The Eye Risk Is Not Spread Evenly — It Concentrates in People Who Already Have Retinopathy

"GLP-1s can worsen diabetic eye disease" is true, widely repeated, and shaped in a way that helps almost nobody. The labels are more specific, and the specificity is the whole point: the increase is not spread evenly across everyone taking the drug.

The headline number

Ozempic, label revised 10 June 2026:

"5.3 Diabetic Retinopathy Complications In a 2-year trial involving patients with type 2 diabetes and high cardiovascular risk, more events of diabetic retinopathy complications occurred in patients treated with OZEMPIC (3.0%) compared to pl"

Three percent against placebo, over two years, in people already at high cardiovascular risk. That is the figure that travels.

The sentence that follows it, which does not travel

"The absolute risk increase for diabetic retinopathy complications was larger among patients with a history of diabetic retinopathy at baseline (OZEMPIC 8.2%, placebo 5.2%) than among patients without a known history of diabetic retinopathy"

Read those two numbers next to the headline. Among people who already had retinopathy, the rate was 8.2% on the drug against 5.2% on placebo. Among people who did not, the label says the increase was smaller.

So the 3.0% is an average across two quite different groups, and which group you are in changes what the number means for you.

The same shape appears in a different molecule

Trulicity, revised 10 August 2026 — dulaglutide, not semaglutide, and a different manufacturer:

"The proportion of patients with diabetic retinopathy complications was larger among patients with a history of diabetic retinopathy at baseline (TRULICITY 8.5%, placebo 6.2%) than among patients without a known history of diabetic retinopat"

Eight and a half against six point two. Almost the same split as Ozempic's, arrived at independently.

Two labels, two molecules, two manufacturers, the same conditional. That is a stronger signal than either number alone, and it is the part that a summary reading "GLP-1s may worsen retinopathy" throws away.

Wegovy reports it differently again

Wegovy, revised 30 June 2026, in a weight-reduction trial rather than a glycemic one:

"In a trial of adult patients with type 2 diabetes and BMI greater than or equal to 27 kg/m 2 for weight reduction (Study 3), diabetic retinopathy was reported by 4% of WEGOVY injection-treated patients and 2.7% placebo-treated patients [see"

Four percent against 2.7%. Different trial design, different population, same direction. The Wegovy label does not break this one out by baseline history the way the other two do.

Why the mechanism matters here

The leading explanation is not that these drugs damage the retina. It is that rapid improvement in blood glucose can transiently worsen existing retinopathy — an effect documented long before this drug class, in intensive insulin therapy. That is consistent with what the labels show: the risk concentrating in eyes that already have disease, and appearing during the period when glucose is falling fastest.

It is also why the answer is not "avoid the drug". It is "look before you start, and look again while it is working".

What to do with this

Find out whether you have retinopathy before starting, not after. The labels split their numbers on exactly that question, and most people do not know the answer without a dilated exam.

If you have it, this is a scheduling conversation. Not whether to take the drug, but how closely the eyes get watched while glucose moves.

Do not read the headline percentage as your risk. 3.0% is an average over a group that includes people at 8.2% and people well below it.

Check the revision date. Ozempic's label is June 2026, Trulicity's August 2026, Wegovy's June 2026. Retinopathy language in this class has been revised more than once as trial data accumulated.

Where this comes from

Every quotation is from the manufacturer's own Structured Product Label as published on DailyMed, the National Library of Medicine's repository. We use the manufacturer's label rather than a repackager's copy, which can lag by several revisions.

The mechanism paragraph is the standard explanation and is not quoted from a label — it is flagged as such rather than dressed up as a citation.

This is information, not medical advice. Whether and how to monitor your eyes on a GLP-1 is a decision for you and your prescriber; the labels give you the number that applies to your situation rather than the average.