Treatment-Emergent Epiretinal Membrane in Diabetic Macular Edema: A Real-World OCT-Based Comparative Pharmacotherapy...

Purpose: To evaluate the incidence, predictors, and clinical impact of treatment-emergent epiretinal membrane (ERM) in diabetic macular edema (DME) treated with dexamethasone implant, aflibercept, ranibizumab, bevacizumab, or faricimab. Methods: In this multicenter, retrospective, nonrandomized study, 501 eyes with center-involving DME were assessed and 423 eyes without ERM on pretreatment OCT and with 24-month follow-up were included. Eyes were classified by the intravitreal agent selected by the treating physician in routine care and maintained throughout follow-up; eyes requiring a drug switch or retinal laser/vitrectomy during follow-up were excluded. The last eligible pre-index SD-OCT confirmed the absence of ERM; BCVA, central subfield thickness (CST), and vitreomacular-interface (VMI) morphology were assessed at baseline and 6, 12, 18, and 24 months. Crude comparisons, multivariable logistic regression, and Kaplan-Meier/log-rank analyses were performed. Results: ERM developed in 36/423 eyes (8.5%). Incidence was 16.0% with dexamethasone implant, 7.5% with aflibercept, 7.7% with ranibizumab, 7.4% with bevacizumab, and 4.3% with faricimab. Dexamethasone implant showed higher ERM odds than pooled non-dexamethasone therapy after adjustment (OR 2.24, 95% CI 1.02-4.92; p = 0.044). Baseline VMI abnormality, severe NPDR/PDR, prior PRP, and higher baseline CST were also associated with ERM. Eyes developing ERM had smaller BCVA gains (+2.8 vs. +7.1 ETDRS letters; p < 0.001) and CST reductions (88.6 vs. 130.8 µm; p = 0.004). Conclusions: Treatment-emergent ERM was associated with baseline ocular characteristics and treatment category and with poorer 24-month outcomes. These observational drug-group signals require prospective confirmation.