Preoperative Anti-VEGF and the Cumulative Risk of Post-operative Vitreous Hemorrhage in PDR: A 2-Year Survival Analysis and Evaluation of Surgical Burden

Purpose To investigate the long-term risk factors for post-operative vitreous hemorrhage (VH) within 24 months following vitrectomy for proliferative diabetic retinopathy (PDR), evaluate the protective efficacy of preoperative anti-VEGF therapy, and quantify the associated secondary surgical burden. Methods This retrospective cohort study included 735 eyes from 735 PDR patients. A 24-month survival analysis was employed to categorize patients into VH (n = 179) and non-VH (n = 556) groups. Independent predictors were identified using univariate and multivariate Cox proportional hazards models. A sensitivity analysis was performed on a "pure hemorrhage model" (n = 676) by excluding eyes with post-operative tractional retinal detachment (TRD) to isolate the direct vascular-stabilizing effect of anti-VEGF. Results The overall cumulative incidence of post-operative VH was 18.4% at 12 months and 25.7% at 24 months. Multivariate Cox analysis revealed that preoperative anti-VEGF injection was the most potent protective factor, associated with a 63% reduction in VH hazard (HR 0.37; 95% CI, 0.27–0.51; P < 0.001). This protective effect remained robust in the sensitivity analysis (HR 0.32; P < 0.001), indicating a direct vascular stabilization independent of anatomical success. Independent risk factors included severe fibrovascular traction (HR 1.93; P = 0.003), younger age, higher HbA1c, and elevated serum creatinine (all P < 0.05). Patients in the VH group had significantly poorer final visual outcomes (1.33 ± 0.75 vs. 0.92 ± 0.60 LogMAR; P < 0.001) and a markedly higher reoperation rate (31.3% vs. 7.2%). Notably, hemorrhage-related indications accounted for 58.4% of the total secondary surgical workload. Conclusion Post-operative VH is a dominant driver of long-term visual impairment and secondary surgical burden in PDR. Preoperative anti-VEGF confers a direct, sustained protective effect by stabilizing the retinal microvasculature. Effective prevention of VH could potentially eliminate nearly 60% of the reoperation requirements in this population.