Highlights
-
•
Levodopa exposure associated with lower risk of neovascular AMD conversion.
-
•
Dopamine receptor agonists showed no significant protective association.
-
•
Target trial emulation using a large electronic health record database.
-
•
Dopaminergic signaling may influence progression of age-related macular degeneration.
Purpose
Age-related macular degeneration (AMD) is a leading cause of central vision loss worldwide. Emerging evidence suggests that targeting dopaminergic pathways may influence AMD progression. This study evaluates whether levodopa ± carbidopa or dopamine receptor D2 (DRD2) agonists are associated with reduced risk of conversion to neovascular AMD (nAMD).
Design
Population-based clinical cohort study.
Participants
Adults aged ≥18 years diagnosed with non-nAMD between May 2005 and May 2025, within the TriNetX US Collaborative Network, a federated electronic health record database spanning 69 healthcare organizations.
Methods
This retrospective cohort study emulated four distinct target trials comparing new users of (1) levodopa (± carbidopa) or (2) DRD2 agonists (pramipexole, ropinirole, bromocriptine, rotigotine, or cabergoline) to new users of two comparators (pantoprazole or gabapentin). Patients with prior nAMD or prescriptions of other dopamine-enhancing agents (eg , selegiline, rasagiline, tolcapone) were excluded. Each exposure group was independently matched to comparators using 1:1 propensity score matching for selected demographics, social factors, comorbidities, and AMD stage.
Main Outcome Measures
The primary outcome was 3-year risk of conversion from non-nAMD to nAMD. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. An α level of 0.05 was used to determine statistical significance.
Results
Patients prescribed levodopa ± carbidopa had a reduced 3-year risk of conversion to nAMD relative to their matched counterparts prescribed pantoprazole (levodopa n = 1312, control n = 1312; HR 0.67; 95% CI, 0.45-0.98) and gabapentin (levodopa n = 1675, control n = 1675; HR 0.69; 95% CI, 0.50-0.95). No significant difference was observed in 3-year risk of conversion to nAMD between DRD2 agonists use relative to pantoprazole (DRD2 n = 1603, control n = 1603; HR 0.81; 95% CI, 0.58-1.13) or gabapentin (DRD2 n = 2779, control n = 2779; HR 0.92; 95% CI, 0.72-1.19).
Conclusions
In this cohort study, levodopa ± carbidopa use was associated with lower 3-year risk of conversion to nAMD in two independent matched comparisons, whereas DRD2 agonists were not associated with significant differences in risk. These findings suggest dopaminergic signaling may influence AMD progression and warrant further prospective investigation.
INTRODUCTION
Age-related macular degeneration (AMD) is the leading cause of irreversible central vision loss in older adults worldwide. Recent investigations have explored the potential neurovascular and immunomodulatory roles of dopamine signaling in AMD pathogenesis. ,,,,,,, Experimental and pharmacoepidemiologic data suggest that dopamine pathways via (1) G-protein coupled receptor 143 (GPR143) and (2) dopamine receptor D2 (DRD2) may modulate oxidative stress, angiogenesis, and retinal pigment epithelium stability through the downstream regulation of vascular endothelial growth factor (VEGF) and exosome release. ,,,,,,,
Levodopa, a dopamine precursor and GPR143 ligand, has been associated with delayed neovascular AMD (nAMD) onset. reduced progression to geographic atrophy. and reduced risk of conversion to nAMD. in retrospective analyses. Moreover, emerging evidence from animal models and national insurance data demonstrated antiangiogenic activity associated with DRD2 activation and reduced anti-VEGF treatment burden in nAMD patients receiving DRD2 agonist therapy. Though these findings are promising, comparative evidence evaluating the potential protective effects of levodopa and DRD2 agonists in AMD progression remains limited.
Herein, we evaluate whether dopamine-enhancing therapies, including levodopa ± carbidopa and DRD2 agonists, are associated with reduced risk of progression from non-nAMD to nAMD. Using the TriNetX research platform, we defined two primary exposures: (1) levodopa with or without carbidopa, and (2) DRD2 agonists (pramipexole, ropinirole, bromocriptine, rotigotine, or cabergoline). Each exposure group was compared to two separate control cohorts consisting of patients prescribed pantoprazole or gabapentin, medications commonly used in older adults with no known effect on AMD risk. We conducted four pairwise comparisons to evaluate conversion risk over a 3-year period.
METHODS
We conducted a retrospective cohort study using the TriNetX US Collaborative Network, a federated research platform aggregating deidentified electronic health record data from 69 healthcare organizations. The study included data from May 2005 through May 2025. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Because the data were deidentified and analyzed at the population level, the study was determined to be exempt from review by the Western Institutional Review Board, and informed consent was not required. The study was conducted in compliance with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule and in accordance with the tenets of the Declaration of Helsinki.
We designed the study to retrospectively emulate a target trial comparing use of dopamine-enhancing agents to two comparators for prevention of conversion from non-neovascular to nAMD. Trial protocol elements, including eligibility criteria, treatment strategies, emulated assignment procedures, start of follow-up, outcome definition, and causal contrasts, were prespecified to mirror the design of a hypothetical randomized trial, consistent with best practices for target trial emulation as described by Hernán and Robins.
