Analysis of Risk Factors for Revitrectomy in Eyes with Diabetic Vitreous Hemorrhage.

Abstract

We aimed to investigate the risk factors associated with revitrectomy in eyes with diabetic vitreous hemorrhage and to determine the prognosis of these patients at least one year postoperatively. This retrospective case-control study had a minimum follow-up period of one year. Patients were divided into single vitrectomy group (control group, n=202) and revitrectomy group (case group, n=36) for analysis. The indications, number, and timing of revitrectomies were documented. And the revitrectomy group was further divided into two vitrectomies group (n=30) and three or more vitrectomies group (n=6). The best-corrected visual acuity (BCVA) at the last follow-up and the occurrence of neovascular glaucoma (NVG) were compared among the single vitrectomy, two vitrectomies and three or more vitrectomies groups. We conducted a thorough collection of patient data and used univariate and binary logistic regression analyses to identify the risk factors associated with revitrectomy. A total of 197 patients (238 eyes) were included. Thirty-six eyes (15.1%) required revitrectomy with six eyes (2.5%) undergoing three or more vitrectomies during the follow-up period. The median duration of the second vitrectomy was 3 (2-6) months. The indications for a second vitrectomy included 28 eyes (77.8%) of postoperative vitreous hemorrhage and 7 eyes (22.2%) combined with tractional retinal detachment. Patients undergoing three or more vitrectomies had significantly worse postoperative BCVA and a higher incidence of NVG (P<0.01). Fibrinogen> 4 g/L (P<0.001) and preoperative anti-vascular endothelial growth factor intravitreal injection (P=0.015) were independent risk factors for revitrectomy, and glycated hemoglobin A1c (HbA1c)>10% (P=0.049) showed significant difference only in univariate analysis. Patients requiring revitrectomy tended to have higher fibrinogen levels, tightly adhered fibrovascular membranes, higher HbA1c levels, and worse prognoses.