Does Semaglutide Cause Diabetic Retinopathy? Evidence from SUSTAIN-6, FOCUS, and the FDA Labels

Every semaglutide FDA label — Wegovy, Ozempic, and Rybelsus — carries a Section 5 warning about diabetic retinopathy complications. The signal originates from a single trial, SUSTAIN-6 (Marso et al., NEJM 2016)^[1], where retinopathy complications in 3,297 type 2 diabetics with established cardiovascular disease occurred at hazard ratio 1.76 (95% CI 1.11 to 2.78, P=0.02) versus placebo. The Vilsbøll 2018 post-hoc analysis^[2] attributed the majority of that effect to the magnitude and rapidity of HbA1c reduction during the first 16 weeks of treatment in patients with pre-existing retinopathy and insulin use — the same “early worsening” phenomenon documented in the DCCT (NEJM 1993)^[3] with intensive insulin therapy. Read narrowly, that means the signal is real and load-bearing for one specific subgroup of type 2 diabetics with pre-existing retinopathy. Read broadly, it has been misinterpreted into a generalized “Wegovy hurts your eyes” concern that has no support in the non-diabetic obesity trials. This article walks through the verified trial data, the FDA label text, the ADA 2026 and AAO 2024 society guidance, and a tier-stratified guide for what the evidence means for three different patient profiles.

The SUSTAIN-6 result, in plain language

SUSTAIN-6 randomized 3,297 adults with type 2 diabetes and high cardiovascular risk to once-weekly semaglutide (0.5 mg or 1.0 mg) or matching placebo, and followed them for 104 weeks^[1]. The primary endpoint was the composite of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke (MACE-3). It was met for non-inferiority and trended toward superiority — primary outcome 6.6% on semaglutide vs 8.9% on placebo, hazard ratio 0.74 (95% CI 0.58 to 0.95, P<0.001 for noninferiority). That alone was a major positive result for the drug.

Buried inside the safety section was the finding that put a warning on every semaglutide label since: rates of retinopathy complications — defined verbatim in the paper as “vitreous hemorrhage, blindness, or conditions requiring treatment with an intravitreal agent or photocoagulation” — were significantly higher in the semaglutide group. Hazard ratio 1.76, 95% confidence interval 1.11 to 2.78, P=0.02. In absolute terms, 50 of 1,648 semaglutide patients (3%) had a retinopathy complication versus 29 of 1,649 placebo patients (1.8%) over two years.

Why the signal is largely a glycemic-velocity effect

The mechanism is not unique to semaglutide and was not discovered with semaglutide. The Diabetes Control and Complications Trial^[3] showed in 1993 that intensive insulin therapy in type 1 diabetics produced an early, transient worsening of retinopathy that resolved into long-term net benefit. The accepted explanation is that rapid normalization of glucose acutely worsens retinal microvascular pathology in eyes with pre-existing retinopathy — through retinal hemodynamics, growth-factor shifts, and vasoregulatory changes — before the long-term protective effect of better glycemic control accrues.

Vilsbøll et al. (2018, Diabetes Obes Metab)^[2]published a post-hoc analysis of the SUSTAIN-6 retinopathy cohort that asked the “is this the early-worsening phenomenon, applied to semaglutide?” question directly. Their conclusion, in their own words: “The majority of the effect with semaglutide vs placebo in SUSTAIN 6 may be attributed to the magnitude and rapidity of HbA1c reduction during the first 16 weeks of treatment in patients who had pre-existing DR and poor glycaemic control at baseline, and who were treated with insulin.” And: “Early worsening of DR is a known phenomenon associated with the rapidity and magnitude of improvement in glycaemic control with insulin; the DRC findings in SUSTAIN 6 are consistent with this.”

That reframing is the most important interpretive fact in this article. SUSTAIN-6 enrolled a population dramatically enriched for the early-worsening risk profile: long-duration diabetes (mean 13.9 years), high baseline HbA1c, established cardiovascular disease, and high insulin use. Patients who already had pre-existing diabetic retinopathy at baseline were the dominant contributors to the absolute risk increase — and they happened to also be the patients in whom HbA1c fell the fastest on semaglutide.

