Association of ocular antihypertensive medications and the development and progression of age-related macular degeneration in a U.S. insurance claims database

Abstract

Purpose/Aim:

To assess whether ocular antihypertensives are associated with development and progression of age-related macular degeneration (AMD).

Materials and Methods:

This retrospective, observational cohort study using healthcare claims data from a U.S. nationwide managed-care network between January 1, 2006 and December 31, 2016 included enrollees ≥40 years old with primary open angle glaucoma with or without a diagnosis of nonexudative AMD at the index date. Hazard ratios (HR) for developing AMD or progressing from nonexudative to exudative AMD with exposure to ocular antihypertensive medications were analyzed.

Results:

Of 132 963 eligible enrollees, 118 174 (87.5%) had no diagnosis of AMD at baseline while 14 789 (12.5%) had a diagnosis of nonexudative AMD. Prostaglandin analog exposure had a decreased hazard of developing AMD among individuals without baseline disease (HR, 0.90; 95% CI, 0.87–0.94; p<0.0001), while topical alpha₂-agonist exposure demonstrated an increased hazard of AMD development (HR, 1.08; 95% CI, 1.03–1.14; p=0.004). Among patients with baseline nonexudative AMD, topical carbonic anhydrase inhibitor exposure was associated with a decreased hazard of progressing to exudative disease (HR, 0.84; 95% CI, 0.71–0.99; p=0.04) while topical alpha₂-agonists had increased hazard (HR, 1.17; 95% CI, 1.01–1.36; p=0.04).

Conclusions:

Certain ocular antihypertensive medications may be associated with development or progression of AMD. Their role in AMD pathogenesis should be better understood as they are considered for therapeutics in this disease.

Introduction

Age-related macular degeneration (AMD) is a leading cause of blindness in industrialized nations, affecting 6.5% of Americans over the age of 40 and estimated to impact 200 million patients globally.^1–4 Additionally, advanced AMD is associated with increased rates of depression and functional disability among the elderly.⁵ AMD is classified as either nonexudative or exudative, with the exudative variant being an advanced form characterized by the presence of choroidal neovascularization. Geographic atrophy represents another form of advanced AMD characterized by irreversible retinal pigment epithelium, photoreceptor, and choriocapillaris loss,⁶ leading to significant rates of blindness and visual disability.⁷ The advent of anti-vascular endothelial growth factor (anti-VEGF) therapy has vastly improved the ability to manage exudative AMD, but there remains an unmet need for therapeutics in nonexudative and exudative AMD, and associated geographic atrophy.

Primary open angle glaucoma (POAG) is the second leading cause of blindness globally, affecting over 44 million patients.^8,9 Ocular antihypertensive medications are first line therapy for treatment of POAG. Recently, these drugs have been investigated as therapeutics for the different forms of AMD. Topical dorzolamide-timolol has been used in treatment of refractory exudation in AMD with small pilot studies and a recent randomized control trial demonstrating decreased central macular thickness when used as an adjuvant with intravitreal anti-VEGF.^10–12 In addition, quarterly intravitreal brimonidine injections were shown in a phase 2b clinical trial (NCT02087085) to modestly decrease geographic atrophy expansion at 24 and 30 months, with a phase 3 trial planned.¹³

While select ocular antihypertensives may show promise in treating specific subsets of AMD, the mechanism by which they modulate AMD pathogenesis and their effect at different stages of the disease have not been well studied. Given this, we sought to assess in a nationwide healthcare claims database whether the use of ocular antihypertensive medications may be associated with the development of AMD or its progression from the nonexudative to exudative form.

Materials and methods

Data Source

We accessed Clinformatics DataMart (OptumInsight, Eden Prairie, MN), which captures insurance claims records of enrollees in a U.S. nationwide managed-care network from 01/01/2006– 12/31/2016. Claims from outpatient healthcare encounters and associated International Classification of Disease (ICD9 or ICD10) diagnosis codes¹⁴ linked to each encounter were available. Age, sex, race/ethnicity, annual household income, and education level of each enrollee was available. Pharmacy records of all outpatient medication fills for all enrollees were available. Those considered in this analysis were prescription fills for the following classes of ocular antihypertensive medications: topical prostaglandin analogs, beta blockers, carbonic anhydrase inhibitors, and alpha₂-agonists, as well as oral carbonic anhydrase inhibitors. The Stanford University Institutional Review Board exempted this study involving de-identified data from review.

