To provide clinical management guidelines for eyes with central retinal vein occlusion. Prospective cohort study with randomized clinical trials of specific subgroups of patients. Three-year follow-up every 4 months. Nine ophthalmology practices. Seven hundred twenty-five patients with central vein occlusion. Iris neovascularization (INV), neovascular glaucoma, and visual acuity. Visual acuity outcome was largely dependent on initial acuity. Sixty-five percent of patients with initially good visual acuity (20/40 or better) maintained visual acuity in the same range at the end of the study. Patients with intermediate initial acuity (20/50-20/200) showed a variable outcome: 19% improved to better than 20/50, 44% stayed in the intermediate group, and 37% had final visual acuity worse than 20/200. Patients who had poor visual acuity at the first visit (< 20/200) had an 80% chance of having a visual acuity less than 20/200 at final visit, whether perfused or nonperfused initially. In the first 4 months of follow-up, 81 (15%) of the 547 eyes with perfusion converted to ischemia. During the next 32 months of follow-up, an additional 19% of eyes were found to have converted to ischemia for a total of 34% after 3 years. The development of nonperfusion or ischemia was most rapid in the first 4 months and progressed continuously throughout the entire duration of follow-up. Iris neovascularization of at least 2-clock hours, and/or angle neovascularization (ANV) developed in 117 (16%) of the 714 eyes. Sixty-one of the 117 eyes that had INV/ANV were initially categorized as nonperfused or indeterminate; 56 of the 117 eyes were initially categorized as perfused. When INV/ANV occurred, it was treated promptly with panretinal photocoagulation. The strongest predictors of INV/ANV were visual acuity (P < .001) and the amount of nonperfusion seen by fluorescein angiogram (P < .001). For eyes initially categorized as nonperfused or indeterminate, 35% (61/176) developed INV/ANV, compared with 10% (56/538) for eyes initially categorized as perfused. Other risk factors were venous tortuosity (P = .02), extensive retinal hemorrhage (P = .07), and duration less than 1 month (P = .08). Neovascular glaucoma that was unsuccessfully managed with medical treatment developed in only 10 eyes. No eye was enucleated. Visual acuity at baseline is a strong predictor of visual acuity at 3 years for eyes with good vision and eyes with poor vision, but a poor predictor for intermediate acuities. Visual acuity is also a strong predictor for the development of INV/ANV, as is nonperfusion. During the course of follow-up, one third of the eyes with perfusion converted to eyes with ischemia. Clinical management guidelines, developed from these and previously reported Central Vein Occlusion Study data, are presented.