Study links GLP-1 drugs to rare vision loss that usually lasts

A Rutgers-led study has linked GLP-1 medicines to a small increase in a potentially permanent vision-loss condition, putting a rare safety concern beside these drugs' expanding use. ^[1][2]

The researchers examined US adults aged 18 to 65 with Type 2 diabetes who had recently started medication. Chintan Dave, a Rutgers researcher involved in the study, estimated approximately three to four additional cases of ischemic optic neuropathy per 10,000 patients treated with GLP-1 medicines over 18 months. ^[1][2]

That is a small difference with a serious consequence. The condition involves reduced blood flow to the optic nerve, the connection that carries visual information from the eye to the brain. It can cause sudden loss of vision, and the impairment often lasts. ^[1][2]

The medicines considered include semaglutide, sold as Ozempic and Wegovy, and tirzepatide, sold as Mounjaro and Zepbound. Their reach is already substantial. KFF found that two million people enrolled in the US Medicare Part D prescription-drug programme took Ozempic in 2024, compared with fewer than 150,000 in 2019. That count covers one programme, not all users. ^[1][2][4]

The right response is proportionate vigilance. A rare threat to sight deserves a clear place in conversations about treatment, while the decision to use these drugs must still weigh their potential benefits.

A small risk does not become a large individual risk because a medicine is popular. It does become a public-health question worth answering carefully.

Three to four extra cases, not the total risk

The study, published in Annals of Internal Medicine, used a large US database of adults starting treatment for Type 2 diabetes. Its central finding was an association between GLP-1 use and a higher risk of ischemic optic neuropathy. ^[1][2]

The estimate needs to retain all its parts: approximately three to four additional cases, among 10,000 patients treated, over 18 months. Expressed as a difference in percentages, that is about 0.03 to 0.04 percentage points. It is neither the total number of cases nor the total probability that a patient taking one of these medicines will develop the condition. ^[1][2]

"Additional" does much of the work. The researchers were describing a difference between treatment groups. People can develop the condition without taking these medicines, so the question is how much more often it appeared among those receiving them. ^[1][2]

There are important boundaries to that finding. The study establishes an association, not causation, and the researchers said differences in patients' underlying health could help explain it. Its reported population had Type 2 diabetes, so the supplied accounts do not establish the risk for people without diabetes taking these medicines solely for weight loss.

The full journal paper was not accessible for this dossier, and the institutional summaries do not provide the sample size, case counts, comparison medicines or detailed methods. ^[1][2]

Those gaps prevent a firm verdict on how convincing the estimated difference is. In an observational comparison, the task is to separate the effect of treatment from the reasons people received that treatment. If the groups begin with different health risks, a later difference in outcomes may reflect some of those starting conditions.

This is the strongest objection to treating the finding as a drug warning: the medicines may not be responsible for the excess cases. It is a substantial objection. It should govern the wording of the finding and the next research questions. It does not make the observed association worthless.

A safety signal earns attention before it earns a causal explanation. The sensible course is to investigate it without asking patients to mistake a research finding for a diagnosis.

What happens when the optic nerve loses blood flow

Ischemic optic neuropathy is a blood-supply problem affecting the optic nerve. The eye gathers visual information, but the nerve must carry it onward to the brain. Damage to that connection can impair sight even when the problem is not in the eye's focusing system. ^[1][2]

Rutgers says more than four in five cases are estimated to involve the non-arteritic anterior form. That form can produce blurred or dimmed vision, or the loss of part of the visual field in one eye. It does not usually cause total blindness. ^[1][2]

For a patient, however, losing part of the view is hardly a minor outcome. A risk estimate describes frequency. It does not describe how disruptive the event is to the person who experiences it.

According to Rutgers, the vision loss is usually permanent. About one-third of affected patients may experience some improvement, generally within six months, but some impairment often remains. Improvement and recovery are different promises. ^[1][2]

That persistence gives the finding its weight. A brief and reversible side effect can be assessed differently from damage that may remain after the immediate event has passed. Here, the low estimated frequency sits beside the possibility of lasting loss.

Dave said most cases in the study occurred among men or people aged 50 to 65. He also said prompt assessment by an eye specialist may be important for patients with higher baseline risk. Rutgers advises prompt clinical evaluation for sudden visual changes. ^[1][2]

That is the practical message patients can use now. Sudden blurring, dimming or a missing part of the visual field deserves attention. Turning the finding into a reason to panic would be poor risk communication. Leaving patients unaware of the symptom would be poor risk communication too.

Scale changes the safety task

Ozempic use among Medicare Part D enrollees grew from fewer than 150,000 people in 2019 to two million in 2024. That is more than a 13-fold increase in five years within the population KFF measured. ^[4]

The growth demonstrates reach, but it cannot supply a case count for the Rutgers finding. The Medicare figure describes a different population and period from the study of adults aged 18 to 65 starting diabetes treatment. Multiplying the study's estimated difference by those two million users would produce a neat number with an untidy foundation. ^[1][2][4]

The broader point needs no such multiplication. As more people receive a treatment, health systems have more reason to look closely for uncommon adverse outcomes. Prescribing at scale requires safety monitoring at scale, including for harms too rare to dominate an individual consultation.

The expansion into obesity treatment also broadens the decisions these medicines serve. In its December 1, 2025 guidance, the World Health Organization supported GLP-1 therapies for long-term treatment of obesity in adults, excluding pregnant women, as part of comprehensive care. WHO describes obesity as a chronic disease. ^[3]

That matters because obesity treatment should not be dismissed as mere body reshaping. WHO recognizes a medical role for these drugs. The benefit side of the decision deserves the same seriousness as the safety side. ^[3]

But a treatment used across different populations needs evidence that follows it into those populations. A diabetes study is valuable evidence for diabetes care. Extending its numerical estimate to everyone seeking weight loss would substitute assumption for measurement.

WHO's recommendation was conditional partly because of limited data on long-term efficacy, safety, maintenance and discontinuation, alongside cost and equity concerns. Its endorsement therefore came with an explicit demand for a broader view of treatment than the prescription alone. ^[3]

The Rutgers finding fits that unfinished safety agenda. It identifies a specific outcome to examine, with an estimated difference small enough to require careful measurement and serious enough to matter.

Benefits, access and the next test

There is another scale problem. WHO said that even with rapid expansion in production, GLP-1 therapies were projected to reach fewer than 10% of people who could benefit by 2030. ^[3]

A treatment can have millions of users and still reach only a small share of those who might benefit. Safety and access are therefore simultaneous responsibilities. Treating them as competing slogans would serve patients badly.

The case for scrutiny is strongest when it stays precise: a reported excess of a rare optic-nerve condition, in adults with Type 2 diabetes, over 18 months. The case for continued access rests on the recognized treatment role of these medicines and the need to judge benefit and risk for the patient in front of the clinician. ^[1][2][3]

The finding should sharpen that judgement. It should encourage clear symptom advice, attention to baseline risk and better evidence, rather than a blanket conclusion about an entire group of medicines.

The next scientific task is to test whether the association persists in further studies, determine whether treatment causes the excess, and identify who is most at risk. Research in people receiving these medicines for obesity without diabetes is especially important. That is where the expanding use of the drugs now demands its own answer.