Highlights
- Target trial emulations in US Veterans and UK primary care populations suggest a 2-fold to 2.5-fold increased risk of nonarteritic anterior ischemic optic neuropathy (NAION) following semaglutide initiation.
- Meta-analyses of randomized controlled trials (RCTs) currently show no statistically significant increase in NAION, likely due to the rarity of the event and the fact that vision outcomes were not prespecified endpoints.
- Higher risks appear associated with rapid HbA1c reduction (>1%), specific formulations (Wegovy vs. Ozempic), and male sex.
- Comparative data suggest tirzepatide may have a more favorable ocular safety profile regarding NAION risk compared to non-GIP-acting GLP-1 RAs.
Background
Glucagon-like peptide-1 receptor agonists (GLP-1 RAs), particularly semaglutide, have transformed the management of type 2 diabetes (T2D) and obesity. By mimicking the incretin hormone GLP-1, these agents promote glucose-dependent insulin secretion, delay gastric emptying, and suppress appetite, leading to significant cardiometabolic benefits. However, as global utilization has surged into the tens of millions, rare but vision-threatening safety signals have emerged.
Nonarteritic anterior ischemic optic neuropathy (NAION) is a leading cause of sudden, painless vision loss due to infarction of the short posterior ciliary arteries supplying the optic nerve head. While its pathogenesis remains incompletely understood, it is often linked to vascular risk factors and anatomical “crowding” of the optic disc. Since July 2024, multiple observational studies have suggested an association between semaglutide and NAION, prompting clinicians to re-evaluate the ocular safety of these essential medications.
Key Content
The US Veterans Target Trial Emulation
A pivotal study by Heberer et al. (2026) conducted a target trial emulation within the Veterans Health Administration (VHA) system, comparing semaglutide initiators with sodium-glucose cotransporter-2 inhibitor (SGLT2i) initiators. Among 102,361 veterans, the study utilized overlap weighting to balance significant baseline confounders, including BMI and baseline HbA1c.
The results demonstrated a hazard ratio (HR) of 2.33 (95% CI, 1.54-3.54; P < .001) for incident NAION over a median follow-up of 2.1 years. While the relative risk was substantially elevated, the absolute risk remained low, with an incidence of 123 per 100,000 person-years in the semaglutide group versus 67 per 100,000 in the SGLT2i group. This large-scale analysis reinforces the concern that semaglutide initiation may be a discrete risk factor for NAION in populations with high metabolic burden.
Global Real-World Evidence and Formulation Specifics
Data from the TriNetX US Collaborative Network (2015–2024) further supports this association. In a study of 799,036 matched pairs, GLP-1 RA users showed a higher cumulative incidence of NAION (0.21%) compared to other antidiabetic agents (0.17%; HR 1.38). When specifically compared to SGLT2i users, the HR was 1.30. Notably, Kaplan-Meier curves in this study showed an early separation, suggesting that the risk may manifest shortly after treatment initiation.
Further granularity was provided by pharmacovigilance analyses of the FDA Adverse Event Reporting System (FAERS). Analysis of over 30 million reports identified that the obesity-specific formulation of semaglutide (Wegovy) carried a significantly higher reporting odds ratio (ROR = 74.89) for ischemic optic neuropathy compared to the diabetes-specific formulation (Ozempic; ROR = 18.81). This suggests a potential dose-response relationship, as Wegovy typically involves higher maintenance doses (2.4 mg) than Ozempic (1.0–2.0 mg).
Discrepancies Between Observational Data and RCTs
In contrast to real-world evidence, a meta-analysis of 20 randomized controlled trials involving 83,2
**Organizing the Content**
I am now extracting all study results into my “Content” section. I will then categorize this information into logical sections, using headers such as “The Veteran Health Administration Target Trial” and “Global Real-World Evidence.” I am also noting the different types of analysis to be included for scientific precision.
88 participants (Diabetes Care 2026) found no statistically significant association between GLP-1 RAs and NAION (OR 1.50, 95% CI 0.49-4.63). However, the authors noted several limitations: NAION was reported as an adverse event rather than a prespecified outcome, and the total number of events was extremely low, leading to wide confidence intervals. This “neutral” finding highlights the challenges of detecting very rare side effects within the controlled environment of clinical trials, where patients with severe pre-existing ocular disease are often excluded.
