Intravitreal Aflibercept for Retinal Nonperfusion in Proliferative Diabetic Retinopathy: Outcomes from the Randomized RECOVERY Trial.

Purpose: Evaluate the impact of intravitreal aflibercept (Eylea; Regeneron, Tarrytown, NY) on retinal nonperfusion (RNP) in eyes with proliferative diabetic retinopathy (PDR).

Design: Prospective, randomized clinical trial.

Participants: Eyes with treatment-naïve PDR and extensive RNP without diabetic macular edema.

Methods: Patients were randomized 1:1 to intravitreal 2 mg aflibercept every 4 weeks (monthly) or every 12 weeks (quarterly).

Main Outcome Measures: The primary outcome measure was change in total RNP area (in square millimeters) from baseline to year 1. Secondary outcomes included ischemic index (ISI), diabetic retinopathy severity scale (DRSS) scores, visual acuity, central retinal thickness, and adverse events. The mean and 95% confidence interval were calculated for each outcome.

Results: Through 1 year, the monthly (n = 20) and quarterly (n = 20) cohorts received 11.0 and 3.95 mean aflibercept injections, and DRSS scores improved 2 steps or more in 74% and 67% of patients, respectively. Among all patients through 1 year, mean total area of RNP increased from 235 mm to 266 mm (P = 0.18) and ISI increased from 25.8% to 31.9% (P = 0.004). Retinal nonperfusion outcomes favored monthly dosing. Mean total RNP increased from 207 mm at baseline to 268 mm (P = 0.01) at 1 year in the quarterly cohort and remained stable at 264 mm at baseline and 1 year (P = 0.70) in the monthly cohort (P = 0.05, monthly vs. quarterly cohorts). Although many eyes demonstrated increased areas of RNP longitudinally (n = 24 [66.7%]), this was more common with quarterly dosing (n = 14 [77.8%]), and a proportion of eyes (n = 12 [33.3%]) demonstrated localized areas of apparent reperfusion of nonperfused retina, more commonly in the monthly cohort (n = 8 [44.4%]).

Conclusions: Widespread evidence of retinal reperfusion with aflibercept dosing of PDR eyes with extensive RNP was not identified, and therefore the primary outcome of the current study was not met. Nevertheless, zones of apparent reperfusion were detected in some patients, and a dose response was identified with a reduction of RNP progression with monthly compared to quarterly dosing.