Does Semaglutide or Tirzepatide Increase the Risk of NAION — and What Is the Absolute Excess Risk Over 18 Months in 2026?

Current evidence indicates that semaglutide is associated with a statistically significant increase in the risk of nonarteritic anterior ischemic optic neuropathy (NAION), particularly in patients with type 2 diabetes or obesity. The absolute excess risk is approximately 3–4 additional cases per 10,000 treated patients over 18 months. Tirzepatide carries a similar signal. Causality has not been established.

What Is NAION and Why Does It Matter for GLP-1 Therapy Patients?

NAION is the most common acute optic neuropathy in adults over 50, caused by ischaemia of the short posterior ciliary arteries supplying the anterior optic nerve head. It presents as sudden, painless, monocular vision loss with an altitudinal visual field defect. No proven treatment exists, and roughly 40% of affected eyes retain permanent visual impairment.

The condition's annual background incidence ranges from 2 to 10 per 100,000 individuals in the general population, rising substantially in patients with diabetes, hypertension, obstructive sleep apnoea, and a structurally crowded optic disc — the so-called "disc at risk." Because GLP-1 receptor agonists are now prescribed to tens of millions of patients globally, even a modest relative risk increase translates to a clinically meaningful absolute number of cases.

The anatomical substrate for NAION is a small cup-to-disc ratio (below 0.2), which creates a compartment-syndrome-like environment when the optic disc swells. Ischaemia in this crowded space can propagate axonal damage rapidly. Patients with pre-existing vascular risk factors — diabetes, hypertension, dyslipidaemia — carry the highest baseline susceptibility.

What Did the Hathaway 2024 JAMA Ophthalmology Study Find?

The Hathaway 2024 matched cohort study of 16,827 patients found that semaglutide was associated with a hazard ratio of approximately 4·28 for NAION in type 2 diabetes patients and approximately 7·64 in patients with overweight or obesity, versus non-GLP-1 comparators. Cumulative incidence at 36 months reached 8·9% in the diabetes semaglutide cohort versus 1·8% in controls.

The study was retrospective and conducted at a single academic neuro-ophthalmology practice, which introduces substantial referral bias. Patients presenting to a specialist service are not representative of the broader semaglutide-treated population. The absolute incidence figures in this cohort are therefore far higher than population-level estimates.

Despite these constraints, the study was the first to quantify the association with a matched comparator design. The hazard ratios remained significant after adjustment for age, sex, diabetes status, hypertension, and body mass index. The signal was not attenuated by removing patients with pre-existing optic disc pathology, suggesting the association is not entirely explained by confounding by indication.

What Did the Wang 2025 JAMA Network Open TriNetX Cohort Add?

Wang and colleagues (2025), using the TriNetX federated database of over 37 million patients, found that both semaglutide and tirzepatide were associated with increased risk of optic nerve disorders including NAION in type 2 diabetes patients. This study extended the signal to tirzepatide and confirmed the association in a far larger, less-selected population than Hathaway 2024.

The TriNetX analysis reported an absolute incidence difference of approximately 1·5 additional NAION cases per 10,000 patients with type 2 diabetes over the study period for semaglutide versus SGLT-2 inhibitor comparators. The tirzepatide signal was directionally consistent but had wider confidence intervals, reflecting shorter market exposure and smaller treated numbers at the time of data extraction.

Importantly, Wang and colleagues found no significant NAION signal in the obesity-without-diabetes subgroup when using SGLT-2 inhibitors as the active comparator. This suggests that diabetes itself — and its associated microvascular disease — may be a necessary co-factor rather than semaglutide acting independently on optic nerve vasculature.

What Is the Absolute Excess Risk Over 18 Months?

Synthesising across the Hathaway 2024 cohort, the Wang 2025 TriNetX analysis, and the Chrzanowski 2026 PLOS Medicine meta-analysis, the best current estimate is approximately 3–4 additional NAION cases per 10,000 treated type 2 diabetes patients over 18 months. The EMA's 2025 PRAC review classified NAION as a "very rare" adverse reaction, defined as affecting up to 1 in 10,000 patients.

The 3–4 per 10,000 figure must be contextualised against the background rate. NAION affects roughly 2 to 10 per 100,000 individuals annually in the general population. The absolute baseline over 18 months is therefore approximately 0·3 to 1·5 per 10,000 — meaning the excess risk represents a two- to tenfold increase over background, concentrated in patients with pre-existing vascular risk factors.

The Chrzanowski 2026 PLOS Medicine meta-analysis pooled observational studies and RCT data and found a possible association in T2DM patients. RCT data showed only four NAION events in semaglutide arms versus controls — insufficient to confirm causality. The pooled observational hazard ratio from Chen and colleagues (2025) was approximately 2·6 (95% CI: 1·8 to 3·8), consistent with the Hathaway point estimate after accounting for referral-bias inflation.

Is the Tirzepatide Signal as Strong as Semaglutide?

The tirzepatide NAION signal is directionally consistent with semaglutide but currently weaker in statistical precision. Wang and colleagues (2025) identified an elevated optic nerve disorder risk for tirzepatide in T2DM patients, but tirzepatide's shorter commercial history means the exposed-patient-years are substantially fewer. No dedicated tirzepatide NAION cohort study has been published independently of the Wang analysis.

