Study links GLP-1 diabetes drugs to a small rise in sudden vision loss

What happened

A study in Annals of Internal Medicine found a small increase in a rare optic nerve condition among people with type 2 diabetes who started GLP-1 drugs. The absolute difference was roughly three to four extra cases per 10,000 people over 18 months.

The headline going around, as first reported by ScienceDaily · Health & medicine (ScienceDaily · Health & medicine), names Ozempic and Wegovy. That is the outlet's shorthand. The paper studied the GLP-1 receptor agonist class as a whole.

This also is not brand-new. The ScienceDaily item is dated Oct 15, 2026 and rewrites a Rutgers press release. The paper went online ahead of print on July 14, 2026. Think of it as a fresh round of coverage rather than a fresh finding.

What the study actually found

The team (Reynolds, O'Malley, Roy and Dave) used a US commercial insurance claims database covering January 2017 to December 2022. They emulated a target trial, a method that tries to mimic a randomized experiment using real-world records. Their abstract says they included adults aged 18 to 65 with type 2 diabetes who started a GLP-1 drug, an SGLT2 inhibitor or a DPP4 inhibitor.

The outcome was ischemic optic neuropathy, used as a stand-in for NAION, a blockage-type injury to the optic nerve. The analysis adjusted for more than 80 characteristics using inverse probability weighting.

Against SGLT2 inhibitors, the risk difference was 3.0 per 10,000 (95% CI 0.4 to 5.7). That interval is wide and its low end sits close to zero. The numbers needed to harm were 3,333 and 2,778 for the two comparisons, respectively.

Among GLP-1 users, 81 cases occurred. Of those, 69 (85.2%) were in people over 50 and 57 (70.4%) were in men. Risk differences were smaller in people on metformin alone and larger in people taking two or more diabetes drugs.

The American College of Physicians release puts the scale in perspective: the condition occurred in fewer than 1 in 1,000 patients in every group.

What it does not show

Rutgers says in its press release that the study does not show GLP-1 drugs directly cause optic nerve damage. The journal's own release says the findings should be interpreted cautiously because of residual confounding, and may partly reflect differences in patients' underlying health.

• It is observational. People prescribed GLP-1 drugs can differ from those prescribed other diabetes drugs in ways that claims data miss, even after adjustment.
• It is about diabetes. Every participant had type 2 diabetes, was 18 to 65 and was commercially insured. Obesity-only Wegovy users, older adults and Medicare patients were not studied, even though ScienceDaily's text notes the drugs are widely used by people without diabetes.
• It is class-level. We could not confirm from the abstract whether results were broken out for Ozempic or Wegovy specifically, or whether tirzepatide (Mounjaro, Zepbound) was analyzed separately.
• The outcome is a proxy. A claims-coded diagnosis of ischemic optic neuropathy stands in for NAION, so some misclassification is possible.
• We could not access the full paper or supplement, so sample size and hazard ratios are not covered here.

One more correction: the ScienceDaily text garbles who was affected. The abstract says most GLP-1 cases were in people over 50 and in men.

What it means for you

If you take a GLP-1 drug, the odds of this happening to you are very low, and this study does not change that. It is still worth knowing what the problem looks like.

WHO describes NAION as typically sudden, painless vision loss in one eye that is generally irreversible, with no effective treatment. Its advice is to contact your doctor without delay if you notice sudden vision loss or rapidly worsening eyesight. WHO says semaglutide should be stopped if NAION is confirmed, and any decision to stop it belongs with your prescriber.

For US readers, we did not verify how the current Ozempic or Wegovy label addresses NAION. Check the prescribing information for the exact wording.

Context: this signal has been building

Regulators got here first. In June 2025, WHO relayed that the European Medicines Agency's safety committee concluded NAION is a very rare side effect of semaglutide (Ozempic, Rybelsus, Wegovy), affecting up to 1 in 10,000 users. The agency recommended updating the product information.

A companion nationwide cohort study was published alongside the US paper. It compared GLP-1 and SGLT2 inhibitor starters aged 35 to 84 with type 2 diabetes. The ACP summarizes it as showing a slightly higher NAION risk with GLP-1 drugs, with small absolute risks. We could not retrieve its effect sizes.

A Department of Veterans Affairs cohort used a similar design over three years, but we only saw its methods. A meta-analysis put the excess at about 1 additional case per 7,000 patients treated for a year, though it rates the certainty of evidence as low. It also notes that meta-analyses of randomized trials consistently reported fewer NAION events than observational studies did.

Most of the studies point the same way, with small numbers and unresolved questions about cause, though meta-analyses of randomized trials reported fewer events and the meta-analysis rates the evidence as low certainty. That combination calls for awareness, not alarm.