The eye-risk signal is not settled
What the study found
Published in Frontiers in Pharmacology in 2026, the review by Anatriello and colleagues combined 28 observational studies. Six examined semaglutide and 22 examined GLP-1 receptor agonists as a class, with the included evidence drawn from people with type 2 diabetes.
Compared with other antidiabetic treatments, the pooled estimates showed no statistically significant increase in the outcomes examined:
- NAION: Risk ratio 1.01 (95% CI 0.62–1.64) — no statistically significant increase
- Glaucoma: Hazard ratio 0.84 (95% CI 0.71–1.00) — no statistically significant increase
- Diabetic retinopathy, new onset: Hazard ratio 0.96 (95% CI 0.85–1.08) — no statistically significant increase
- Diabetic retinopathy, progression: Hazard ratio 0.97 (95% CI 0.83–1.14) — no statistically significant increase
That is reassuring evidence. It is not proof that risk is absent.
Where the uncertainty sits
The NAION estimate is the clearest reason for caution. Its confidence interval runs from 0.62 to 1.64, so the pooled evidence does not exclude a clinically meaningful increase or decrease. Heterogeneity was also high: I² was 89% for NAION and 91% for glaucoma and new-onset retinopathy.
The evidence base was observational rather than randomised, and it centred on type 2 diabetes. That limits how confidently the result can be carried across to obesity-only patients using GLP-1 medicines through private weight-management services.
The authors' own conclusion keeps monitoring on the table. The defensible reading is “no statistically significant class-wide increase was detected in these studies,” not “GLP-1 medicines cannot affect eye health.”
The MHRA position still governs
In February 2026, the MHRA advised that semaglutide may be very rarely associated with NAION. Its safety update tells healthcare professionals to advise patients about sudden vision loss and to arrange urgent ophthalmology assessment when NAION is suspected.
That position is not automatically cancelled by the meta-analysis. The review pools several medicines and observational designs; the MHRA warning is semaglutide-specific and reflects a regulatory assessment of the wider safety evidence.
For UK clinics, current regulator guidance and the authorised medicine information remain the operational baseline. The new paper is useful context for risk communication, but it should not be presented to patients as evidence that the warning no longer matters.
What good governance looks like
- Use precise language: say the review detected no statistically significant overall increase; do not say it proved there is no risk.
- Keep safety material current: align patient information and staff escalation pathways with the latest MHRA advice.
- Separate class evidence from medicine-specific warnings: a pooled GLP-1 estimate and a semaglutide regulatory warning answer different questions.
- Record uncertainty: high heterogeneity and the NAION confidence interval belong in any serious internal evidence review.
Key takeaway
The review is reassuring, but it does not close the GLP-1 eye-safety question. Clinics should hold both facts at once: no statistically significant class-wide increase was detected in the pooled observational evidence, and the MHRA still carries a very-rare NAION warning for semaglutide.