OBJECTIVES: First line treatment for diabetic macular edema (DME) patients tends to be intravitreal anti-vascular endothelial growth factor (anti-VEGF) therapy. However, many patients do not adequately respond to anti-VEGF therapy. Corticosteroid implants may be considered a second line therapy for these patients. No clear guidelines for switching a patient from anti-VEGF to corticosteroid implant therapy have been established.

METHODS: Electronic health record (EHR) data from 3 specialty ophthalmology networks and 6 integrated delivery networks in the OMNY Health real-world data platform from 2017 to 2023 were accessed. DME patients who initiated anti-VEGF therapy on or after their index diagnosis were selected if they had no prior corticosteroid implants. Switching was defined as initiation of corticosteroid implant therapy after anti-VEGF. Demographic characteristics were tabulated at index diagnosis. A cumulative incidence plot per index therapy was generated to characterize time from anti-VEGF initiation to switching or censoring.

RESULTS: Among 39,158 DME patients, 9,923 patients without history of corticosteroid implant initiated anti-VEGF therapy on or after their index diagnosis. Distributions of gender (55% male), race (25% nonwhite among known categories), and age (50% 65+ years, 43% 45-64 years, 7% ≤ 30 years) were tabulated. 70% of patients had index anti-VEGF of bevacizumab, 20% aflibercept, 5% ranibizumab, and 4% faricimab. Overall, 8% of patients experienced a therapy switch. Proportions for switching at 3, 6, 9, and 12 months were 1.5%, 3.2%, 4.5%, and 5.8% for bevacizumab; 2.7%, 5.7%, 8.6%, and 10% for aflibercept; 2.7%, 5.2%, 6.7%, and 8.3% for ranibizumab; and 1.7%, 3.7%, 7.8%, and 15% for faricimab, respectively.

CONCLUSIONS: Results provide insights into the real-world treatment experience of anti-VEGF initiators for DME. Bevacizumab had the lowest proportion of patients switching to corticosteroid implant therapy. Additional analyses accounting for severity would be beneficial for understanding the association with time to switching.