An estimated 27% of U.S. adults with diabetes are using glucagon-like peptide-1 receptor agonists (GLP-1 RAs) — a type of medication that mimics the GLP-1 hormone — to lower blood sugar and support weight loss. Some research has suggested that their use can reduce the risk of developing other diseases.
A new, federally funded retrospective study led by Johns Hopkins Medicine strongly suggests that semaglutide and other GLP-1 RAs did not statistically change the risk of developing neovascular age-related macular degeneration (NVAMD) — a fast-progressing, blinding condition that is caused by uncontrolled, abnormal blood vessel growth in the back of the eye — in adults with type 2 diabetes without a prior history of GLP-1 use.
A peer-reviewed report of the work, funded by the National Institutes of Health, published online June 2 in Ophthalmology.
Cindy Cai, M.D., principal investigator and Jonathan and Marcia Javitt Rising Professor of Ophthalmology at Johns Hopkins Medicine, says the study was designed to resolve existing, conflicting research on the possible link between GLP-1 RAs and age-related macular degeneration.
"Prior to our study, GLP-1s were reported to both increase AND decrease the risk of developing AMD in the literature," Cai says. "We wanted to resolve the lack of consensus with our work."
Analyzing de-identified patient data collected from December 2017 through December 2024 across 12 databases managed by the Observational Health Data Sciences and Informatics (OHDSI) network — an international and interdisciplinary collaborative of researchers and observational health databases — the researchers followed the health outcomes of adults with type 2 diabetes who had been prescribed semaglutide (227,971), dulaglutide (68,588), exenatide (5,460), empagliflozin (252,356), sitagliptin (100,083) or glipizide (213,515) for the first time. First-time users were defined as individuals who did not have GLP-1 RAs listed on their health records for at least 365 days and who only initiated the medications as a second-line treatment to metformin.
"We included other GLP-1 receptor agonists in our analysis to show our findings weren't specific to a single drug," says Cai.
People who developed neovascular AMD were identified based on specific medical codes present in their database profiles. Two researcher-devised definitions of the disease were used to ensure they did not miss any cases: 1) condition-procedure NVAMD (NVAMD-CP) and 2) condition-only NVAMD (NVAMD-C). Under the first definition, medical codes for both NVAMD and NVAMD treatment were required on a medical profile to have the condition. Under the second definition, a patient only needed an NVAMD diagnostic code to qualify.
Using both NVAMD definitions, Cai's team performed two sets of analyses to calculate the rates of neovascular age-related macular degeneration onset for semaglutide and the other medications: an active-comparator cohort analysis and a self-controlled case-series analysis.
In the active-comparator cohort analysis, the researchers compared the risk of neovascular age-related macular degeneration onset between statistically similar patients on each medication to see whether people were more or less likely to develop NVAMD. They found that the risk of developing neovascular age-related macular degeneration while taking semaglutide was comparable to the other GLP-1s and non-GLP-1 treatments.
In the self-controlled case-series analysis, the researchers only studied patients who developed neovascular age-related macular degeneration. Cai's team determined that the incidence-risk ratio — the likelihood of developing NVAMD while taking a treatment versus the likelihood of developing the disease while not on the medication — for semaglutide was 1.02 using the NVAMD-CP definition and 0.92 using the NVAMD-C definition. A risk ratio of 1 suggests no difference between the two groups.
Overall, in both analyses, the risk of developing neovascular age-related macular degeneration while taking semaglutide or the other study medications for type 2 diabetes did not differ enough to be considered greater or less than random chance, the researchers concluded.
While the study clarifies some conflicting findings on NVAMD risk in adults taking GLP-1s for type 2 diabetes, Cai warns that her group's findings should not be applied to other groups, such as people taking GLP-1 drugs primarily for weight loss, and additional research should be done in people without type 2 diabetes.
"Our study only looked at patients with existing type 2 diabetes who were prescribed semaglutide and other GLP-1 RAs," she says. "But you don't need a type 2 diabetes diagnosis to take these medications, so we can't say if our findings hold true beyond this patient group."
In addition to Cindy Cai, researchers who authored the study, in alphabetical order, are Mohammed Adibuzzaman, Thamir Alshammari, Karen Armbrust, Andrew J. Barkmeier, Sally Baxter, Clair Blacketer, Michael V. Boland, William J. O'Brien, Eric N. Brown, Aiyin Chen, Hannah Morgan-Cooper, Priya Desai, David A. Dorr, Scott DuVall, Thomas Falconer, Ruochong Fan, Kerry Goetz, Michelle Hribar, Izabelle Humes, George Hripcsak, Ghazala O'Keefe, Yashwant Kothari, Cecilia S. Lee, Haeun Lee, Theodore Leng, Rumel Mahmood, Nestoras Mathioudakis, Benjamin Martin, David McCoy, Nitish Mehta, Paul Nagy, Akihiko Nishimura, Shikha Anna Ostropolets, Philip R.O. Payne, Gowtham Rao, Patrick Ryan, Anthony Sena, Azza Shoaibi, Jacqueline C. Stocking, Marc A. Suchard, Brian Toy, Diep Tran, Edmund Tsui, Sophia Wang, Erik Westlund and Linying Zhang.
The study authors have the following conflict-of-interest disclosures to share: Cai receives grant support from Regeneron Pharmaceuticals, Inc. and equipment support from Optomed. Goetz is an employee of Topcon Healthcare, Inc. and Baxter has received consulting fees from Topcon. Boland is a speaker for Carl Zeiss Meditec and provides consulting services to the company as well as Abbvie. Mehta serves on the 4DMT, Regeneron Pharmaceuticals and Eyepoint Pharmaceuticals advisory boards. DuVall is an employee and shareholder of PurpleLab. Sena, Blacketer and Ostropoiets. Shoaibi and Ryan are employees and shareholders of Johnson & Johnson. Finally, Suchard receives contacts and grants from Janssen Research & Development and Gilead Sciences, Inc., which are outside the scope of this work.
The study was supported by the National Institutes of Health (K23EY033440; K23EY032985, UL1TR001855, T15LM013979, R35GM160458, U24HD113136, R01DK125780, R01DK134955, K12GCAEI0492E, P30EY022589, OT2OD032644, K23EY03263501, 1K01HL168222, R01AG060942, OT2OD32644, UL1TR002345, UL1TR002369, UL1TR002369, UL1TR002369, UL1TR002369, and R01HL169954, R01HL167858).
DOI: 10.1016/j.ophtha.2026.05.034
