Residual vitreous cortex after surgical posterior vitreous separation visualized by intravitreous triamcinolone acetonide

Objective: To visualize the residual vitreous cortex (VC) on the retinal surface after surgical posterior vitreous separation (PVS) during a pars plana vitrectomy (PPV), especially in patients with diabetic retinopathy. Design: Case-control study. Participants: Patients with proliferative diabetic retinopathy (PDR), diabetic macular edema (DME), branch retinal vein occlusion (BRVO), and rhegmatogenous retinal detachment (RRD). Methods: A triamcinolone acetonide (TA)-assisted vitrectomy was performed on patients with the following diseases: PDR (40 eyes), DME (26 eyes), BRVO (11 eyes), and RRD (17 eyes). Eyes with no apparent preoperative posterior vitreous detachment were enrolled in this study. After performance of surgical PVS, the residual VC was visualized as a white gel highlighted by TA. Based on this finding, the residual VC pattern was then divided into 3 groups: (1) diffuse type (VC was diffusely present in the temporal vascular arcade), (2) focal type (a small island of VC was left), and (3) no residual VC. A multivariate analysis using analysis of variance was performed regarding the residual VC pattern, disease type, age, and the 3 different surgeons. Main Outcome Measures: Each surgeon determined the type of residual VC during the operation, and the results were confirmed by a postoperative review of the videotape records judged by the other 2 surgeons. Results: Eighty percent of the PDR eyes demonstrated the diffuse type; 10%, the focal type; and 10%, no residual VC. Fifty-eight percent of the eyes with DME demonstrated the diffuse type; 19%, the focal type; and 23%, no residual VC. Eighteen percent of the BRVO eyes showed the diffuse type; 24%, the focal type; and 59%, no residual VC. Thirty percent of the RRD eyes showed the diffuse type; 30%, the focal type; and 40%, no residual VC. A multivariate logistic regression analysis showed that PDR was a predictor of the diffuse type of residual VC in comparison to RRD (odds ratio = 8.42, 95% confidence interval = 2.07-34.3). Neither age nor the surgeon was a significant factor for a specific type of residual VC. Conclusions: Diabetic eyes more often demonstrated the diffuse type of residual VC, even after surgical PVS. This information may be valuable for surgeons performing a PPV on patients with diabetic retinopathy. (C) 2004 by the American Academy of Ophthalmology.