Effects of vitrectomy on age-related macular degeneration.

Abstract

To determine whether vitrectomy alters the long-term progression of age-related macular degeneration (AMD). Retrospective case-control study. Forty-four eyes of 22 patients with AMD who underwent vitrectomy in 1 eye were included in the study. The progression of AMD at follow-up in the 22 eyes that underwent vitrectomy was compared with the 22 fellow, nonvitrectomized eyes. The charts and photographs of subjects with Age-Related Eye Disease Study category 3 AMD in both eyes who previously underwent vitrectomy surgery for an epiretinal membrane or macular hole were reviewed. Subjects were excluded if they had had a vitrectomy in both eyes, had <2 years of follow-up, had previous choroidal neovascularization (CNV), retinal detachment, diabetic retinopathy, angioid streaks, high myopia, vascular occlusions, or extensive macular scarring in either eye, or insufficient hospital records or photographs to determine the extent of AMD. Clinical notes throughout the follow-up interval were reviewed. Two vitreoretinal specialists independently graded pre- and postvitrectomy fundus photographs of all eyes in a masked fashion. The development or progression of geographic atrophy of the retinal pigment epithelium and the development of CNV. Twenty-two patients were included. The average follow up interval was 5.5 years (range, 2-15). Choroidal neovascularization developed in 5 control eyes and in 2 vitrectomized eyes, and atrophy developed in 7 control and 4 vitrectomized eyes. The difference between vitrectomized eyes and fellow eyes for the combined end points of RPE geographic atrophy or CNV was significant (P = 0.02). In this pilot study, we did not detect that vitrectomy increased the progression of AMD. In fact, it was associated with a reduced progression to geographic atrophy or CNV. Additional studies are needed to confirm or refute this association. The authors have no proprietary or commercial interest in any of the materials discussed in this article.