Review suggests semaglutide eye risk is not shared by all GLP-1 drugs

A review accepted September 10, 2026 in the journal Frontiers in Medicine argues that the eye-related risks reported with some GLP-1 medications are not a shared effect of the entire drug class. The authors say longer-acting semaglutide — sold as Ozempic, Wegovy and Rybelsus — has been repeatedly linked to non-arteritic anterior ischemic optic neuropathy (NAION) and, when blood sugar drops quickly, to early worsening of diabetic retinopathy, while shorter-acting drugs such as liraglutide and exenatide show much weaker signals ^[1].

The paper is a narrative review, not new primary data from patients, and it was received August 3, 2026 and accepted September 10, 2026. It does not report new incidence rates, case counts or clinical outcomes; it pulls together existing evidence and proposes an explanation for why the signals appear to differ between drugs ^[1].

According to the review, the split may come down to how long each drug keeps its receptor engaged and how fast it corrects blood sugar, rather than a single mechanism that all GLP-1 receptor agonists share ^[1]. The authors build their case from three lines of evidence: biased agonism and receptor trafficking at the cellular level, signaling differences between specific cell types in the eye, and a recently proposed hemodynamic mechanism involving GLP-1 receptor agonist-related changes in intracranial pressure ^[1].

In that model, the eye liabilities of these drugs may reflect a mix of factors: the receptor occupancy profile dictated by each individual drug molecule, the signaling environment of particular cell types, and a person's own anatomical vulnerabilities ^[1]. The review states that this model makes testable predictions and has implications for drug choice, risk-stratified monitoring, and development of future agonists with better eye safety ^[1]. It does not claim to prove that semaglutide causes these eye problems, and it does not quantify how much risk any drug carries.

The review lists two priority areas for future research: head-to-head comparisons of how different GLP-1 receptor agonists signal in specific eye cell types, and the risk-benefit balance in people who do not have diabetes and are using these drugs for weight loss ^[1]. The authors are affiliated with The First Affiliated Hospital of Chongqing University of Traditional Chinese Medicine in Chongqing, China ^[1].

Why it matters for patients

Much of the public discussion about GLP-1 drugs and vision has treated eye risk as a property of the whole class. This review argues the opposite: that the evidence points to differences between molecules, with the strongest signals attached to longer-acting semaglutide and much weaker ones attached to liraglutide and exenatide ^[1]. If that framework holds up, it would matter for how doctors weigh one drug against another, and for how closely people on different GLP-1 drugs are monitored for eye problems ^[1].

It is worth being clear about what this paper is not. It is a review of existing evidence, not a trial or a safety study, so it does not change any prescribing label, and no regulator action is reported in it ^[1]. It also does not give advice on who should stop, switch or avoid a medication, and it does not address the risk-benefit question for people using these drugs without diabetes — the authors name that as an open question ^[1]. Anyone with vision changes while taking a GLP-1 medication has a reason to contact a clinician, and that clinical conversation is where individual risk gets sorted out.

The review also leaves an important gap: it does not establish how often these eye events occur, which patients are most vulnerable, or whether the proposed mechanisms are correct. Those answers are not yet known from the evidence it reviews ^[1].

What happens next

The journal notes that the final, formatted version of the article will be published soon, after the accepted manuscript posted September 10, 2026 ^[1]. The authors describe their framework as generating testable predictions, which means the next step would be studies designed to confirm or refute it, including the head-to-head comparisons of drug signaling in eye cells that the review calls for ^[1]. No specific trials, timelines or regulatory reviews tied to this paper have been announced ^[1].