Eye Condition Risk Elevated With Semaglutide, but Still Low

Semaglutide is linked to a small but significantly elevated risk for nonarteritic anterior ischemic optic neuropathy (NAION), according to the results of a new study.

The risk for NAION was approximately twofold greater among people taking semaglutide than among those using an SGLT2 inhibitor, but the absolute risk was still only about 29 per 10,000 people, study author Jennifer S. Lee, MD, PhD, professor of medicine at Stanford School of Medicine in Palo Alto, California, told Medscape Medical News.

In NAION, the optic nerve suddenly loses its blood supply, leading to vision loss, sometimes described as a “stroke of the eye.”

“Clinicians should balance this rare but serious vision-loss risk against semaglutide’s meaningful cardiometabolic benefits,” said Lee, who is also chief of Research and Clinical Innovations and director of the VA National Center for Collaborative Healthcare Innovation at the VA Palo Alto Healthcare System, US Department of Veterans Affairs.

“We recommend counseling patients to seek prompt evaluation if they experience visual symptoms, particularly sudden vision changes,” she added.

The new findings, recently published in JAMA Ophthalmology, are the latest from several studies examining a possible association between NAION and semaglutide, some of which have been negative.

In the current study, the investigators analyzed US Veterans Health Administration data for a total of 11,478 veterans with type 2 diabetes (T2D) who initiated semaglutide and 90,883 who initiated an SGLT2 inhibitor (mostly empagliflozin) between March 1, 2018, and March 1, 2025. Overlap weighting was used to balance the two groups by age, BMI, A1c, sex, and race.

Over a median follow-up of 2.1 years, NAION developed in 123 initiating semaglutide vs 143 initiating SGLT2 inhibitors, with rates of 123 vs 67 per 100,000 person-years. After overlap weighting, the risk for NAION was 2.3-fold higher with semaglutide (hazard ratio [HR], 2.33, P < .001).

Over a maximum of 7.5 years of follow-up, the overlap-weighted NAION incidence was 0.29% vs 0.13% for semaglutide initiators vs SGLT2 initiators.

While the mechanism(s) for the association remains unclear, hypotheses include hypotension, volume depletion from gastrointestinal side effects, rapid glucose improvement leading to stress on the eye microvasculature, and impaired vascular autoregulation at the head of the optic nerve, Lee said.

Different Studies, Different Answers

“Growing observational evidence links GLP-1 receptor agonists, particularly semaglutide, to NAION, with similar twofold to threefold risks reported in several large analyses,” Lee told Medscape Medical News.

“However, other studies have observed attenuated or null associations,” she said. “The mechanism remains unknown and requires further investigation. More research is needed to determine whether this is a class effect or specific to semaglutide.”

A possible reason that some other studies have not found an association may be due to methodological differences, particularly in how NAION cases are identified, noted Joseph F. Rizzo III, MD, professor of ophthalmology and director of the Neuro-ophthalmology Service at Harvard Medical School in Boston.

Rizzo and colleagues were the first to conduct a study looking at this association in their own neuro-ophthalmology population, thereby guaranteeing that the diagnosis was correct but also introducing selection bias. Database studies rely on International Classification of Diseases, 10th Revision codes, and there isn’t one specifically for NAION — only for the broader category of ischemic optic neuropathy (H47.01), Rizzo told Medscape Medical News.

“There are ways to mitigate that effect, but they’re imperfect…so you may get different answers.” Rizzo previously conducted a similar large database study in people with T2D that also produced a twofold difference, although in a shorter follow-up time (6.7 months). Of the new study, he said, “It’s an active comparator study, and it’s beautiful. This SGLT2 approach is so good because GLP-1 and SGLT2 drugs tend to be used for patients with the same level of diabetes, so you’re minimizing residual confounding. I think it’s a great approach.”

Lee added that her team is looking at whether some patients “are at greater risk of NAION than others. To do this, we are refining our definition of NAION and evaluating the risks of other eye conditions that cause impaired vision. We are evaluating the risk of NAION related to other GLP-1s as well.”

Consider Ocular Risk Factors

“These [GLP-1] medicines have generally been really helpful for a lot of people. I never discourage patients from taking them,” said Rizzo.

“But if they’ve had visual loss for whatever reason…I really feel a moral responsibility to say, ‘Look, just so you’re aware, here’s what we found. The absolute risk is low, but I just want to inform you.’ Then they can make their own judgement about the level of risk.”

Lee noted that “clinicians should consider ocular risk factors such as preexisting optic nerve disease or prior NAION when prescribing.”

Rizzo extended that consideration to any patient who has experienced any type of vision loss in the past, such as due to glaucoma, diabetic retinopathy, or injury.

The study was conducted by the US Veterans Affairs with support from the VA Cooperative Studies Program. Lee reported having no further disclosures, and Rizzo reported having no relevant disclosures.

Miriam E. Tucker is a freelance journalist based in the Washington, DC, area. She is a regular contributor to Medscape, with other work appearing in the Washington Post, NPR’s Shots blog, and Diatribe. She is on X @MiriamETucker and BlueSky @miriametucker.bsky.social.

A version of this article first appeared on Medscape.com.