Inclusion/Exclusion Criteria

We included individuals with a baseline diagnosis of primary open angle glaucoma, as identified by ICD9 or ICD10 billing codes (ICD9: 365.10–365.15, 365.89, 365.9; ICD10: H40.10–12-, H40.15-) associated with outpatient encounters with ophthalmologists during their first 3 years of continuous enrollment in the plan (e.g. baseline 3-year lookback period). Patients with only a glaucoma suspect diagnosis or billing codes associated with other forms of glaucoma over the lookback period were excluded. We included only primary open angle glaucoma patients to decreased confounding, as other forms of glaucoma with different pathogeneses could theoretically influence AMD disease course. Patients must have been at least 40 years of age at the beginning of follow-up (after the 3-year lookback period) to be included. Patients were divided into two cohorts: the no AMD cohort consisting of patients without any billing codes associated with AMD (nonexudative or exudative) at baseline over the lookback period (nonexudative: ICD9 362.50–1, 362.57, ICD10 H35.30–31-; exudative: ICD9 362.52, ICD10 H35.32-), and the nonexudative AMD cohort with nonexudative AMD at baseline but not exudative AMD.

Outcomes

The primary outcome was development of any form of macular degeneration in the no AMD cohort, and progression from nonexudative to exudative AMD in the nonexudative AMD cohort. These outcomes were identified by the occurrence of an ICD code for AMD (nonexudative or exudative) at any point from the end of the lookback period onward.

Data Analysis

Data analyses were performed using SAS software, version 9.4 (SAS, Inc., Cary, NC). Characteristics of the study population were summarized using mean/standard deviation (SD) or frequency/percentage as appropriate. Multivariable Cox regression modeling was performed to analyze predictors for incidence of AMD among patients without AMD at baseline, and predictors for progression to exudative AMD among patients with nonexudative AMD at baseline. Predictors included in the multivariable models were age, sex, race/ethnicity, education, annual household income, and exposure to topical prostaglandin analogs, beta-blockers, carbonic anhydrase inhibitors, and alpha₂-agonists, and exposure to oral carbonic anhydrase inhibitors. Many patients had exposure to more than one ocular antihypertensive class. Because patients have complex drug exposures that change over time, we considered ocular antihypertensive exposure to be a time-varying predictor such that drug exposures which may be changing over time were accounted for, rather than only the drug exposure calculated at baseline. This is a standard method for including time-varying predictors in Cox proportional hazards models. Ocular antihypertensive exposure at any given point in time was determined through prescription fill records during the preceding 3 years, dichotomized in two different ways: 1) having any exposure to that medication class versus no exposure to that medication class, or 2) having at least 30 days of exposure to that medication class versus less than 30 days of exposure to that medication class. For all analyses P<0.05 was considered statistically significant.

Results

Population Characteristics

There were 132 963 enrollees with POAG who met inclusion criteria. Table 1 shows the baseline characteristics of all POAG patients, of whom 118 174 (87.5%) had no AMD at baseline (no AMD cohort) and 14 789 (12.5%) had nonexudative AMD (nonexudative AMD cohort). The no AMD and nonexudative AMD cohorts both had slight female predominance (56.7% and 61.6%, respectively) and were majority white (68.4% and 76.1%, respectively). Patients with baseline nonexudative AMD tended to be older than their non-AMD counterparts (mean 76.5 and 68.8 years, respectively). In the no AMD cohort, 11 941 (10.1%) developed nonexudative or exudative AMD. Among patients with nonexudative AMD, 1 508 (10.2%) developed the exudative variant.

Table 1:

Population Characteristics of Individuals with POAG on Intraocular Pressure Lowering Medications With or Without Baseline Nonexudative AMD.

