Excessive accumulation of retinol-based toxins has been implicated in the pathogenesis of geographic atrophy (GA). Fenretinide, an orally available drug that reduces retinol delivery to the eye through antagonism of serum retinol-binding protein (RBP), was used in a 2-year trial to determine whether retinol reduction would be effective in the management of geographic atrophy. The efficacy of fenretinide (100 and 300 mg daily, orally) to slow lesion growth in geographic atrophy patients was examined in a 2-year, placebo-controlled double-masked trial that enrolled 246 patients at 30 clinical sites in the United States. Fenretinide treatment produced dose-dependent reversible reductions in serum RBP-retinol that were associated with trends in reduced lesion growth rates. Patients in the 300 mg group who achieved serum retinol levels of ≤ 1 μM (≤ 2 mg/dL RBP) showed a mean reduction of 0.33 mm in the yearly lesion growth rate compared with subjects in the placebo group (1.70 mm/year vs. 2.03 mm/year, respectively, P = 0.1848). Retinol-binding protein reductions <2 mg/dL correlated with further reductions in lesion growth rates (r = 0.478). Fenretinide treatment also reduced the incidence of choroidal neovascularization (approximately 45% reduction in incidence rate in the combined fenretinide groups vs. placebo, P = 0.0606). This therapeutic effect was not dose dependent and is consistent with anti-angiogenic properties of fenretinide, which have been observed in other disease states. The findings of this study and the established safety profile of fenretinide in chronic dosing regimens warrant further study of fenretinide in the treatment of geographic atrophy.