Evidence for Vision Loss with Semaglutide
Semaglutide carries an FDA-labeled warning for diabetic retinopathy complications, with increased risk particularly in patients with pre-existing retinopathy, but the evidence for nonarteritic anterior ischemic optic neuropathy (NAION) remains inconclusive and likely reflects confounding by underlying disease severity rather than direct drug causation. 1
FDA-Labeled Diabetic Retinopathy Risk
The FDA label explicitly warns that in a 2-year cardiovascular outcomes trial, diabetic retinopathy complications occurred in 3.0% of semaglutide-treated patients versus 1.8% on placebo. The risk was substantially higher in patients with baseline diabetic retinopathy (8.2% vs 5.2%) compared to those without known retinopathy (0.7% vs 0.4%). This worsening is attributed to rapid glucose reduction rather than direct drug toxicity. 1
Key clinical action: Assess retinopathy status before initiating semaglutide, particularly when intensifying glucose-lowering therapy, as rapid A1C reductions can temporarily worsen diabetic retinopathy. 2
NAION: Conflicting Evidence Without Causal Proof
The NAION controversy centers on observational studies showing divergent results based on comparator selection:
Studies Suggesting Increased Risk
- Observational data in diabetes populations show hazard ratios of 1.76-1.85 when comparing semaglutide to non-GLP-1 RA therapies 3, 4
- One meta-analysis found pooled OR of 3.92 for NAION, though with wide confidence intervals (1.02-15.02) 5
- Absolute incidence remains extremely low: 26.7 per 100,000 person-years on semaglutide versus 18.9 per 100,000 person-years off drug 3
Studies Finding No Association
- Randomized controlled trials show no significant NAION risk (RR 1.76,95% CI 0.43-7.25 for diabetes; RR 2.18,95% CI 0.33-14.34 for weight loss) 3
- Studies using active comparators with similar metabolic profiles (e.g., SGLT2 inhibitors) found no association (OR 0.72) or even protective effects 6, 7
- When semaglutide was compared to other diabetes medications in matched cohorts, the signal disappeared or reversed 6
Critical Interpretation: Confounding by Indication
The most plausible explanation for positive observational signals is confounding by indication: patients prescribed semaglutide typically have more severe diabetes, obesity, hypertension, and cardiovascular disease—all independent NAION risk factors. Studies using untreated or healthier comparators consistently show elevated risk, while those using metabolically similar active comparators do not. 7
The 2025 ADA guidelines note that meta-analyses of cardiovascular outcomes trials show no association between GLP-1 RAs and retinopathy except through A1C reduction effects, and that rapid glucose improvement can temporarily worsen retinopathy regardless of agent used. 2
Clinical Recommendations
For patients with pre-existing diabetic retinopathy:
- Perform dilated fundoscopic examination before initiating semaglutide 2
- Monitor for retinopathy progression with annual ophthalmology follow-up 1
- Titrate semaglutide gradually to avoid precipitous A1C drops that can worsen retinopathy 2
For all patients:
- The absolute NAION risk remains extremely low (0.02-0.04% over 2 years) even in positive studies 4
- Do not withhold semaglutide based on NAION concerns alone, as cardiovascular and renal benefits substantially outweigh the uncertain and minimal absolute risk 7
- Counsel patients about sudden vision loss symptoms but emphasize the lack of proven causation
- Document baseline optic disc anatomy if concerned (crowded discs increase baseline NAION risk independent of medication) 8
Common pitfall: Overinterpreting observational NAION signals without recognizing that diabetes severity itself—not the medication—likely drives the association. The FDA label does not include NAION as a warning, only diabetic retinopathy complications related to rapid glucose improvement. 1