The FDA label, verbatim

Section 5.8 of the current Wegovy prescribing information^[4] reads:

“In a 2-year trial with semaglutide 0.5 mg and 1 mg in patients with type 2 diabetes and high cardiovascular risk, diabetic retinopathy complications occurred in 3% of semaglutide-treated patients compared to 1.8% with placebo. The absolute risk increase was larger among patients with a history of diabetic retinopathy at baseline (semaglutide 8.2%, placebo 5.2%) than among patients without a known history (semaglutide 0.7%, placebo 0.4%). Rapid improvement in glucose control has been associated with a temporary worsening of diabetic retinopathy. Patients with a history of diabetic retinopathy should be monitored for progression of diabetic retinopathy.”

Two things to notice. First, the label itself sub-stratifies the risk: in patients without pre-existing retinopathy, the absolute event rate on semaglutide was 0.7% over two years, versus 0.4% on placebo. That is the rate that applies to the typical type 2 diabetic without a prior retinopathy diagnosis. Second, the label explicitly attributes the effect to “rapid improvement in glucose control” — not to a drug-specific retinopathic mechanism. The Ozempic and Rybelsus labels carry analogous Section 5 language. The Mounjaro, Zepbound, and Foundayo labels carry parallel class-level language for tirzepatide and orforglipron.

What about non-diabetic patients on Wegovy?

The SELECT cardiovascular outcomes trial (Lincoff et al., NEJM 2023) randomized 17,604 adults with overweight or obesity and pre-existing cardiovascular disease — but without diabetes — to semaglutide 2.4 mg or placebo for a median follow-up of about 40 months. SELECT did not report a retinopathy signal as an adverse event of interest. Pooled meta-analytic work in non-diabetic semaglutide cohorts likewise has not found a diabetic retinopathy signal.

That makes mechanistic sense: a non-diabetic patient does not have diabetic retinopathy to acutely worsen, and the early-worsening phenomenon is a glycemic-velocity effect on pre-existing diabetic microvascular pathology. The SUSTAIN-6 signal does not transport to the non-diabetic Wegovy population, and the Wegovy label Section 5.8 itself describes the signal as occurring “in patients with type 2 diabetes.”

Tirzepatide and orforglipron — the class effect

SURPASS-2^[5], the head-to-head tirzepatide vs semaglutide trial in type 2 diabetics, excluded patients with proliferative diabetic retinopathy or diabetic retinopathy requiring acute therapy from enrollment. So did SURPASS-4 and the broader tirzepatide pivotal program. Because of those exclusions, tirzepatide trials do not have a direct retinopathy-progression endpoint analogous to SUSTAIN-6, and comparisons of tirzepatide vs semaglutide on this specific outcome should be made with that asymmetry explicit.

The Mounjaro and Zepbound FDA labels nevertheless carry the same class language as the semaglutide labels: “Rapid improvement in glucose control has been associated with a temporary worsening of diabetic retinopathy.” The Foundayo (orforglipron) label, approved April 1, 2026, carries parallel class-level language. The class-level FDA framing is consistent with the early-worsening interpretation rather than a semaglutide-specific intrinsic effect.

What ADA and AAO actually recommend

The American Diabetes Association Standards of Care 2026, Section 12^[6], recommends an initial dilated and comprehensive eye examination by an ophthalmologist or optometrist at the time of type 2 diabetes diagnosis, with annual screening thereafter (or every 1-2 years if no retinopathy is present and glycemic targets are met). On GLP-1 RAs specifically the 2026 guidance states: “Retinopathy status should be assessed when glucose-lowering therapies are intensified, such as those using GLP-1 RAs, since rapid reductions in A1C have been shown to be associated with a risk of initial worsening of retinopathy.”

The American Academy of Ophthalmology Diabetic Retinopathy Preferred Practice Pattern, published in Ophthalmology February 2025^[7], confirms the same baseline-screening rhythm and explicitly notes that rapid tight glycemic control can accelerate retinopathy onset, naming the semaglutides among newer agents. Neither the ADA nor the AAO recommends withholding GLP-1 therapy from a patient with pre-existing retinopathy. Both recommend close coordination with an ophthalmologist when intensifying therapy in patients with significant baseline retinopathy.

