The risk of neovascular age-related macular degeneration (NVAMD) associated with the use of semaglutide is similar to that with other glucagon-like peptide-1 receptor agonists (GLP-1 RAs) or non-GLP-1 RAs among adults with type 2 diabetes (T2D), suggests a study.
“[W]e believe this study provides evidence against large differences in the risk of NVAMD associated with semaglutide use among patients with T2D,” the researchers said.
A total of 227,971 new users of semaglutide were included in this retrospective study conducted across 12 databases in the Observational Health Data Sciences and Informatics (OHDSI) network from 1 December 2017 through 31 December 2024.
Two approaches were used to assess the relationship between semaglutide use and NVAMD: an active-comparator cohort design and a self-controlled case series analysis. Propensity score-adjusted Cox proportional hazards model was used in the former to estimate hazard ratios (HRs), while conditional Poisson regression models were used in the latter to estimate incidence rate ratios.
The researchers generated network-wide HR and IRR estimates using a random-effects meta-analysis. They also used two definitions of NVAMD, one based on condition codes alone (NVAMD-C) and one based on condition codes and procedures (NVAMD-CP).
The risk of NVAMD among semaglutide users was comparable to those on dulaglutide (NVMAD-C: HR, 0.57, 95 percent confidence interval [CI], 0.21‒1.57; p=0.28; NVAMD-CP: HR, 0.25, 95 percent CI, 0.05‒1.27; p=0.10) and glipizide (NVAMD-C: HR, 0.83, 95 percent CI, 0.35‒2.02; p=0.69; NVAMD-CP: HR, 0.50, 95 percent CI, 0.21‒1.19; p=0.12). [Ophthalmology 2026;133:1221-1229]
Likewise, the NVAMD risk associated with semaglutide was similar to that with empagliflozin (NVAMD-C: HR, 0.98, 95 percent CI, 0.54‒1.79; p=0.94; NVAMD-CP: HR, 0.79, 95 percent CI, 0.38‒1.64; p=0.52) and sitagliptin (NVAMD-C: HR, 2.08, 95 percent CI, 0.90‒4.83; p=0.09; NVAMD-CP: HR, 1.80, 95 percent CI, 0.55‒5.86; p=0.33).
Notably, evidence of an increased or decreased risk for NVAMD associated with semaglutide exposure (NVAMD-C: IRR, 0.92, 95 percent CI, 0.67‒1.26; p=0.60; NVAMD-CP: IRR, 1.02, 95 percent CI, 0.76‒1.36; p=0.92) nor with other GLP-1 RAs or non-GLP-1 RAs was not observed.
“Our study contributes an important finding to the literature by showing that semaglutide is associated with neither a large increased risk nor a decreased risk of NVAMD, at least among patients with T2D,” the researchers said. “This finding contradicts some of the previously published studies in this area.”
Diverse findings
The conflicting conclusions seen across observational studies were potentially driven by a combination of data sources used, study design choices, choice of target and comparison cohorts, covariates used in propensity score matching, and analytic approaches. [JAMA Ophthalmol 2025;143:408-409]
“A strength of OHDSI network studies is the ability to evaluate the impact of altering each of these study components on conclusions,” the researchers said.
In the current study, results were assessed across several databases ranging from commercial administrative claims, Medicare and Medicaid administrative claims, individual health system electronic health records, and aggregated electronic health records.
The research team also examined the impact of two study designs, namely an active-comparator new user cohort and a self-controlled case series.
“For the self-controlled case series design, the main assumption relevant to this work is that NVAMD does not affect the patient’s probability of exposure to semaglutide, which is evaluated as part of the study diagnostics,” the researchers said. [BMJ 2016;354:i4515; https://ohdsi.github.io/SelfControlledCaseSeries/articles/SingleStudies.html]
“We tested the robustness of our findings against varying definitions of our target and comparison medications,” they added.