Intravitreal bevacizumab (IVB) versus IVB in combination with pars plana vitrectomy for vitreous hemorrhage secondary to proliferative diabetic retinopathy: a randomized clinical trial

BACKGROUND: The main purpose of this study is to compare the vitreous hemorrhage (VH) score reduction and visual acuity outcomes in patients with VH secondary to proliferative diabetic retinopathy (PDR) treated with intravitreal injections of bevacizumab (IVB) versus IVB and pars plana vitrectomy (IVB and PPV). METHODS: Patients with VH secondary to PDR were randomized into 2 groups: in Group A, patients were treated with a total of 3 IVB (1.5 mg/0.06 ml) at 8-week intervals; and in Group B, patients received a single IVB (1.5 mg/0.06 ml) and, 7 days later, underwent PPV. Patients received an ophthalmic evaluation that included best-corrected visual acuity (BCVA), indirect ophthalmoscopy, and mode B echography at weeks 8, 16 and 24. VH was classified according to the Diabetic Retinopathy Vitrectomy Study classification as grade 1, 2 or 3. Change in VH score was the primary outcome measure and change in BCVA was the secondary outcome. RESULTS: Seventy-three eyes of 66 patients were randomized and 70 eyes completed the 24-week follow-up visit. Mean VH score reduction (± SEM) of 0.4571 ± 0.0283 (p = 0.0014), 1.3429 ± 0.0393 (p < 0.0001) and 1.8286 ± 0.0438 (p < 0.001) was observed in Group A at 8, 16 and 24 weeks after treatment, respectively (Table 2; Fig. 2). In Group B, the reduction of VH score (± SEM) was 2.2571 ± 0.0720 (p = 0.0014), 2.2857 ± 0.0606 (p < 0.0001) and 2.2286 ± 0.0726 (p < 0.001) at 8, 16 and 24 weeks after treatment, respectively. Group comparison revealed a significantly greater reduction in mean VH score in Group B at 8 and 16 weeks after treatment (p < 0.0001). However, at 24 weeks this difference was no longer statistically significant (p = 0.1854). In Group A, mean (± SEM) BCVA showed an improvement of 0.00285 ± 0.0004 (p = 0.971), 0.5371 ± 0.0072 (p < 0.0001), 0.8143 ± 0.0001 (p < 0.0001) and 0.8543 ± 0.0008 (p < 0.0001) compared to baseline at 1, 8, 16 and 24 weeks after treatment, respectively. In Group B, mean (± SEM) BCVA showed an improvement of 0.3657 ± 0.0507 (p = 0.0002), 0.8857 ± 0.0385 (p < 0.0001), 0.9457 ± 0.0499 (p < 0.0001) and 0.9629 ± 0477 (p < 0.0001) compared to baseline at 1, 8, 16 and 24 weeks after treatment, respectively. No significant difference in BCVA improvement was observed between groups at 24 weeks after treatment. CONCLUSION: PPV with preoperative IVB is associated with more rapid clearance of VH and improvement in BCVA than IVB injections alone. However, after 24 weeks of follow-up, the reduction in VH score and BCVA were similar between both treatment strategies. Trial Registration The project is registered in Plataforma Brasil with CAAE number 927354.7.0000.5440 and was approved by the Ethics Committee of the Clinics Hospital of Ribeirao Preto Medicine School of São Paulo University—Ribeirão Preto, São Paulo, Brazil (appreciation number 3.053.397 gave the approval).