Ozempic and blindness (NAION): the research

In 2024, a single-center study out of Boston put a frightening idea into the headlines: that Ozempic and Wegovy might raise the risk of a rare type of vision loss called NAION. Since then, regulators on two continents have weighed in, large national databases have been mined, and clinical trial data has been pooled. The picture that has emerged is more nuanced than the first headlines suggested. This article walks through what NAION actually is, what each study found, where the evidence is strong and where it is shaky, and what the numbers mean for a real person deciding whether to start or stay on a GLP-1 drug.

What NAION Is, in Plain Terms

NAION stands for non-arteritic anterior ischemic optic neuropathy. Strip away the jargon and it means a small "stroke" of the optic nerve. The front part of the optic nerve loses its blood supply for long enough that nerve fibers are damaged. The result is sudden, usually painless vision loss in one eye, often noticed on waking. It does not typically cause total blindness, but it can leave a permanent blind spot or blur, most often in the lower half of the field of vision.

NAION is the most common cause of sudden optic nerve-related vision loss in people over 50. Even so, it is uncommon. In the general population, the yearly rate is roughly 2 to 10 cases per 100,000 people over age 50. There is no proven treatment that reliably restores vision once it happens, which is part of why any possible new risk factor gets serious attention.

The damage is not reversible in most cases. About 15 to 20 percent of people who get NAION in one eye will eventually get it in the other eye over the following years. That risk of the second eye being affected is a big reason doctors take the condition seriously.

Why Some People Are More Vulnerable

NAION almost always strikes eyes with a specific anatomy. The optic nerve head, where the nerve enters the eye, can be small and "crowded," with little or no central cup. Ophthalmologists call this a disc at risk. In a crowded disc, the nerve fibers are packed tightly inside a narrow channel in the sclera. When even minor swelling happens, there is no room for it. The swelling squeezes the fibers, chokes off more blood flow, and a self-feeding cycle of compression and ischemia sets in. A disc at risk is the single strongest predictor of who develops NAION.

On top of that anatomy, the classic vascular risk factors stack up: age over 50, type 2 diabetes, high blood pressure, high cholesterol, smoking, and obstructive sleep apnea. A sudden drop in blood pressure overnight can be the final trigger. That is why some cases of NAION are noticed on waking, after a night when blood pressure naturally dips. Certain blood-pressure medications taken at bedtime, sleep apnea that goes untreated, and even drugs like sildenafil have all been studied as possible contributors for the same reason. You can read more on the full background in StatPearls' overview of NAION.

It helps to know what NAION is not. It is not the same as a retinal detachment, a cataract, or the gradual blur of an uncorrected prescription. It is also distinct from the diabetic eye disease that GLP-1 drugs are sometimes blamed for. Diabetic retinopathy affects the retina's small blood vessels; NAION affects the optic nerve head. They can coexist, but they are different problems with different mechanisms. Keeping them separate matters, because the headlines often blur "eye damage" into one scary bucket when the evidence for each is different.

Notice something about that list. Type 2 diabetes and obesity are themselves NAION risk factors. And type 2 diabetes and obesity are exactly the conditions GLP-1 drugs treat. That overlap is the central problem in untangling whether the drug or the disease is doing the damage.

The Proposed Mechanism: Plausible but Unproven

Here is the honest starting point. Nobody has shown, at the level of cells and tissue, exactly how semaglutide could cause NAION. The mechanism is a hypothesis, not an established fact.

The leading idea goes like this. GLP-1 receptors exist in the retina and on the optic nerve. Rapid changes in blood sugar, blood pressure, or fluid balance after starting a GLP-1 drug could alter blood flow to the already-vulnerable optic nerve head. There is a known parallel with diabetes treatment in general: when blood sugar drops quickly after years of being high, diabetic retinopathy can temporarily worsen. Some researchers think a similar fast-correction effect could stress the optic nerve in people who already have a disc at risk.

That is biologically reasonable. It is not proof. Animal and lab studies have not nailed down a clear pathway, and the EMA's own review noted that non-clinical data did not establish a mechanism. So when you read that semaglutide "causes" NAION, treat that as shorthand. What the data actually shows is an association in some studies, with causation still open.