FOCUS — the trial we are still waiting on

FOCUS^[8] (NCT03811561, “Long-term Effects of Semaglutide on Diabetic Retinopathy in Subjects With Type 2 Diabetes”) is the Phase 3 randomized, quadruple-masked trial Novo Nordisk committed to as a regulatory follow-up to the SUSTAIN-6 signal. It enrolls approximately 1,500 type 2 diabetic adults with HbA1c 7.0-10.0% and ETDRS retinopathy level 10-75 (i.e., must have at least mild non-proliferative retinopathy at baseline), and randomizes them to once-weekly semaglutide 1.0 mg or matching placebo. The primary outcome is the proportion of subjects with at least 3-step progression on the ETDRS Diabetic Retinopathy Severity Scale at five years. The trial started May 8, 2019.

FOCUS estimated primary completion is November 7, 2027 and no peer-reviewed primary outcome publication exists as of April 2026. Until FOCUS reads out, the SUSTAIN-6 two-year signal remains the load-bearing trial-level evidence and Section 5.8 of the Wegovy label remains unchanged. Press releases and conference abstracts that pre-announce FOCUS results should be cited cautiously and do not change the regulatory record.

Tier-stratified guide — what to actually do

The evidence does not support a one-size-fits-all rule. The right action depends on the patient profile.

• Tier 1 — Non-diabetic patient on Wegovy or Zepbound for chronic weight management. No meaningful retinopathy concern from SUSTAIN-6. The signal does not transport to the non-diabetic population, and SELECT did not report a retinopathy signal. No special eye-exam protocol is mandated by ADA or AAO for this group. Standard age-appropriate eye care continues.
• Tier 2 — Type 2 diabetic patient on Ozempic, Mounjaro, or Foundayo, no pre-existing retinopathy or mild non-proliferative disease. Continue the standard ADA dilated eye exam at diagnosis with annual (or 1-2 year) screening. The absolute event rate in this group from the Wegovy label was 0.7% on semaglutide vs 0.4% on placebo over two years — a real but small increment that does not meaningfully change the benefit-risk balance for a patient who needs glycemic improvement and weight management.
• Tier 3 — Type 2 diabetic patient with pre-existing significant retinopathy (especially proliferative disease), HbA1c above 9%, and current insulin treatment. This is the SUSTAIN-6 high-risk subgroup. Coordinate the GLP-1 initiation with an ophthalmologist and consider stabilizing the eye (anti-VEGF or photocoagulation) before or alongside intensification. A slower titration may be reasonable. Do not withhold GLP-1 therapy on retinopathy grounds alone — no specialty body recommends that — but do not initiate without ophthalmology input.

The defensible YMYL framing for this article: the SUSTAIN-6 signal is real and FDA-labeled, the mechanism is best explained as rapid-glycemic-improvement early worsening rather than an intrinsic retinopathic effect of semaglutide, and the practical implications stratify cleanly by patient profile rather than collapsing into a single warning that misleads either the non-diabetic Wegovy user (who is overcautioned) or the type 2 diabetic with proliferative retinopathy (who is undercounseled).

Related research

• GLP-1 side effects — what STEP-1 and SURMOUNT-1 actually showed — the broader trial-level safety picture this retinopathy section sits inside.
• SELECT trial — semaglutide cardiovascular benefits in non-diabetics — the trial that establishes Wegovy's safety profile in the non-diabetic obesity population.
• FLOW — how semaglutide became the first GLP-1 approved for diabetic kidney disease — adjacent SUSTAIN-6-era diabetes trial program.
• Tirzepatide vs semaglutide — SURPASS and STEP head-to-head — the trials that excluded proliferative DR and what that asymmetry means.
• Does GLP-1 cause cancer? MTC and pancreatic evidence — companion piece on the other big FDA-label safety question.
• GLP-1s and blood pressure — adjacent vascular-effect article using the same SUSTAIN-6 / SELECT evidence base.