Mechanistically, tirzepatide activates both GLP-1 and GIP receptors, whereas semaglutide is a selective GLP-1 receptor agonist. If the NAION signal is mediated through GLP-1 receptor activation — whether via haemodynamic effects, intracranial pressure modulation, or optic nerve vasculature — tirzepatide would be expected to carry a comparable or slightly attenuated risk. No head-to-head ophthalmic safety comparison between the two agents has been published.

The EMA's June 2025 PRAC review and the UK MHRA Drug Safety Update of February 2026 both addressed semaglutide specifically; neither regulatory body has yet issued a parallel label update for tirzepatide. This asymmetry reflects the evidence timeline rather than a confirmed safety difference between the two molecules.

What Mechanisms Have Been Proposed to Explain the Association?

Three mechanistic hypotheses have been advanced: rapid weight loss may transiently alter intracranial pressure dynamics affecting optic nerve perfusion; GLP-1 receptors expressed on optic nerve vasculature may directly modulate posterior ciliary artery autoregulation; and disc-at-risk anatomy may interact with haemodynamic fluctuations during early treatment. None of these hypotheses has been confirmed in human mechanistic studies.

The intracranial pressure hypothesis draws on evidence that GLP-1 receptor agonists reduce intracranial pressure in patients with idiopathic intracranial hypertension. However, rapid ICP reduction in patients with a crowded optic disc could theoretically alter the pressure gradient across the lamina cribrosa, transiently compromising axoplasmic flow. This remains speculative and has not been tested prospectively.

The direct vascular hypothesis is supported by the known expression of GLP-1 receptors in vascular endothelium. Acute GLP-1 receptor activation alters nitric oxide bioavailability and vascular tone in multiple vascular beds. Whether this extends to the short posterior ciliary arteries — the vessels whose occlusion causes NAION — has not been demonstrated in human ophthalmic studies.

Confounding by indication remains a competing explanation. Patients prescribed semaglutide or tirzepatide carry higher baseline vascular risk than comparator populations on metformin or SGLT-2 inhibitors. Residual confounding is difficult to eliminate in observational designs, and the AAO/NANOS consensus statement explicitly states that causality has not been established.

How Have Regulators Responded to the NAION Signal?

The EMA's PRAC concluded in June 2025 that NAION should be listed as a "very rare" adverse reaction in the product information for Ozempic, Rybelsus, and Wegovy. The UK MHRA issued a Drug Safety Update in February 2026 warning of a likely association. The US FDA had not issued a parallel label update for semaglutide or tirzepatide as of mid-2026.

The EMA review also recommended that patients experiencing sudden vision loss or rapidly deteriorating vision while taking semaglutide should be evaluated urgently by an ophthalmologist. If NAION is confirmed, European guidance suggests that the treating physician consider whether continued semaglutide use is appropriate — a risk-benefit reassessment rather than an automatic discontinuation mandate.

The AAO and the North American Neuro-Ophthalmology Society issued a joint consensus statement in 2025 explicitly stating that routine pre-treatment ophthalmic examination is not recommended solely to assess NAION risk before initiating a GLP-1 receptor agonist. The societies noted that the absolute risk remains low and that the cardiovascular and metabolic benefits of these agents are well-established in high-risk populations.

What Safety Monitoring Is Warranted for Patients on GLP-1 Receptor Agonists?

Current AAO/NANOS consensus does not support mandatory pre-treatment ophthalmic screening for all GLP-1 receptor agonist users. Clinicians should document any history of prior NAION or crowded optic disc anatomy, as these represent the highest-risk substrate. Patients should be counselled to report sudden, painless monocular vision loss or new visual field defects immediately.

Patients with established risk factors — a known "disc at risk" on prior fundoscopy, prior NAION in the fellow eye, uncontrolled hypertension, or severe obstructive sleep apnoea — represent a subgroup where a baseline ophthalmic assessment may be clinically prudent before initiating therapy, even absent a formal guideline recommendation. This is a clinical judgment, not a regulatory requirement.

The safety profile of semaglutide and tirzepatide across their full indication spectrum remains strongly net-positive. The SELECT trial demonstrated a 20% relative reduction in major adverse cardiovascular events with semaglutide 2·4 mg in high-risk patients. The NAION signal, while warranting vigilance, does not alter the overall risk-benefit calculus for the vast majority of indicated patients based on current evidence.

Any patient who develops sudden monocular vision loss while receiving a GLP-1 receptor agonist should be referred urgently to ophthalmology. NAION is diagnosed clinically by the characteristic altitudinal field defect and optic disc oedema on fundoscopy; optical coherence tomography can confirm the diagnosis. There is no proven pharmacological treatment for acute NAION. Does the Epidemiological Signal for Semaglutide and Dementia Risk Survive Methodological Scrutiny in 2026? What Does 2026 Research Reveal About Semaglutide's Oncogenic Potential and Cardiotoxicity Mitigation Beyond Glycemic Control? What Do Semaglutide's 2026 Oncogenic and Cardiotoxicity Data Mean for Metabolic Performance Users?