FDA adverse event data links semaglutide and dulaglutide to diabetic retinopathy reports | Research

The takeaway

FAERS disproportionality analysis shows strong diabetic retinopathy signal for semaglutide (PRR 19.43) and dulaglutide (PRR 9.01), weaker for tirzepatide and liraglutide, and none for lixisenatide.

PRR 19.43 for semaglutide
FDA adverse event data shows disproportionately high diabetic retinopathy reports with semaglutide—the strongest signal among GLP-1 drugs

What the researchers found

Semaglutide: PRR 19.43, ROR 19.48, Chi² 1078. Dulaglutide: PRR 9.01, ROR 9.02, Chi² 478. Tirzepatide and liraglutide: weaker but significant. Lixisenatide: no significant association.

Why it matters

With millions using GLP-1 drugs, the diabetic retinopathy signal—especially for semaglutide—warrants clinical attention. While causation is unclear, the strong disproportionality suggests either a true risk or significant reporting bias that needs investigation.

How the study worked

Disproportionality analysis of FDA FAERS database Q4/2003-Q2/2024 via OpenVigil 2.1. Calculated PRR, ROR with 95% CI. Evans' criteria for significance.

What this study cannot tell us

FAERS data is spontaneous reports—subject to reporting bias, confounding, and cannot establish causation or incidence. Semaglutide's high reporting may reflect greater market share and awareness. No denominator (number of users) available.

How to read the evidence

Disproportionality analysis of spontaneous reports—hypothesis-generating only. Cannot establish causation or true incidence. Subject to multiple biases.

When this study was published

Published in 2025; FAERS data through Q2/2024.

The bigger picture

The DR signal may reflect rapid glucose improvement causing retinopathy worsening ("early worsening" phenomenon) rather than a direct drug effect. However, the varying signal strength across different GLP-1 drugs complicates this explanation.

Questions still open

Is the DR signal due to rapid glucose improvement or a direct drug effect?
Why is the signal stronger for semaglutide than other GLP-1 drugs?
Should diabetic retinopathy screening be mandatory before starting GLP-1 therapy?

Common questions

Does semaglutide damage eyesight?

FDA adverse event reports show a disproportionately high number of diabetic retinopathy cases with semaglutide. However, this analysis cannot prove the drug causes eye damage—the reports may reflect that patients with more severe diabetes (higher retinopathy risk) preferentially receive semaglutide, or that rapid blood sugar improvement temporarily worsens existing eye disease.

Should I get an eye exam before starting a GLP-1 drug?

Based on this and other safety signals, it is reasonable to have a comprehensive eye exam before starting GLP-1 therapy, especially if you have existing diabetic eye disease. Close monitoring during treatment is also recommended, particularly during the initial months of rapid blood sugar improvement.

Read the original research

Association between glucagon-like peptide-1 agonists and risk of diabetic retinopathy: a disproportionality analysis using FDA adverse event reporting system data.

Expert review of endocrinology & metabolism, 20(2), 147-152

View on PubMed ↗Publisher’s paper ↗

Citation

Singh, Harmanjit; Natt, Navreet Kaur; Nim, Dwividendra Kumar. (2025). Association between glucagon-like peptide-1 agonists and risk of diabetic retinopathy: a disproportionality analysis using FDA adverse event reporting system data.. Expert review of endocrinology & metabolism, 20(2), 147-152. https://doi.org/10.1080/17446651.2025.2459720