More Evidence that GLP-1s May Reduce Risk of Vision-Robbing Eye Diseases

Patients taking GLP-1 receptor agonists (GLP-1RAs) had a lower likelihood of developing age-related macular degeneration (AMD) compared with patients using other glucose-lowering medications or lipid-lowering drugs, a large retrospective cohort study showed.

Propensity-matched comparisons showed that patients on GLP-1 drugs had 21-25% lower 3-year hazards for non-neovascular (nn)AMD as compared with patients taking other glucose-lowering medications. Even larger reductions occurred in hazard ratios for neovascular (n)AMD, on the order of 35-50%.

Compared with users of lipid-lowering medications, patients taking GLP-1 agonists had 2- and 3-year hazard reductions for nnAMD of 16% and 20%, respectively. The impacts on hazards for nAMD were even greater, 30-40% reductions for each year. The rate of conversion from nnAMD to nAMD did not differ significantly between groups.

The findings add to other reports and warrant prospective clinical trials to validate the observations, reported Aleksandra V. Rachitskaya, MD, of the Cleveland Clinic in Ohio, and colleagues in Ophthalmology Retina.

“We recapitulated previous results that GLP-1RA prescription was associated with a reduced risk of developing both nonneovascular and neovascular AMD in older adults with documented ophthalmology follow-up,” the authors stated. “However, here, we show that these medications seem to have no impact on conversion to neovascular AMD in patients with established nonneovascular disease at baseline. Together, these findings suggest that GLP-1RAs … do not promote conversion to neovascular AMD as has been recently reported.”

“Glucagon-like peptide-1 receptor agonists have been shown to exert protective effects across numerous facets of medicine,” they added. “With the exponential increase in patients being prescribed GLP-1RAs, it is critical to understand how this medication class influences ocular disease.”

Still No Direct Evidence

Most of the evidence accumulated to date has come from retrospective studies, and as such cannot be used to support recommendations regarding use of GLP-1 drugs in clinical practice, said Rudrani Banik, MD, of the Icahn School of Medicine at Mount Sinai in New York City.

“I think that the effects we’re seeing in these studies are indirect. And until we have prospective trials, we really can’t say this is a first-line or even second-line treatment strategy,” Banik, who was not involved in the study, told MedPage Today. “It’s promising, for sure, but I think it’s way too early to say ‘This drug is going to help’ reduce risk of AMD.”

Similarly, Sarah DeParis, MD, director of quality and clinical standards for the American Academy of Ophthalmology, said the findings should be interpreted with caution “because this type of study can identify associations but cannot show a cause-and-effect relationship.”

“There can be differences between the groups of patients in the study that aren’t fully accounted for, which may influence the results,” said DeParis. “In addition, electronic health record databases like TriNetX, which was used in this study, have important limitations. They often lack detailed clinical data, such as imaging, visual acuity, and precise disease staging, and they underrepresent care delivered in private ophthalmology practices. That makes it more difficult to fully understand the true ophthalmic outcomes.”

GLP-1 receptor agonists have demonstrated favorable effects in diabetes and obesity, as well as cardiovascular and renal protection, Rachitskaya and colleagues noted. The agents affect insulin signaling and neurohormonal axes and appear to have anti-inflammatory and antioxidant properties, which has spurred interest in a variety of disease states.

In ophthalmology, studies involving preclinical models of diabetic retinopathy and ocular hypertension have shown that GLP-1 agents mitigate damage to retinal ganglion cells, the authors continued. Numerous retrospective studies have suggested a protective role in glaucoma. But, studies of the drugs’ effects on AMD pathogenesis have yielded mixed results, leaving the issue in a state of uncertainty.

Background, Key Results

The authors searched the TriNetX database for 2005-2024 to identify patients older than 60 who had a history of taking GLP-1 drugs, other glucose-lowering medications, or lipid-lowering medications. Eligible patients were followed for 3 years, and investigators performed propensity matching for each year. Additionally, each patient had documentation of an ophthalmology visit. The investigators used diagnostic codes to determine diagnoses of nnAMD and nAMD.

The primary outcomes were the risk of developing nnAMD or nAMD at 1, 2, and 3 years after initial prescription. Conversion from nnAMD to nAMD was a secondary outcome.

In year 1, the propensity-matched groups for GLP-1RAs and other glucose-lowering medications had 43,532 patients each. For the comparisons of GLP-1RAs and lipid-lowering medications, the propensity-matched cohorts had 42,464 patients each.

The cumulative incidence of nnAMD in the GLP-1RA cohort rose from about 0.5% at 1 year to about 1.25% at 3 years but was significantly higher at the end of each year in the patients treated with other glucose-lowering agents, topping out at more than 1.5% after 3 years (P<0.05 to 0.001). For nAMD, the cumulative incidence in the GLP-1RA cohort increased from less than 0.2% after year 1 to about 0.5% at year 3, significantly lower at all three time points versus other glucose-lowering medications, even though the cumulative incidence remained less than 1% at year 3 (P<0.001 for all comparisons).

The nnAMD comparisons with lipid-lowering agents showed no difference between groups after year 1 (about 0.5% in each cohort) but was significantly lower in the GLP-1RA cohort at years 2 (P<0.05) and 3 (P<0.01). The cumulative incidence at year 3 was about 1.5% with lipid-lowering medications and about 1.2% with GLP-1RAs. For nAMD, the cumulative incidence in the lipid-lowering cohort remained under 0.8% at year 3 but was still significantly higher at each time point as compared with the GLP-1RA cohort (maximum of about 0.5%).

The authors acknowledged several limitations of the study, including the inability to perform comparisons for each eye, to match patients reliably by severity or duration of nnAMD at baseline, and to control for drug dosage or adherence.

Please enable JavaScript to view the comments powered by Disqus.