Demographics	POAG (n=118 174) No. (%)	POAG + NE-AMD (n=14 789) No. (%)
Age, mean (SD), y	68.8 (10.6)	76.5 (8.0)
Sex
Male	51 108 (43.3%)	5 687 (38.5%)
Female	67 066 (56.7%)	9 102 (61.6%)
Race/Ethnicity
White	80 771 (68.4%)	11 258 (76.1%)
Black	20 695 (17.5%)	1 372 (9.3%)
Hispanic	8 744 (7.4%)	928 (6.3%)
Asian	4 202 (3.6%)	688 (4.7%)
Unknown	3 762 (3.2%)	543 (3.7%)
Education
<12th grade	584 (0.5%)	56 (0.4%)
High school grad	35 521 (30.1%)	4 210 (28.5%)
College attendee	63 228 (53.5%)	8 317 (56.2%)
College grad	18 841 (15.9%)	2 206 (14.9%)
Annual household income ($)
<40,000	40 596 (34.4%)	6 827 (46.2%)
40–49,999	11 428 (9.7%)	1 510 (10.2%)
50–59,999	10 234 (8.7%)	1 308 (8.8%)
60–74,999	12 409 (10.5%)	1 380 (9.3%)
75–99,999	15 467 (13.1%)	1 550 (10.5%)
≥100,000	28 040 (23.7%)	2 214 (15.0%)
Disease progression*
Number of events	11 941 (10.1%)	1 508 (10.2%)
Days of follow-up, median (range), d	963.0 (4017)	732 (4014)

Abbreviations: POAG=primary open angle glaucoma; AMD= age-related macular degeneration; NE-AMD=nonexudative age-related macular degeneration

^*Progression for POAG patients is the development of any form of AMD and for POAG + AMD patients is progression from to exudative AMD.

Development of AMD

Multivariable Cox proportional hazards modeling evaluated the role of ocular antihypertensive medications in the development of either exudative or nonexudative AMD in the no AMD cohort, while adjusting for demographic characteristics (Table 2). Age and female sex were both associated with increased hazard of developing any AMD (HR, 1.07 per 1-year increase in age; 95% CI, 1.07–1.08; HR, 1.09; 95% CI, 1.05–1.13; respectively). While black or Hispanic persons had statistically significant decrease in hazard of developing AMD (HR, 0.61; 95% CI, 0.57–0.64; HR, 0.85; 95% CI, 0.79–0.92; respectively), Asian persons showed an increased hazard of AMD development (HR, 1.21; 95% CI, 1.10–1.34) relative to white persons. Household income did not predict incidence of AMD, while education level greater than high school graduate was associated with an increased hazard of AMD development (high school graduate HR, 1.42; 95% CI, 1.05–1.91; college attendee HR, 1.42; 95% CI, 1.05–1.92) compared with persons with the lowest education level.

Table 2:

Multivariable Cox Proportional Hazards Model for Development of AMD Among POAG Patients Without Baseline AMD.

