Importance This study could facilitate the generation of automated quantitative assessments, providing clinicians and clinical trials with a continuous severity scale of retinal leakage to improve the prediction of future diabetic retinopathy (DR) worsening. Objective To examine ultra-widefield fluorescein angiography (UWF-FA) retinal leakage associated with DR progression and complications. Design, Setting, and Participants This cohort study was a post hoc analysis of the DRCR Retina Network Protocol AA clinical trial. Participants were adults with nonproliferative diabetic retinopathy (NPDR). Data was collected from February 2015 to March 2020. Analysis was performed June 2023 to April 2024. Interventions Treatment of DR or diabetic macular edema was at the investigator’s discretion. Main Outcomes and Measures Time to Diabetic Retinopathy Severity Scale (DRSS) worsening of 2 or more steps, vitreous hemorrhage (VH), and proliferative diabetic retinopathy (PDR) over 4 years. Baseline and 1-year leakage index, calculated as the area with leakage divided by the total area in the analyzable retina on UWF-FA, were evaluated as risk factors. Results Among 537 eyes (363 participants; mean SD age, 61 12 years; 183 50% male, 180 50% female), the baseline leakage index (mean SD, 3.5% 3.9%) was associated with DRSS worsening over 4 years (hazard ratio HR for 1% increase, 1.09; 95% CI, 1.05-1.13; P lt; .001). At 1 year, 114 of 453 eyes (25%) had an absolute leakage increase 1% or greater from baseline, which was associated with a higher risk of DRSS worsening over 4 years (61% vs 33%; HR, 2.63; 95% CI, 1.90-3.64; P lt; .001). Baseline leakage and a 1% or greater increase at 1 year were also associated with development of VH and PDR. Conclusions and Relevance In eyes with NPDR, higher leakage at baseline and an increase from baseline to 1 year were associated with greater risk of DRSS worsening and vision-threatening complications, including VH and PDR. These findings support quantitative leakage assessment at baseline and over time as a promising biomarker for predicting the severity and progression of diabetic retinopathy.
Reese et al. (Thu,) studied this question.