BackgroundProliferative diabetic retinopathy is the most common cause of severe sight impairment in people with diabetes. Proliferative diabetic retinopathy has been managed by panretinal laser photocoagulation (PRP) for the past 40 years. We report the 1 year safety and efficacy of intravitreal aflibercept.MethodsIn this phase 2b, single-blind, non-inferiority trial (CLARITY), adults (aged ≥18 years) with type 1 or 2 diabetes and previously untreated or post-laser treated active proliferative diabetic retinopathy were recruited from 22 UK ophthalmic centres. Patients were randomly assigned (1:1) to repeated intravitreal aflibercept (2 mg/0·05 mL at baseline, 4 weeks, and 8 weeks, and from week 12 patients were reviewed every 4 weeks and aflibercept injections were given as needed) or PRP standard care (single spot or mutlispot laser at baseline, fractionated fortnightly thereafter, and from week 12 patients were assessed every 8 weeks and treated with PRP as needed) for 52 weeks. Randomisation was by minimisation with a web-based computer generated system. Primary outcome assessors were masked optometrists. The treating ophthalmologists and participants were not masked. The primary outcome was defined as a change in best-corrected visual acuity at 52 weeks with a linear mixed-effect model that estimated adjusted treatment effects at both 12 weeks and 52 weeks, having excluded fluctuations in best corrected visual acuity owing to vitreous haemorrhage. This modified intention-to-treat analysis was reapplied to the per protocol participants. The non-inferiority margin was prespecified as −5 Early Treatment Diabetic Retinopathy Study letters. Safety was assessed in all participants. This trial is registered with ISRCTN registry, number 32207582.FindingsWe recruited 232 participants (116 per group) between Aug 22, 2014 and Nov 30, 2015. 221 participants (112 in aflibercept group, 109 in PRP group) contributed to the modified intention-to-treat model, and 210 participants (104 in aflibercept group and 106 in PRP group) within per protocol. Aflibercept was non-inferior and superior to PRP in both the modified intention-to-treat population (mean best corrected visual acuity difference 3·9 letters [95% CI 2·3–5·6], p<0·0001) and the per-protocol population (4·0 letters [2·4–5·7], p<0·0001). There were no safety concerns. The 95% CI adjusted difference between groups was more than the prespecified acceptable margin of −5 letters at both 12 weeks and 52 weeks.InterpretationPatients with proliferative diabetic retinopathy who were treated with intravitreal aflibercept had an improved outcome at 1 year compared with those treated with PRP standard care.FundingThe Efficacy and Mechanism Evaluation Programme, a Medical Research Council and National Institute for Health Research partnership.