Predictors of Risk: Glycemic Trajectory and Demographics
A population-based study using the UK Clinical Practice Research Datalink identified that the risk of NAION was most pronounced within the first 6 months of GLP-1 RA use. Crucially, patients experiencing a hemoglobin A1c reduction of ≥1% were at significantly higher risk. This mirrors the “early worsening” phenomenon seen in diabetic retinopathy, where rapid shifts in osmotic pressure and vascular endothelial growth factor (VEGF) levels due to sudden glycemic control can trigger microvascular complications. Demographic analyses across multiple studies also consistently indicate higher risk in men and patients younger than 50 years.
Comparative Safety: The Role of GIP Agonism
Emerging data suggest that not all incretin-based therapies carry the same ocular risk profile. A retrospective cohort study comparing tirzepatide (a dual GIP/GLP-1 RA) with traditional GLP-1 RAs found that tirzepatide was associated with a 55% lower risk of NAION (HR 0.45, 95% CI 0.27-0.86). This finding suggests that GIP agonism might exert a protective effect or mitigate some of the vascular dysregulation potentially caused by pure GLP-1 agonism, although prospective confirmation is required.
Expert Commentary
Mechanistic Insights
The biological rationale for a semaglutide-NAION link likely involves multi-factorial vascular and metabolic pathways. GLP-1 receptors are expressed in the human retina and optic nerve. Rapid metabolic shifts following semaglutide initiation may alter autoregulation of the microvasculature supplying the optic nerve head. Some researchers hypothesize that GLP-1 RAs may induce transient hypotension or nocturnal dips in blood pressure, which are known triggers for NAION in anatomically susceptible individuals (those with a “disc at risk”). Furthermore, the potential for rapid HbA1c lowering to induce a transient pro-inflammatory or pro-angiogenic state cannot be ignored.
Clinical Applicability and Guidelines
Despite the nearly two-fold increase in relative risk, the absolute risk of NAION remains very low (less than 0.3% in most cohorts). Therefore, semaglutide should not be reflexively withheld from patients who stand to benefit significantly from its cardiometabolic effects. Instead, clinicians should adopt a risk-stratification approach:
- Baseline Assessment: Patients with a history of NAION in one eye, severe diabetic retinopathy, or a known “crowded” optic disc should be counseled extensively.
- Gradual Titration: To avoid the risks associated with rapid glycemic shifts, careful adherence to standard titration schedules is essential.
- Monitoring: Patients should be instructed to report any sudden changes in vision immediately.
Controversy remains regarding the “certainty of evidence.” As noted in the Ophthalmology 2026 systematic review, the current evidence is characterized by high heterogeneity and potential for misclassification in electronic health records. The medical community awaits the results of prospective trials with dedicated ocular endpoints.
Conclusion
The association between semaglutide and NAION represents a critical junction in incretin-based therapy research. While real-world data from large cohorts like the US Veterans suggest a significant increase in relative risk, the absolute incidence remains rare. The discrepancy between observational data and RCT meta-analyses underscores the need for standardized ophthalmic monitoring in future trials. Current evidence points toward a risk profile influenced by formulation dose, the speed of glycemic improvement, and patient-specific anatomical factors. Clinicians must balance these rare ocular risks against the robust and life-saving systemic benefits of GLP-1 RAs, ensuring that patient counseling is both evidence-based and individualized.
References
- Heberer K, et al. New-Onset Nonarteritic Anterior Ischemic Optic Neuropathy and Initiators of Semaglutide in US Veterans With Type 2 Diabetes. JAMA Ophthalmol. 2026;144(3):259-264. PMID: 41678180.
- Diabetes Care. GLP-1 Receptor Agonists and Risk of Optic Nerve or Vision-Threatening Events: A Meta-analysis of Randomized Controlled Trials. 2026;49(3):526-535. PMID: 41587563.
- Diabetes Metab. Glucagon-like peptide-1 receptor agonists and the risk of nonarteritic anterior ischemic optic neuropathy: Evidence from a global real-world cohort. 2026;52(2):101740. PMID: 41687959.
- Ophthalmol Retina. Ocular Outcomes with Tirzepatide versus Glucagon-like Peptide-1 Receptor Agonists in Type 2 Diabetes. 2026. PMID: 41655764.
- Br J Ophthalmol. Ischemic optic neuropathy with semaglutide: global observational analysis of sex- and formulation-specific risk. 2026. PMID: 41807083.
- Ophthalmology. Semaglutide and the Risk of Nonarteritic Ischemic Optic Neuropathy: A Systematic Review and Certainty of Evidence Meta-Analysis. 2026. PMID: 41692115.