The Evidence, Study by Study

This is where being careful pays off. The studies do not all agree, and the type of study matters enormously. Observational studies (looking back at who took what and who got sick) can find associations but are prone to confounding. Randomized trials are the gold standard for proving cause, but NAION is so rare that trials struggle to catch enough cases. Here is the landscape.

Study	Type	Population	Key finding	Strength of evidence
Hathaway et al., JAMA Ophthalmology 2024	Retrospective single-center cohort	710 diabetes, 979 overweight/obese (Mass Eye and Ear)	HR 4.28 (diabetes), HR 7.64 (overweight/obese)	Weak: one specialty clinic, small numbers, referral bias
Danish national cohort, 2024	Population registry	424,152 people with type 2 diabetes	HR 2.19 (95% CI 1.54–3.12); risk more than doubled over 5 years	Stronger: nationwide, but observational
Danish-Norwegian cohort, 2025	Population registry, active comparator	61,377 semaglutide users vs SGLT-2i users	Pooled HR 2.81 (95% CI 1.67–4.75); absolute risk low	Stronger: active comparator design
TriNetX cohort, JAMA Ophthalmology 2025	Multi-center database	174,584 matched pairs with diabetes	HR ~2.4 at 2–3 years	Moderate: large, but database coding limits
Diabetes Care meta-analysis, 2026	Pooled randomized trials	83,288 participants, 20 RCTs	No significant association with optic nerve or vision-threatening events	Strongest design, but few NAION events

The Study That Started It: Hathaway 2024

The alarm came from a retrospective cohort at Massachusetts Eye and Ear, published in JAMA Ophthalmology in 2024. Researchers looked back at patients seen at their neuro-ophthalmology clinic. Among 710 people with type 2 diabetes, those prescribed semaglutide had a 36-month NAION rate of 8.9 percent versus 1.8 percent for those on other diabetes drugs, a hazard ratio of 4.28. Among 979 overweight or obese patients, the rate was 6.7 percent versus 0.8 percent, a hazard ratio of 7.64. Risk was highest in the first year.

Those hazard ratios are big and grabbed every headline. But read the fine print. This was a single specialty eye clinic, the kind of place people go precisely because they have eye problems. That creates referral bias. The total number of NAION events was tiny, just 17 and 20 cases in the semaglutide groups. The authors themselves wrote that because the study was observational, "future study is required to assess causality." It was a flag, not a verdict.

The National Registries: Bigger and Harder to Dismiss

The single-clinic concern would mean little if larger, population-wide data showed nothing. It did not. A Danish national cohort of 424,152 people with type 2 diabetes found that once-weekly semaglutide more than doubled the five-year risk of NAION, with a hazard ratio of 2.19 (95% CI 1.54–3.12). There were 67 cases among users and 151 among non-users. The incidence was 0.228 per 1,000 person-years on semaglutide versus 0.093 without it. The median time from starting the drug to the event was about 22 months.

A follow-up Danish-Norwegian cohort study, published in 2025, used a smarter design. Instead of comparing semaglutide to a grab-bag of other drugs, it compared it head-to-head against SGLT-2 inhibitors, another modern diabetes drug used in similar patients. That helps control for how sick the patients are. The pooled hazard ratio was 2.81 (95% CI 1.67–4.75). The incidence difference was about 1.4 extra cases per 10,000 person-years. The authors' bottom line is worth quoting plainly: semaglutide "is associated with an increased risk of NAION compared with the use of SGLT-2is. However, the absolute risk remains low."

A separate large US database study using TriNetX, also in JAMA Ophthalmology in 2025, matched 174,584 semaglutide users to non-GLP-1 users and found roughly a 2.4-fold higher risk at two to three years. So three independent large datasets, on two continents, landed in the same general range: roughly two to three times the relative risk, off a very small base rate.

The Counterweight: Pooled Trial Data Found Nothing

Now the part that the scary headlines usually skip. When researchers pooled the randomized controlled trials, the signal disappeared. A 2026 meta-analysis in Diabetes Care combined 20 randomized trials with 83,288 participants. It found GLP-1 receptor agonists were not associated with a higher risk of optic nerve or vision-threatening events, including ischemic optic neuropathy, blindness, or visual impairment. The result held across every subgroup.