	Any Exposure to Ocular Antihypertensives		>30 Days Exposure to Ocular Antihypertensives
Covariates	Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P
Age	1.07 (1.07–1.08)	<0.0001	1.07 (1.07–1.08)	<0.0001
Female	1.09 (1.05–1.13)	<0.0001	1.09 (1.05–1.13)	<0.0001
Race/Ethnicity
Black	0.61 (0.57–0.64)	<0.0001	0.61 (0.57–0.65)	<0.0001
Hispanic	0.85 (0.79–0.92)	<0.0001	0.86 (0.79–0.93)	0.0001
Asian	1.21 (1.10–1.34)	0.0001	1.22 (1.11–1.35)	<0.0001
Unknown	0.95 (0.86–1.06)	0.36	0.96 (0.86–1.06)	0.40
Annual household income ($)
40–49 999	0.98 (0.92–1.05)	0.56	0.98 (0.92–1.05)	0.61
50–59 999	1.01 (0.95–1.08)	0.76	1.00 (0.94–1.07)	0.99
60–74 999	0.96 (0.89–1.02)	0.18	0.96 (0.90–1.03)	0.22
75–99 999	0.98 (0.92–1.04)	0.51	0.98 (0.92–1.05)	0.61
≥100 000	1.00 (0.94–1.06)	0.96	1.00 (0.94–1.06)	0.94
Education
High school grad	1.42 (1.05–1.91)	0.02	1.41 (1.05–1.90)	0.02
College attendee	1.42 (1.05–1.92)	0.02	1.41 (1.05–1.90)	0.02
College graduate	1.35 (0.99–1.82)	0.06	1.34 (0.99–1.81)	0.06
Ocular antihypertensive medications
Topical PGA	0.90 (0.87–0.94)	<0.0001	0.89 (0.86–0.93)	<0.001
Topical BB	0.96 (0.91–1.00)	0.05	0.96 (0.91–1.00)	0.05
Topical AA	1.08 (1.03–1.14)	0.004	1.08 (1.02–1.15)	0.005
Topical CAI	0.99 (0.93–1.05)	0.67	0.97 (0.92–1.03)	0.37
Oral CAI	1.17 (1.02–1.34)	0.03	0.98 (0.77–1.24)	0.85

Abbreviations: AMD= age-related macular degeneration; CAI= carbonic anhydrase inhibitor; PGA= prostaglandin analog; BB= beta-blocker; AA= alpha₂-agonist

Exposure to topical prostaglandin analogs conferred a significantly decreased hazard of developing any form of macular degeneration (HR, 0.90; 95% CI, 0.87–0.94) compared to no exposure. In contrast, POAG patients with any use of topical alpha₂-agonists or oral carbonic anhydrase inhibitors had increased hazard of developing nonexudative or exudative AMD over the follow-up period (HR, 1.08; 95% CI, 1.03–1.14; HR, 1.17; 95% CI, 1.02–1.34; respectively), compared to no usage of these medications. Sensitivity analyses for regressions were carried out controlling for a variety of systemic comorbidities including diabetes, hypertension, congestive heart failure, stroke, myocardial infarctions, renal disease, dementia and depression, which was found to not significantly affect the results of the primary predictors of medication usage.

Additionally, to account for the duration of therapy and the fact that some medications, especially oral carbonic anhydrase inhibitors, tend to be prescribed for only short durations, a similar analysis was conducted that considered exposure to a particular class of medication only if pharmacy records demonstrated the enrollee had been prescribed the ocular antihypertensive medication for greater than 30 days. As seen in Table 2, the multivariable Cox proportional hazards model for the development of AMD among POAG patients without baseline AMD demonstrated very little change in the associations between the covariates and the development of either nonexudative or exudative AMD when controlling for the duration of therapy in this manner. Importantly, though, controlling for a potential short duration of therapy resulted in oral carbonic anhydrase inhibitors no longer being associated with an increased hazard of developing nonexudative or exudative AMD over the follow-up period (HR, 0.98; 95% CI, 0.77–1.24).

Progression from nonexudative to exudative AMD

The cohort of enrollees (n = 14 789) who had diagnoses of both POAG and nonexudative AMD at the beginning of the follow-up period was evaluated for progression to exudative AMD using similar Cox proportional hazards modeling, adjusted for demographic characteristics and duration of therapy (Table 3). Age was associated with a statistically significant increased hazard of progressing from nonexudative to exudative AMD (HR, 1.05 per 1 year increase in age; 95% CI, 1.04–1.06). Female sex, while predictive of AMD development above, was not significantly associated with AMD progression among patients with existing nonexudative disease. Black persons, again, appeared protected from AMD progression (HR, 0.61; 95% CI, 0.48–0.76) and Asian persons similarly demonstrated a decreased hazard of progression to exudative AMD (HR, 0.48; 95% CI 0.33–0.71) relative to white persons. There was no significant association between income or education level and AMD progression.

Table 3:

Multivariable Cox Proportional Hazards Model for Progression From Nonexudative to Exudative AMD among POAG patients.

