To investigate the role of increased polyol pathway activity and hemodynamic deficits in the pathogenesis of diabetic retinopathy in non-insulin-dependent diabetes mellitus (NIDDM), Otsuka Long-Evans Tokushima fatty (OLETF) rats, an animal model of human NIDDM, were given water with or without 30% sucrose and some of them were fed laboratory chow containing 0.03% cilostazol, an anticoagulant, or 0.05% [5-(3-thienyl)tetrazol-1-yl] acetic acid monohydrate (TAT), an aldose reductase inhibitor, for 8 wk. Long-Evans Tokushima Otsuka (LETO) rats were used as nondiabetic controls. The peak latencies of oscillatory potentials of the electroretinogram in sucrose-fed OLETF rats were significantly prolonged compared with those in OLETF rats without sucrose feeding and LETO rats. There was a marked increase in platelet aggregability and a significant decrease in erythrocyte 2,3-diphosphoglycerate in sucrose-fed OLETF rats. Cilostazol significantly improved these parameters without changes in retinal levels of sorbitol and fructose. TAT, however, ameliorated all of these parameters. These findings confirm that the sucrose-fed OLETF rat is a useful animal model of retinopathy in human NIDDM and suggest that cilostazol improved diabetic retinopathy by modifying vascular factors, not by altering polyol pathway activity.