Semaglutide, NAION, and a retracted study

A paper published in the journal Retina in late 2025 asking whether semaglutide offers therapeutic benefit or poses an ocular risk has been retracted^[1]. The retraction notice, authored by Suhayya Batool and published in the August 2026 issue of Retina (PMID 42468029), formally withdraws the original article (DOI: 10.1097/IAE.0000000000004709). No stated rationale appears in the publicly available PubMed record. Retractions without disclosed reasons leave a gap in the scientific record, and this one lands in the middle of an active, contested debate about GLP-1 receptor agonists and optic nerve safety.

What non-arteritic anterior ischemic optic neuropathy is

Non-arteritic anterior ischemic optic neuropathy, usually abbreviated NAION, is the most common acute optic nerve disease in adults over 50. It occurs when blood supply to the anterior portion of the optic nerve is interrupted, typically causing sudden, painless vision loss in one eye. The disc appears swollen on fundoscopy. Risk factors include a structurally small optic disc, hypertension, dyslipidaemia, diabetes, obstructive sleep apnoea, and smoking. Unlike the arteritic form (giant-cell arteritis), there is no systemic inflammation. Cases of atypical optic nerve injury following semaglutide have also been described -- see our coverage of retrobulbar optic neuropathy after semaglutide escalation for a different presentation pattern.

How semaglutide entered the NAION conversation

The signal emerged formally in a 2024 retrospective study by Hathaway et al. at Massachusetts Eye and Ear, published in JAMA Ophthalmology. Among neuro-ophthalmology patients with type 2 diabetes, those prescribed semaglutide had a hazard ratio for NAION of 4.28 (95% CI 1.62-11.29) compared with those on non-GLP-1 receptor agonist treatments. In the overweight or obese group the hazard ratio was 7.64 (95% CI 2.21-26.36)^[2]. The absolute numbers were small (17 and 20 NAION events in the semaglutide arms, respectively), and the study was retrospective, single-centre, and enriched by selecting from patients already attending a neuro-ophthalmology service. The authors explicitly stated that future study was required to assess causality.

A 2025 retrospective cohort analysis by Hsu et al. in JAMA Ophthalmology, using the TriNetX database and 174,584 diabetic patients, found no statistically significant NAION risk increase at one year or under, but hazard ratios of 2.39 at two years, 2.44 at three years, and 2.05 at four years^[3]. A 2026 systematic review and meta-analysis by Lampsas et al., drawing on eight retrospective cohorts totalling over 14 million participants, calculated a pooled hazard ratio of 3.36 (95% CI 1.44-7.84) at one-year follow-up and 2.37 (95% CI 1.46-3.85) at two years^[4]. Both studies used real-world data, which limits their ability to separate the drug effect from the underlying disease profile of people taking semaglutide.

Evidence pulling the other way

Against those observational signals sits a 2026 analysis by Vilsboll et al. in the British Journal of Ophthalmology, which pooled safety data from randomised placebo-controlled trials of liraglutide and semaglutide involving 96,829 participants and 205,777 participant-years of observation. Confirmed NAION cases: three in the GLP-1 receptor agonist group, five in the placebo group, giving event rates of approximately three and six per 100,000 participant-years respectively^[5]. The trial evidence, covering longer exposure in more controlled settings, showed no elevated risk. The contradiction between observational and trial data has not been resolved.

The current consensus position

In May 2026, the North American Neuro-Ophthalmology Society and the American Academy of Ophthalmology published a joint consensus statement in Ophthalmology (DeParis et al.). The statement acknowledged possible association reports but noted that other studies found no correlation and that the overall magnitude of NAION risk remains low. The societies stopped short of recommending against prescribing GLP-1 receptor agonists, instead calling for shared decision-making between each patient and their care team^[6].

Why a retraction matters here

When a paper is retracted without a publicly disclosed reason, readers and researchers lose the ability to evaluate whether its findings should be discarded entirely, or whether some specific element (a methodological flaw, a data error, an authorship dispute) invalidates only part of its claims. In a rapidly evolving safety debate, unpublished retractions make the literature harder to read. Anyone synthesising evidence on semaglutide and NAION now has to treat the 2025 Batool paper as withdrawn from the record without knowing exactly why.

Retraction Watch and PubMed both index retraction notices, which helps. But the absence of an explanatory note from the journal is a gap the field would benefit from filling. Retina (the publisher) has not, as of this writing, released a statement beyond the notice itself.

What this does not change

The retraction removes one paper from the semaglutide-NAION literature. It does not invalidate the Hathaway, Hsu, Lampsas, or Vilsboll datasets, which are from independent research groups. The joint consensus statement from the two ophthalmology societies was published after the original Batool article and does not cite it as a key source. The central clinical question -- whether semaglutide meaningfully raises NAION risk above background rates -- remains open and contested. A 2026 case report in BMC Ophthalmology describing bilateral NAION and branch retinal artery occlusion in a 37-year-old man after 11 months of semaglutide use continues to add to the case-report literature^[7].

What we do not yet know

The mechanism by which semaglutide could theoretically affect optic nerve perfusion is not established. Whether specific patient populations (small optic disc, poorly controlled vascular risk factors) face disproportionate risk is not resolved. Whether the NAION association, if real, is a class effect shared by all GLP-1 receptor agonists or specific to semaglutide is under study. And the reason for the Batool retraction itself is publicly unknown.

If you are prescribed semaglutide and have any concern about vision changes, raise them with your prescriber. This article is for educational purposes and does not substitute for advice from a qualified healthcare professional. Any decision to start, continue, or stop a prescribed medicine belongs with a clinician who knows your individual circumstances.