How can both be true? A few reasons. NAION is so rare that even 83,000 trial participants produce very few events, so trials may simply be underpowered to detect a small absolute increase. Trial populations are also healthier and more closely watched than the real-world patients in registries. And observational studies, no matter how large, can never fully rule out that the underlying disease, not the drug, drives the risk. The trials are the strongest study design for proving cause, but on this specific rare outcome they may lack the statistical muscle. That tension is the real state of the science.

What Regulators Concluded

Drug regulators do not need proof of causation to act. They weigh the signal. And here, two major bodies reached the same measured conclusion.

In June 2025, the European Medicines Agency's safety committee finished its review. After looking at lab studies, clinical trials, post-marketing reports, and the published literature, it concluded that NAION is a "very rare" side effect of semaglutide. In regulatory terms, "very rare" means it may affect up to 1 in 10,000 people taking the drug. The EMA recommended that the product information for Ozempic, Rybelsus, and Wegovy be updated to list NAION.

The World Health Organization issued its own advisory in June 2025, aligning with the EMA and adding NAION to semaglutide's risk management plan based on case reports from multiple countries. Both bodies gave the same practical guidance: if you suddenly lose vision or your eyesight worsens fast while on a semaglutide drug, contact your doctor right away, and if NAION is confirmed, the drug should be stopped.

Note the careful wording. "Very rare." "Associated." Not "causes." Regulators added a warning while explicitly keeping causation an open question. That is the appropriate level of confidence for the current evidence.

Body	Date	Conclusion	Frequency label
EMA (PRAC)	June 2025	NAION is a side effect of semaglutide; update product information	Very rare (up to 1 in 10,000)
WHO	June 2025	Added NAION to semaglutide risk management plan	Very rare
US FDA	As of mid-2026	Reviewing data; had not mandated a NAION boxed warning	Not formally classified

Putting the Numbers in Perspective

Relative risk and absolute risk are not the same thing, and conflating them is where most of the fear comes from. A hazard ratio of 2 to 3 sounds huge. But it is multiplying a tiny baseline.

Work the Danish national numbers. The rate on semaglutide was about 0.228 per 1,000 person-years. That is roughly 1 extra case for every 10,000 people taking the drug for a year, on top of a background rate that is already low. The EMA's "up to 1 in 10,000" framing says the same thing. For comparison, the gastrointestinal side effects of these drugs, the nausea and the rest, affect a large share of users. NAION affects a sliver of a percent.

So a fair summary is this: the relative increase is real and consistent across several large studies, but the absolute risk to any one person is small. For most people, the cardiovascular, metabolic, and weight benefits of these drugs, which are well established in large trials, are likely to outweigh a rare eye risk. The math changes if you personally carry extra NAION risk factors.

There is one more honest caveat about the evidence base. None of the large studies that found a signal were randomized. They mined registries and electronic records, which means the people who got semaglutide may have differed from those who did not in ways the data could not fully capture. Maybe they were sicker, or saw eye doctors more often, or had their NAION coded differently. The Danish-Norwegian study tried to blunt this by comparing against SGLT-2 inhibitor users rather than untreated patients, which is a real strength, but even an active comparator cannot erase every difference. Meanwhile, the studies that were randomized, the gold standard for causation, did not find a significant signal. So the cleanest reading is that there is a consistent association in observational data, a plausible but unproven mechanism, and an unresolved gap between what registries suggest and what trials confirm. Anyone who tells you the question is settled, in either direction, is overstating the evidence.

It is also worth naming the conflict-of-interest landscape, because it cuts both ways. Several of the reassuring meta-analyses include authors with industry ties or were conducted in the period when manufacturers were facing litigation, which is a reason to read their conclusions critically. At the same time, the alarming single-clinic study had its own limits and generated a wave of lawsuits, which creates incentives to overstate harm. The most trustworthy anchors are the independent national registries and the regulators, who have no product to sell and concluded the same careful thing: a real but very rare association.

Who Should Be More Cautious

The evidence does not support panic, and it does not support ignoring the issue either. Some people sit higher on the risk curve and deserve a real conversation with their doctor.

People who have already had NAION in one eye. They face a meaningful risk to the second eye regardless of any drug, and most specialists would avoid adding a possible risk factor.
People with a known "disc at risk." If an eye exam has shown a small, crowded optic disc, the underlying vulnerability is already there.
People with stacked vascular risk factors. Poorly controlled blood pressure, sleep apnea, smoking, and significant cardiovascular disease all add up.
Anyone who has had unexplained vision loss. This warrants a neuro-ophthalmology look before starting.