	Any Exposure to Ocular Antihypertensives		>30 Days Exposure to Ocular Antihypertensives
Covariates	Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P
Age	1.05 (1.04–1.06)	<0.0001	1.05 (1.04–1.06)	<0.0001
Female	1.06 (0.95–1.18)	0.31	1.07 (0.96–1.19)	0.24
Race/Ethnicity
Black	0.61 (0.48–0.76)	<0.0001	0.61 (0.49–0.77)	<0.0001
Hispanic	0.98 (0.78–1.23)	0.86	0.97 (0.77–1.22)	0.79
Asian	0.48 (0.33–0.71)	0.0002	0.48 (0.33–0.70)	0.0002
Unknown	0.99 (0.76–1.30)	0.95	1.01 (0.77–1.31)	0.99
Annual household income ($)
40–49,999	1.01 (0.85–1.20)	0.92	1.00 (0.84–1.18)	0.99
50–59,999	0.94 (0.77–1.15)	0.55	0.97 (0.80–1.17)	0.99
60–74,999	1.02 (0.85–1.22)	0.87	1.01 (0.85–1.21)	0.91
75–99,999	0.99 (0.82–1.19)	0.91	0.98 (0.82–1.18)	0.87
≥100 000	0.87 (0.72–1.05)	0.15	0.88 (0.73–1.06)	0.17
Education
High school grad	1.11 (0.41–3.00)	0.84	1.12 (0.42–3.02)	0.82
College attendee	1.21 (0.45–3.26)	0.70	1.22 (0.45–3.27)	0.70
College graduate	1.39 (0.51–3.77)	0.52	1.41 (0.52–3.83)	0.50
Ocular antihypertensive medications
Topical PGA	0.93 (0.83–1.04)	0.20	0.93 (0.83–1.04)	0.20
Topical BB	0.91 (0.80–1.04)	0.15	0.90 (0.79–1.02)	0.11
Topical AA	1.17 (1.01–1.36)	0.04	1.19 (1.02–1.39)	0.03
Topical CAI	0.84 (0.71–0.99)	0.04	0.79 (0.66–0.94)	0.009
Oral CAI	0.60 (0.34–1.06)	0.08	0.97 (0.46–2.05)	0.94

Abbreviations: AMD= age-related macular degeneration; POAG=primary open angle glaucoma; CAI= carbonic anhydrase inhibitor; PGA= prostaglandin analog; BB= beta-blocker; AA= alpha₂-agonist

While topical prostaglandin analogs were associated with a decreased hazard of AMD development, the effect on AMD progression to exudation was not statistically significant. Exposure to alpha₂-agonists again demonstrated an association with AMD, showing a statistically significant increased hazard of progressing to exudative disease (HR, 1.17; 95% CI, 1.01–1.36). Topical carbonic anhydrase inhibitors were associated with a decreased hazard of AMD progression (HR, 0.84; 95% CI, 0.71–0.99) (Table 3). Accounting for the duration of therapy had little effect on the associations between the covariates and the progression to exudative AMD in the cohort of enrollees who had diagnoses of both POAG and nonexudative AMD at the beginning of the follow-up period (Table 3).

Discussion

To our knowledge, this is the first study to comprehensively query the association of topical and oral ocular antihypertensive medications with age-related macular degeneration development and progression. This study demonstrated that topical alpha₂-agonist exposure was associated with an increased hazard of AMD development while topical prostaglandin analog exposure was protective. Among those with nonexudative AMD at baseline, topical alpha₂-agonist exposure was again associated with an increased hazard of progression to exudative AMD whereas topical carbonic anhydrase inhibitor exposure showed a decreased hazard. Topical beta-blockers did not have a statistically significant association with development or progression of AMD. Prior work done by Kolomeyer et al. similarly did not show an increased hazard of progression to exudative AMD among patients using oral beta blockers.¹⁵

Identifying factors that may protect against or promote AMD are of significant interest at a population-health level given the prevalence of this disease. Our results are particularly pertinent in light of the recent and ongoing clinical trials investigating dorzolamide-timolol¹² and intravitreal brimonidine (NCT02087085)²⁷ as treatment agents in exudative AMD and geographic atrophy, respectively. Despite interest in using ocular antihypertensives as novel therapeutics in AMD, existing studies of their potential effects on this disease are lacking. Though small, the associations demonstrated in this study underscore the need to better understand how these medications may modulate the course of AMD.

