PMID-40383360 — Lakhani 180 Countries: GLP-1 RA Ocular Adverse Events

[DRAFT — authored 2026-04-18. Requires Medical Director review.]

Citation

Lakhani M, Kwan ATH, Mihalache A, et al. Association of Glucagon-Like Peptide-1 Receptor Agonists With Optic Nerve and Retinal Adverse Events: A Population-Based Observational Study Across 180 Countries. American Journal of Ophthalmology. 2025 Sep;277:148-168. doi: 10.1016/j.ajo.2025.05.013. PMID: 40383360.

Study Design

Design: Population-based observational pharmacovigilance analysis across 180 countries
Databases: US FDA Adverse Event Reporting System (FAERS, n=12,936,341) + WHO VigiBase (n>35,000,000)
Signal detection methodology: Reporting Odds Ratio (ROR) with 95% CI
Comparator: Metformin (primary); empagliflozin and dulaglutide (sensitivity)
Exposure: Semaglutide, tirzepatide, and other GLP-1 RAs

Population

Reports in FAERS and VigiBase for semaglutide and tirzepatide among patients with T2D and obesity indications. Semaglutide and tirzepatide accounted for 76,444 cases (0.59%) in FAERS and 118,639 cases (0.34%) in VigiBase.

Key Findings

Semaglutide ocular adverse events

Ischemic optic neuropathy (ION): ROR 11.12 (95% CI 8.15–15.16) in FAERS vs. metformin; ROR 68.58 (95% CI 16.75–280.67) in VigiBase
Diabetic retinopathy: ROR 17.28 in FAERS
Retinal/vitreous detachment, hemorrhage, tear: significantly elevated
VigiBase-specific signals: macular edema (ROR 3.87), macular hole (ROR 20.90), papilledema (ROR 6.97) vs. metformin

Tirzepatide ocular adverse events

Diabetic retinopathy: ROR 3.06 (95% CI 2.05–4.58) — significant
NAION / ION: no significant association detected at the same analytical threshold
This represents a class-differentiated signal: tirzepatide's ocular profile is narrower than semaglutide's in this pharmacovigilance dataset

Sensitivity analyses

Using empagliflozin and dulaglutide as comparators (rather than metformin) revealed continued significant associations of semaglutide with ION and DR; vitreous detachment and hemorrhage remained significant when compared to dulaglutide.

Limitations (Author-acknowledged)

Pharmacovigilance databases have inherent reporting biases (sometimes termed "notoriety bias" when a drug receives attention, reporting increases for that drug)
Spontaneous reporting systems cannot establish causation — only signal
Confounders (baseline ocular disease, glycemic state, co-medication) cannot be controlled at the individual level
Tirzepatide has shorter real-world exposure than semaglutide; absence of signal may reflect exposure rather than true absence of association
Report-case-adjudicated vs. administrative data vary by country

Evidence Level

Level III (Oxford CEBM) — large, well-designed pharmacovigilance observational study; strongest signal-detection evidence available short of RCT with rare-event capture.

Linked Peptides

Related Studies

Orchestrator Notes

Definitive evidence for class-differentiated ocular safety profile: semaglutide has broader/larger signals than tirzepatide for ION, NAION, and several retinal events.
EMA PRAC cited similar data in classifying NAION as "very rare" side effect of semaglutide (June 2025).
FDA has not yet updated US labels as of April 2026.
Clinical implication: For patients with ophthalmic risk factors (crowded disc, history of NAION in fellow eye, hypertension), tirzepatide may be preferred over semaglutide where clinical equipoise exists.
This paper was previously misattributed to "Shao et al" in earlier Academy content; the correct lead author is Lakhani M.

PMID-40383360 – Lakhani 180 Countries GLP-1 Ocular Events