A baseline eye exam before starting is reasonable for higher-risk people, though no guideline currently mandates one for everyone. If you are generally healthy and starting a GLP-1 drug for weight or diabetes, the practical move is awareness: know the warning sign and act on it fast.

Alternatives and Whether They Are Safer

A natural question is whether switching drugs sidesteps the risk. The honest answer is that the data is thinner for the alternatives, which is not the same as proving they are safe.

Tirzepatide (Mounjaro and Zepbound) is a dual GIP/GLP-1 agonist. Because it also hits the GLP-1 receptor, the theoretical mechanism could apply, but the large NAION cohort studies focused on semaglutide specifically. There is not yet a comparable body of NAION evidence for tirzepatide, so "no signal found" partly reflects "not enough study done." The head-to-head comparison of semaglutide and tirzepatide covers their broader differences.

SGLT-2 inhibitors served as the comparison drug in the Danish-Norwegian study and showed a lower NAION rate, but they treat diabetes, not obesity, and are not interchangeable for weight loss. For diabetes specifically, they are a reasonable alternative to discuss. Switching from semaglutide to another agent is a decision to make with a clinician who can weigh your full picture, not a blanket fix. For a fuller map of the trade-offs, see the complete GLP-1 side effects guide and the broader GLP-1 side effects and risks overview.

Warning Signs You Should Never Ignore

NAION is sudden and painless, which is exactly why it is easy to dismiss at first. The vision change does not hurt. Knowing the signs is the most useful thing you can take from this article.

Sudden loss or dimming of vision in one eye, often noticed on waking
A blind spot or curtain, frequently in the lower half of your vision
Colors looking washed out or dimmer in one eye
A blurry or "shaded" area that does not clear

If any of these happen while you are on a semaglutide medication, treat it as an emergency. Contact your doctor or seek urgent eye care the same day. The same advice applies whether or not you think it is related to the drug, because other causes of sudden vision loss are also emergencies. This overlaps with the broader Wegovy blindness risk discussion and the full Ozempic side effects review.

Frequently Asked Questions

Does Ozempic cause blindness?

No study has proven that Ozempic causes blindness, and NAION rarely causes total blindness even when it happens. Several large studies link semaglutide to a roughly two- to three-fold higher risk of NAION, a rare optic nerve condition, but pooled randomized trials found no significant association. Regulators label it a "very rare" side effect affecting up to 1 in 10,000 users. The absolute risk is small, and causation is not settled.

How common is NAION in people taking semaglutide?

Very uncommon. In a Danish national cohort, the rate was about 0.228 cases per 1,000 person-years on semaglutide, which works out to roughly 1 extra case per 10,000 people per year compared with non-users. The EMA describes it as affecting up to 1 in 10,000 people taking the drug. By contrast, common GLP-1 side effects like nausea affect a large share of users.

Should I stop my GLP-1 medication because of the eye risk?

Not on your own. For most people the proven heart, metabolic, and weight benefits outweigh a rare eye risk. Stopping suddenly without a plan can let blood sugar or weight rebound. The exception is if you develop sudden vision changes, in which case you should seek care immediately and the drug may be stopped. Discuss your personal risk factors with your doctor before making any change.

Who is most at risk for NAION on these drugs?

People who already have a "disc at risk" (a small, crowded optic nerve), anyone who has had NAION in one eye, and those with stacked vascular risk factors such as poorly controlled blood pressure, sleep apnea, and smoking. Type 2 diabetes and obesity themselves raise NAION risk, which makes it hard to separate the disease from the drug.

Is the NAION risk the same for Wegovy, Ozempic, and tirzepatide?

Wegovy, Ozempic, and Rybelsus all contain semaglutide, and the EMA's warning covers all three. The large NAION studies focused on semaglutide, so the evidence is strongest there. Tirzepatide (Mounjaro, Zepbound) also acts on the GLP-1 receptor, so the theoretical risk could apply, but it has not been studied as thoroughly for NAION, meaning we have less data rather than proof of safety.

This article is for general information only and is not medical advice. Talk to a qualified healthcare provider before starting, stopping, or changing any medication, and seek urgent care for any sudden change in vision.