Our population demographics were consistent with previously described AMD populations, with macular degeneration patients tending to be older and female, and 10.2% of nonexudative AMD patients progressing to exudation.^16–18 A decreased hazard of developing AMD and progressing to exudation was observed in black relative to white persons, which has previously been shown in several population-based studies.^19–22 The decreased hazard of development of nonexudative AMD in Hispanic persons and the increased hazard in Asian persons demonstrated herein is in part consistent with prior published findings. Vanderbeek et al. described a decreased hazard of nonexudative AMD in Latinos by the age of 80 years old and increased hazard in Asians at age 60 years old, though this effect was not seen by age 80.²² Our current study demonstrated a decreased hazard of progression to exudative AMD in Asian as compared to white persons. This is in agreement with the aforementioned study by Vanderbeek et al. that found a decreased hazard of exudative AMD in elderly Asian American persons.²² In contrast, a study by Klein et al. in the MESA cohort demonstrated Chinese-Americans had higher rates of exudative AMD compared to white persons.²⁰

Our primary analysis examined whether exposure to ocular antihypertensives was associated with the development and progression of AMD. One class of ocular antihypertensive, alpha₂-agonists, are thought by many to have neuroprotective in addition to intraocular pressure lowering effects. Interestingly, we found that exposure to alpha₂-agonists was associated with an increased hazard of both developing AMD and progressing to exudation. While rat models have shown alpha₂-agonists may promote photoreceptor and retinal ganglion cell survival and thus play a neuroprotective role at the level of the retina and optic nerve,^23–25 the mechanism by which this may occur has not been well characterized. Further, there is a paucity of clinical studies demonstrating neuroprotective effects of alpha₂-agonists in human subjects.²⁶ The associations seen with alpha₂-agonists and AMD development and progression in this study in conjunction with an incomplete understanding of these drugs’ mechanism and neuroprotective effects at the level of the retina should prompt further investigation. This is particularly salient as the intravitreal brimonidine implant progresses to phase 3 trials for geographic atrophy,¹³ though we recognize the pathogenesis of nonexudative or exudative AMD is not equivalent to geographic atrophy. Notably, the phase 2b trial showed only modest results with the primary endpoint of change in geographic atrophy area compared to controls just meeting statistical significance (p = 0.047) and secondary outcome of change in best-corrected visual acuity not reaching significance (NCT02087085).²⁷

Carbonic anhydrase inhibitors (CAI) had differing effects on the course of AMD with different administration routes as well as when accounting for duration of therapy. When evaluating progression to exudative AMD, topical carbonic anhydrase inhibitors had a decreased hazard of progression. There was no effect on development of exudation with oral carbonic anhydrase inhibitors, but this lack of observed association may be secondary to the small number of patients in this sub-group. The decrease in progression to wet AMD seen with topical CAIs may represent a delay in recognizing the conversion to exudative AMD. In a randomized control trial, topical dorzolamide-timolol in combination with continued anti-vascular endothelial growth factor (VEGF) injections has been shown to decrease central macular thickness in patients with persistent exudative AMD,¹² with the dorzolamide component of the medication postulated to regulate the pumping of fluid out of the retina by the retinal pigment epithelium.¹⁰ Additionally, dorzolamide is commonly used for treatment of cystoid macular edema in other retinal diseases including retinitis pigmentosa and X-linked retinoschisis.^28,29 This previously established retinal drying effect of dorzolamide may mask detection of early conversion to exudative AMD on examination and explain the decreased hazard of developing exudative AMD in persons exposed to topical CAIs seen in this study.

