Glucagon-like Peptide-1 Receptor Agonists and Risk of Nonarteritic Anterior Ischemic Optic Neuropathy: Systematic Review and Meta-Analysis
Dhivagaran T, Butt F, Arunasalam L, Bae I, Mikhail D, Tao BK
Neurology (2026) · n=1,593,554
Across 1,593,554 patients in a Neurology-published meta-analysis (682,456 semaglutide users vs 911,098 non-GLP-1RA users), semaglutide use was associated with a 2.5-fold increased risk of nonarteritic anterior ischemic optic neuropathy (RR 2.52; 95% CrI 1.56–4.72), with the strongest signal in patients with diabetes (RR 2.41; 95% CrI 1.57–4.10) — the strongest meta-analytic confirmation to date that the NAION safety signal is real.
This is the largest systematic review and meta-analysis to date addressing the question of whether glucagon-like peptide-1 receptor agonists — particularly semaglutide — are associated with nonarteritic anterior ischemic optic neuropathy (NAION), an acute optic-nerve infarction event that can cause permanent visual field loss. The work is published in Neurology by a University of Toronto-led ophthalmology and neurology team, and represents the independent academic synthesis of the conflicting observational literature that has accumulated since the Hathaway 2024 Mass Eye and Ear retrospective cohort study.
Embase, Medline, and Cochrane CENTRAL were searched from inception to April 5, 2025. Observational studies and randomized controlled trials comparing NAION risk in GLP-1RA users versus non-GLP-1RA users were included. Screening was conducted in duplicate by two independent reviewers with adjudication by a third; risk of bias was assessed with ROBINS-I, and evidence certainty was evaluated with GRADE. Pooled relative risks were estimated using Bayesian random-effects models with prespecified priors.
Five studies were included in the primary analysis, encompassing 1,593,554 patients (682,456 semaglutide users and 911,098 non-GLP-1RA users). Semaglutide use was associated with a 2.52-fold increased risk of NAION (RR 2.52; 95% credible interval 1.56–4.72; I² 57.8%; posterior probability of RR > 1: 99.9%). The signal was particularly pronounced among patients with diabetes (RR 2.41; 95% CrI 1.57–4.10; I² 47.2%; posterior probability of RR > 1: 99.9%).
The overall NAION incidence across 5 studies (1,460,760 patients, 272 NAION events, of which semaglutide n=198, dulaglutide n=54, exenatide n=20) was 85 cases per 100,000 GLP-1RA users (95% CrI 29–263; I² 98.8%).
The meta-analysis is composed of five observational studies — the conclusion is the synthesis of observational evidence, not RCT evidence. Confounding by indication, healthcare-utilisation bias, and ascertainment bias remain real concerns at the level of the individual underlying studies, and meta-analytic pooling does not remove these biases — it averages over them. The I² heterogeneity is moderate (57.8%) for the primary semaglutide pooled RR, indicating the individual studies are not unanimous in either magnitude or direction.
The 85-per-100,000 incidence estimate has very wide credible intervals (29–263) and I² of 98.8%, reflecting deep between-study heterogeneity in both the underlying NAION-detection methodology and the comparator groups. This is consistent with a real but small absolute risk that is hard to estimate precisely.
The pooled RCT data — published in Vilsbøll 2026 in the British Journal of Ophthalmology, a pooled analysis of Novo Nordisk-sponsored semaglutide and liraglutide trials covering 96,829 participants — found no increased NAION incidence, with point estimates lower in the GLP-1RA arm than placebo. The Vilsbøll analysis has sponsor-affiliation conflict but is RCT-derived; the Dhivagaran analysis is independent academic but observational. The honest synthesis is that the question is not yet settled and both signals should be reported with their respective methodological frames.
NAION is biologically rare; the absolute increase implied by a 2.52-fold relative risk on a low baseline rate is still small in absolute terms. The clinical decision is not "avoid semaglutide" but "informed consent and ophthalmologic vigilance in patients with NAION risk factors" (age > 50, disc-at-risk anatomy, hypertension, diabetes, sleep apnea, hyperlipidaemia).
07·Member discussion
No member discussion yet.
Member-only conversation lives here — cycle notes, practitioner commentary, pattern-matching. Be the first paying member to start the thread.