Pharmacovigilance data show semaglutide with a reporting odds ratio of 120.8 (95% CI 110.6–132.0) for non-arteritic anterior ischaemic optic neuropathy across 879 FAERS reports, roughly sixty times the relative risk the European regulator accepted. The odds that a report named NAION rose 48.6-fold after the 2024 cohort publication. Mendelian randomisation using six GLP1R expression variants found no association with anterior ischaemic optic neuropathy (0.56, 0.21–1.49), but the test could only detect effects above fourfold.
Spontaneous-report databases cannot estimate risk, because publicity changes what gets reported; this analysis puts a number on that distortion. The 1-in-10,000 person-year figure from the EMA remains the more credible anchor, and it describes a very rare event. The genetic null is weak evidence: variants shifting GLP1R expression don't mimic a drug that pharmacologically activates the receptor, and the study had only 244 cases. The authors themselves conclude the causal question stays open in both directions. For adults on semaglutide for diabetes or obesity, the practical message is unchanged: sudden vision loss warrants urgent assessment, and people with prior NAION or crowded optic discs should discuss risk with their doctor, while the large cardiometabolic benefits still dominate for most users. This is a preprint, not yet peer reviewed, and its conclusions may change. It is a useful caution on interpreting safety signals rather than a finding that overturns anything.