Semaglutide and Diabetic Retinopathy

Semaglutide was associated with more diabetic-retinopathy complications than placebo in SUSTAIN-6, a trial in adults with type 2 diabetes at high cardiovascular risk. The Ozempic label reports 3.0% versus 1.8% over two years. That is a reason to plan appropriate eye monitoring, especially with existing retinopathy, rather than assume every visual symptom has the same cause. ^{1 2}

What diabetic retinopathy means

Diabetic retinopathy affects blood vessels in the retina, the light-sensitive tissue at the back of the eye. It can initially have no symptoms. The National Eye Institute recommends comprehensive dilated eye examinations for people with diabetes and explains that early detection can help protect vision. ³

This is distinct from NAION, an optic-nerve condition discussed in a separate semaglutide and NAION evidence review. Combining both into a single “blindness risk” number changes the question. The conditions, mechanisms and supporting studies are different, even though either can prompt concern about eyesight.

A symptom cannot identify the diagnosis by itself. New floaters, worsening vision or sudden visual loss require clinical assessment; they should not be dismissed as an expected adjustment to treatment. NEI describes floating spots and bleeding-related symptoms as reasons to obtain treatment promptly. ³

What SUSTAIN-6 actually measured

SUSTAIN-6 randomized 3,297 adults with type 2 diabetes to weekly semaglutide or placebo for 104 weeks. Most had established cardiovascular disease, chronic kidney disease or both. The primary question concerned cardiovascular safety, rather than a trial designed specifically to resolve all questions about diabetic eye disease. ¹

The retinopathy-complication outcome included vitreous hemorrhage, blindness, and conditions requiring treatment with an intravitreal agent or photocoagulation. The hazard ratio was 1.76, with a 95% confidence interval of 1.11–2.78. That signals a higher hazard in the semaglutide group during the trial; it does not mean 76% of treated participants developed an eye complication. ¹

The label's absolute figures make the scale clearer: 3.0% with semaglutide and 1.8% with placebo. Their difference is 1.2 percentage points. This was a particular high-risk diabetes population over a particular follow-up period, so it should not be presented as an exact prediction for every person using semaglutide. ²

Why rapid glucose improvement is part of the explanation

A later analysis examined the SUSTAIN program and the SUSTAIN-6 signal. It found no comparable imbalance in the other included SUSTAIN trials and suggested that much of the SUSTAIN-6 difference could be explained by the magnitude and speed of HbA1c reduction during the first 16 weeks in people with existing retinopathy, poor baseline glucose control and insulin treatment. ⁴

This interpretation is consistent with the recognized phenomenon of early worsening of retinopathy when glucose control improves rapidly. However, the analysis was post hoc. It was not a new trial randomizing people to different speeds of glucose reduction in order to isolate the mechanism. “May explain much of the signal” is more defensible than “proves semaglutide cannot affect the eye directly.” ⁴

The distinction has practical value. If risk depends partly on existing disease and changes in glucose control, the useful response is to identify that context and coordinate care. It is not to assume that a general warning applies identically to someone without diabetes and someone with advanced diabetic eye disease.

It also does not justify changing insulin or semaglutide on your own. The balance among glucose control, cardiovascular benefit, eye status and other treatments is individual. A prescriber and eye-care clinician can interpret the relevant history together.

What to clarify before starting or changing treatment

Tell the prescriber about diagnosed retinopathy, previous eye injections or laser treatment, and recent changes in vision. Ask whether the eye-care clinician needs to know about the planned glucose-lowering treatment and how follow-up will be coordinated. These details are more actionable than asking whether semaglutide universally “causes blindness.”

The Ozempic label specifically calls for monitoring patients with a history of diabetic retinopathy. NEI also emphasizes that early retinopathy may be symptom-free. Feeling that your vision is normal is therefore not a substitute for the examinations appropriate to your diabetes care. ^{2 3}

A remote weight-management program should be able to explain who handles this coordination. A video visit does not itself provide a retinal examination, and an app's symptom questionnaire cannot establish that retinopathy is absent. If a service is prescribing for someone with diabetes, ask how it connects with existing diabetes and eye care rather than treating weight loss as the only outcome.

Benefit and risk belong in the same discussion

SUSTAIN-6 also reported fewer primary cardiovascular events with semaglutide, although its design and statistical questions need to be read on their own terms. An eye safety signal does not erase the rest of a medicine's evidence; equally, cardiovascular evidence does not make an eye warning disappear. ¹

A balanced decision asks what the medicine is intended to do for this patient, what alternatives exist, and which risks require monitoring. Someone with both substantial cardiovascular risk and retinopathy may have a more complex decision than a headline suggests. This article cannot make that decision from a diagnosis alone.

The evidence also should not be used to reassure beyond what was measured. The post hoc explanation is informative, but longer-term eye outcomes and individual responses remain clinical questions. Conversely, isolated reports of visual symptoms should not be merged with the randomized retinopathy endpoint as if they measured the same condition.

A clearer way to read the warning

Keep four details together: the studied diabetes population, the defined retinopathy-complication outcome, the absolute event proportions and the importance of existing eye disease. Then add the uncertainty about mechanism without pretending it has been settled.

That approach supports timely examination and coordinated prescribing. It avoids both an alarmist claim that most users will lose vision and an unsupported reassurance that every eye symptom is harmless. If vision changes suddenly, seek prompt medical assessment rather than waiting for a routine weight-management follow-up. ³