Abstract
PURPOSE: To compare the long-term incidence of recurrent vitreous hemorrhage(VH) and neovascular glaucoma (NVG) in patients with proliferative diabetic retinopathy (PDR) undergoing pars plana vitrectomy (PPV) alone versus combined PPV and phacoemulsification (PPV&P). METHODS: This large, single-center, retrospective cohort study analyzed 548 eyes (137 PPV&P, 411 PPV-alone) with a minimum 12-month follow-up. Kaplan-Meier survival analysis and multivariable Cox proportional hazards regression models were used to compare outcomes and identify independent predictors for complications within a 2-year follow-up period. A supplementary propensity score-matched (PSM) analysis was also performed to confirm the findings. RESULTS: Over the entire follow-up, the overall incidence of recurrence and NVG did not differ significantly between the groups. However, multivariable Cox regression analysis for the first two years revealed that PPV-alone was a significant protective factor against recurrence (adjusted Hazard Ratio [aHR], 0.61; 95% CI, 0.40-0.92), while severe preoperative traction was a major risk factor (aHR, 2.24; 95% CI, 1.39-3.60). For NVG development within two years, the surgical approach was not a significant factor. Instead, severe preoperative traction (aHR, 5.88; 95% CI, 1.67-20.68) and older age (aHR, 1.06; 95% CI, 1.01-1.11) were the primary independent risk factors.These findings were confirmed in a 1:1 age-matched PSM cohort. CONCLUSION: Combined PPV with phacoemulsification is associated with an increased risk of recurrence within the first two postoperative years but does not increase the risk of neovascular glaucoma. The severity of baseline proliferative disease, particularly the degree of traction, is a more critical determinant of postoperative complications than the surgical strategy regarding the lens. This supports tailoring the surgical approach based on individual patient cataract status and retinal pathology.