PURPOSE: To evaluate the safety and efficacy of fibrin glue in preventing early recurrence of vitreous hemorrhage following surgery for proliferative diabetic retinopathy (PDR). DESIGN: Single-masked randomized controlled clinical trial. SUBJECTS: Consecutive patients with vitreous hemorrhage due to PDR undergoing primary vitreoretinal surgery were screened. After completing all vitreoretinal maneuvers including endocautery to bleeders, infusion pressure was gradually reduced to 5 mm Hg. Eyes with persistent bleeders at the macula and/or the optic nerve head not amenable to endocautery were included. Eyes with tractional retinal detachment more than 5-disc-diameter areas and/or retinal breaks were excluded. METHODS: Twenty eyes were randomized into 2 groups. Controls were treated with conventional surgery, whereas cases were additionally treated with fibrin glue application over the posterior bleeding sites. MAIN OUTCOME MEASURES: Vitreous haze due to early recurrent vitreous hemorrhage was graded at weeks 1 and 4 as the main efficacy outcome measure. Absence of macular membranes at 3 months was the main safety outcome measure. RESULTS: Mean age, frequency of preoperative anti-vascular endothelial growth factor injection or retinal photocoagulation, antiplatelet usage, and mean intraoperative blood pressures were similar between the 2 groups. Mean surgery duration (54.6 vs 51.6 minutes) and frequency of fibrous proliferation >
2-disc-diameter areas (7 each) was similar. Fine dissection with forceps and/or scissors was required in more cases than controls (6 vs 3). Macular membranes were noted in 1 participant in each group. Mean grade of vitreous haze was significantly higher in controls than cases at week 1 (2.4 vs 0.6, P = .027) and week 4 (2 vs 0, P = .029). Cases had significantly better odds of having optically clear vitreous cavity at both week 1 (odds ratio [OR] 8.167, CI 1.02-64.03
P = .047) and week 4 (OR 33, CI 1.43-760.67
P = .029). Repeat retinal intervention for vitreous hemorrhage within 3 months was required in 5 of 10 controls, and in none of the cases. CONCLUSION: Our findings indicate fibrin glue application under direct visualization to be safe and efficacious in preventing recurrent vitreous hemorrhage for PDR in the early postoperative period.