Factors Affecting the Results of Ahmed Glaucoma Valve Implantation in Diabetic Neovascular Glaucoma With or Without Previous Pars Plana Vitrectomy.

Objectives: The aim of this study was to compare the outcomes of Ahmed glaucoma valve implantation (AGVI) in neovascular glaucoma (NVG) due to proliferative diabetic retinopathy (PDR) with or without a pars plana vitrectomy (PPV) history and to analyze the factors affecting surgical failure.

Methods: Patients with NVG secondary to PDR undergoing AGVI at a single center were reviewed retrospectively. The surgical failure rates and post-operative complications were compared between eyes with (PPV-AGVI group) and without previous PPV (AGVI group). Failure was defined as loss of light perception or intraocular pressure (IOP) >17 mmHg despite maximum medication, or need of additional intervention for IOP control or for the management of complications. Survival analysis was investigated by Kaplan-Meier test. The possible factors for failure were analyzed with logistic regression analysis.

Results: The failure rates were 21.9% during the mean follow-up of 27.56±15.38 months and 14.3% during 23.63±12.38 months, in PPV-AGVI group (n=32) and in AGVI group (n=49), respectively (p=0.37). The frequency of complications and surgical intervention need for management of post-operative complications was similar between groups (p>0.05). There was no significant difference in survival analysis (p=0.13). The history of intravitreal anti-vascular endothelial growth factor (anti-VEGF) injection before AGVI was significantly associated with failure (odds ratio = 26.941, p=0.02).

Conclusion: The results of AGVI performed with long scleral tunnel technique were comparable in terms of failure rates, between NVG patients with and without previous diabetic vitrectomy. The only significant factor for failure was intravitreal anti-VEGF pre-treatment. This may be related to the necessity of anti-VEGF therapy in aggressive PDR, and also, anti-VEGF agents may increase fibrosis in the anterior chamber angle.