In contrast to conversion to exudation, an association between topical CAIs and the development of AMD was not observed. Any exposure to oral CAIs was associated with an increased hazard of developing AMD, but this association was no longer apparent when using greater than 30 days of medication fills as the cutoff for exposure to a particular ocular antihypertensive medication. The effect by oral but not topical CAIs may have to do with increased drug delivery to the outer retina and RPE, where CAIs modulate lactate metabolism, which is important to maintaining outer retinal homeostasis and photoreceptor function and viability.^34–35 Yet, as these medications tend to be utilized to control severe elevations in intraocular pressure that are otherwise uncontrollable with maximal topical therapies, we cannot rule out that other factors associated with these severe intraocular pressure elevations might be underlying the association between oral CAIs and the development of AMD. The lack of association between oral CAIs and the development of AMD when controlling for the duration of therapy may be due to a lack of power in this subgroup as this oral medication is usually prescribed for only short durations.

Lastly, topical prostaglandin analog exposure was associated with a decreased hazard of developing AMD. Prostaglandins have cellular and tissue-level activities that may explain this protective effect. Studies of unoprostone isopropyl showed that this prostaglandin promotes choroidal blood flow.³⁶ This is of relevance because impaired choroidal blood flow has been associated with AMD development.^37,38 Prostaglandins have been investigated to a limited degree for use in AMD. Alprostadil, a prostaglandin given intravenously, has previously been shown to be superior to placebo when evaluating visual acuity changes in patients with dry AMD.³⁹ We saw no effect of topical prostaglandin analog exposure on the progression to exudative AMD, which is consistent with previously published data showing no association between long-term topical prostaglandin analog exposure and development of choroidal neovascular membranes.⁴⁰ Additionally, though, topical prostaglandin analogs are often first-line treatment in POAG. Therefore, these medications may have been used in those patients with less severe glaucoma who require less frequent follow-up. As such, these patients, on average, may have fewer ophthalmic visits, lowering the likelihood that they are diagnosed with macular degeneration.

The strengths of this study include the large, nationwide sample of insured individuals in the United States that allowed for a substantial number of persons exposed to the different classes of ocular antihypertensives. Patients with non-POAG glaucoma diagnoses were excluded, decreasing confounding from baseline variability in comorbid ocular disease. Additionally, we minimized recall bias by using pharmacy records for outpatient prescription fills rather than patient reported medication use.

Several limitations should be considered when interpreting this study. The retrospective nature of the analysis allows for the identification of associations, but limits the ability to draw causal inferences about the effects of ocular antihypertensive medications. Additionally, though using data from prescription fills of a medication decreases recall bias, medication compliance cannot be confirmed. Insurance claims data analysis uses ICD coding to classify diagnoses. While the study design used was stringent in code selection for each diagnosis, misclassification is possible and diagnoses could not be verified through individual medical record review. Certain risk factors for the progression of AMD are not available in the data source, such as smoking status and body mass index.^41–42 Likewise, claims data is unable to account for nonprescription medications, such as the vitamins utilized in the Age-Related Eye Disease Studies.^43–44 Also, patients with advanced glaucoma are more likely to be on second- or third-line ocular antihypertensive agents including beta-blockers, alpha₂-agonists, or carbonic anhydrase inhibitors. Therefore, prescription fills of these drugs may be a surrogate for POAG severity, which is not easily controlled for with claims data as ICD coding for glaucoma severity is often missing or unreliable. While a recent case-control study did show an association of exudative AMD with prior diagnosis of POAG,⁴⁵ an association between AMD and POAG severity has not been well established in the literature.

In conclusion, this retrospective, observational longitudinal cohort study of over 130,000 individuals with POAG demonstrated an association between topical alpha₂-agonist exposure and both the development and progression of AMD. In contrast, topical prostaglandin analogs demonstrated a protective effect on the development of AMD. Though the effect sizes reported were small, these results raise the possibility that ocular antihypertensive medications may play a role in development of AMD and progression from nonexudative to exudative disease. Further studies are needed to better elucidate the effect of ocular antihypertensive medications on AMD pathogenesis, particularly as these medications are being considered for use as AMD therapeutics.