This IRIS Registry cohort study assessed whether initiation of GLP-1 receptor agonists (GLP-1RAs) in patients with diabetic macular edema (DME) is associated with a change in the burden of intravitreal anti-VEGF treatments. The study objective was to evaluate treatment-frequency trajectories after starting GLP-1RA therapy and to examine secondary outcomes of visual acuity (VA) and central subfield thickness (CST).
Researchers used the American Academy of Ophthalmology IRIS® Registry, a de-identified, HIPAA-compliant clinical database, to identify patients with DME. Incident GLP-1RA initiators were matched 1:1 to controls using Mahalanobis distance matching. Matching variables included sociodemographic characteristics, DME risk factors, and comorbid conditions that commonly influence GLP-1RA prescription decisions (for example, hypertension, obesity, and chronic kidney disease). The matched sample comprised 9,896 GLP-1RA initiators and 9,896 matched controls, for a total N = 19,792.
The primary outcome was monthly anti-VEGF injection frequency. Secondary outcomes were visual acuity (VA) and central subfield thickness (CST). Outcomes were evaluated across a 36-month observation window: 12 months before initiation through 24 months after GLP-1RA initiation. The study adjusted analyses for DME duration.
A longitudinal mixed-effects event-study model was used to estimate monthly anti-VEGF injection trajectories and to test for differences between the GLP-1RA cohort and matched controls over time. The model included interaction terms to evaluate whether trajectories diverged after GLP-1RA initiation. Results were reported with interaction coefficients and associated P values. Secondary outcomes (VA and CST) were analyzed within the same matched framework.
Across the matched cohorts, there was no statistically significant change in anti-VEGF injection frequency following GLP-1RA initiation when compared with matched controls. Interaction coefficients reported in the abstract ranged from -0.18 to 1.59 and were not statistically significant (P > 0.05). In other words, the trajectory of anti-VEGF use over the 24 months after initiation did not differ meaningfully between those who started GLP-1RA therapy and those who did not.
No significant differences were observed between cohorts for the secondary structural and functional outcomes. Reported differences in VA were in the range of -0.05 to 0.04 logMAR (P > 0.05). Changes in CST ranged from -14.12 to 33.58 µm, also not statistically significant (P > 0.05). The abstract indicates no measurable improvement in either VA or CST attributable to GLP-1RA initiation within the assessed follow-up.
The analysis used de-identified data from the IRIS Registry. The Partners Human Research institutional review evaluated the project and determined it did not meet the definition of human subjects research; continuing review was not required. The study was funded by the Massachusetts Eye and Ear Clinical Data Science Fund. The manuscript is posted as a medRxiv preprint and has not been peer-reviewed.
The abstract provides key design and outcome results but omits several details that could influence interpretation and generalizability. Specific GLP-1RA agents, dosing regimens, duration of GLP-1RA exposure, information on adherence, and concurrent systemic diabetes management were not reported in the abstract. Likewise, information on baseline anti-VEGF agent types, retreatment criteria, or clinic-level practice patterns is not described in the source abstract. Because the report is a preprint, findings should be interpreted cautiously until peer review and full manuscript details are available.
In a matched cohort of 19,792 patients with DME from the IRIS Registry, initiation of GLP-1RAs was not associated with a statistically significant change in monthly anti-VEGF injection frequency nor with significant changes in visual acuity or central subfield thickness over a 24-month post-initiation period. The study suggests no observable reduction in intravitreal treatment burden associated with GLP-1RA initiation in this registry-based analysis. As the report is a preprint and the abstract omits several granular clinical details, further peer-reviewed publication and additional data (including agent-specific and adherence information) are needed to fully interpret these findings for clinical practice.