Back

Retinal perfusion delay associated with GLP-1 receptor agonists: A possible role of intracranial pressure?

Lee, Y. K.; Chang, Y. Y.; Jang, H. N.; Ma, D. J.; Song, H. B.

2025-09-26 ophthalmology
10.1101/2025.09.25.25336128
Show abstract

Recent studies have suggested an association between GLP-1 receptor agonist (GLP-1RA) treatment and increased risk of non-arteritic anterior ischemic optic neuropathy, particularly within the first year of treatment. This study investigated potential mechanisms by examining retinal perfusion changes in both humans and mice during GLP-1RA therapy. We first compared six patients who underwent fluorescein angiography during GLP-1RA treatment with three patients after treatment completion. Two cases in a during-treatment group showed increased arteriovenous transit time. To validate these findings, we treated wild-type mice with liraglutide and measured retinal perfusion parameters before, during, and after treatment. Mice exhibited delayed injection-to-retina time and increased arteriovenous transit time during treatment, and decreased scotopic b-wave responses and increased hypoxia-related gene expression indicating retinal hypoxia, with recovery after treatment discontinuation. Investigation of mechanisms revealed that GLP-1RA-treated mice had significantly reduced intracranial pressure (ICP) at day 7, corresponding with the timing of retinal perfusion delays, and there was a trend toward a negative correlation between ICP and injection-to-retina time. Reduced ICP may affect ophthalmic vessels in the subarachnoid space, contributing to delayed retinal perfusion. These findings suggest that ocular circulation may be particularly vulnerable to GLP-1RA-related hemodynamic effects, independent of systemic cardiovascular changes. These results indicate that careful monitoring of ocular circulation may be warranted in individuals using GLP-1RA, and suggest the need for further investigation of temporal ICP changes in human patients.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.