity in nonexudative age-related macular degeneration. Ophthalmol Retina. 2021;5:761-774.

25.

Anastasopoulos

Coleman

AL.

Age-related macular degeneration is associated with an increased risk of hip fractures in the Medicare database. Am J Ophthalmol. 2006;142:1081-1083.

26.

Singh

Patel

Nielsen

, et al. Patient-, caregiver-, and eye care professional-reported burden of geographic atrophy secondary to age-related macular degeneration. Am J Ophthalmic Clin Trials. 2019;2:1-6.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB

0.00 MB