We included patients aged 18 years or older with a diagnosis of non-nAMD, identified using the International Classification of Diseases, 10th Revision (ICD-10) code H35.31. Two exposure cohorts were defined: (1) patients prescribed levodopa, with or without carbidopa, and (2) patients prescribed DRD2 agonists, including pramipexole, ropinirole, bromocriptine, rotigotine, or cabergoline. Both cohorts excluded patients with prescriptions for the opposite dopaminergic drug class (eg , DRD2 agonists in the levodopa group and vice versa), as well as other dopamine-enhancing agents including selegiline, rasagiline, safinamide, entacapone, tolcapone, apomorphine, istradefylline, and opicapone.
Each exposure group was compared to two control groups consisting of patients with non-nAMD who were prescribed either (1) pantoprazole or (2) gabapentin. These control medications were selected due to being commonly used, prescription-only medications with no known link to AMD progression. To ensure clear separation between exposure and control groups, we excluded control patients with any prior or subsequent use of dopamine-enhancing therapies (levodopa, DRD2 agonists, or other dopamine-enhancing agents).
The index event was defined as the date of first exposure to the study or comparator medication following a diagnosis of non-nAMD. Patients were excluded if they had prior evidence of nAMD (ICD-10 code H35.32). Those with history of other etiologies of choroidal neovascular membrane (CNVM), such as myopic CNVM (H44.2A), unspecified retinal neovascularization (H35.05), and other specified disorders of the choroid (H31.8) or history of related ocular procedures identified by Current Procedural Terminology (CPT) codes, including intravitreal injection (CPT 67028), photodynamic therapy (CPT 67221), or photocoagulation of CNVM (CPT 67220) were also excluded.
Propensity score matching (PSM) was performed using the TriNetX integrated analytics platform, which applies a greedy nearest-neighbor algorithm with a caliper of 0.1 pooled SDs (SD). We conducted 1:1 matching for each exposure-comparator pair to adjust for potential confounders. We matched for demographic variables such as age at index, gender at birth, race (White, Black or African American, Asian, American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, Other Race, Unknown), and ethnicity (Hispanic or Latino, Not Hispanic or Latino, Unknown). Matched socioeconomic and behavioral factors included tobacco use, exposure to environmental tobacco smoke, low-income status, and alcohol-related disorders. Cardiovascular and metabolic comorbidities that were matched for included essential hypertension, disorders of lipoprotein metabolism and other lipidemias, atherosclerosis, and body mass index categories (≤19.9, 20-29, 30-39, ≥40 kg/m²) along with indicators for overweight and obesity. Ophthalmic variables related to non-nAMD staging, when available, were also included, accounting for laterality and severity across both eyes. This incorporated staging for the right, left, and bilateral eyes, with categories for unspecified, early dry, intermediate dry, and advanced atrophic stages—with or without subfoveal involvement. Covariate balance between matched cohorts was evaluated using standardized mean differences, with values <0.1 indicating acceptable balance across groups.
We assessed the 3-year risk of conversion from non-nAMD to nAMD (ICD-10 H35.32) through four pairwise comparisons: (1) levodopa ± carbidopa vs pantoprazole, (2) levodopa ± carbidopa vs gabapentin, (3) DRD2 agonists vs pantoprazole, and (4) DRD2 agonists vs gabapentin. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated to assess the risk of conversion to nAMD. The proportional hazards assumption was tested and confirmed for all models using the method described by Grambsch and Therneau. Additionally, 3-year cumulative incidence was assessed using the Kaplan–Meier method with log-rank testing.
RESULTS
After PSM, the samples consisted of 2624 patients in the levodopa ± carbidopa ( n = 1312) vs pantoprazole ( n = 1312) analysis, 3350 patients in the levodopa ± carbidopa ( n = 1675) vs gabapentin ( n = 1675) analysis, 3206 patients in the DRD2 agonists ( n = 1603) vs pantoprazole ( n = 1603) analysis, and 5558 patients in the DRD2 agonists ( n = 2779) vs gabapentin ( n = 2779) analysis. Baseline demographic, clinical, and ophthalmic characteristics were well-balanced across all matched cohorts (standardized mean differences <0.1), with summaries provided in Tables S1 to S4 in the Supplementary Materials. An overview of the four emulated trials is summarized in Table .
TABLE
Three-Year Cumulative Incidence (Kaplan–Meier method) of Conversion From Non-Neovascular to Neovascular Age-Related Macular Degeneration in Four Pairwise Comparisons Between Dopamine-Enhancing Therapies and Control Medications
| Comparison | Cumulative Incidence (Treatment) | Cumulative Incidence (Control) | Hazard Ratio (95% CI) | Log-rank P Value |
|---|---|---|---|---|
| Levodopa ± carbidopa ( n = 1312) vs pantoprazole ( n = 1312) | 4.89% | 7.34% | 0.67 (0.45-0.98) | .040 |
| Levodopa ± carbidopa ( n = 1675) vs gabapentin ( n = 1675) | 5.51% | 8.54% | 0.69 (0.50-0.95) | .021 |
| DRD2 agonists ( n = 1603) vs pantoprazole ( n = 1603) | 5.79% | 7.45% | 0.81 (0.58-1.13) | .213 |
| DRD2 agonists ( n = 2779) vs gabapentin ( n = 2779) | 5.68% | 6.33% | 0.92 (0.72-1.19) | .533 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
