One hundred forty-three patients with intraocular pressures (IOPs) above 22 mmHg and without visual field defects or any obvious evidence of optic nerve damage were randomly assigned to either a timolol treatment group or no treatment in a 6-year prospective clinical trial. Endpoints were defined as reproducible visual field defects on automatic perimetry, disc hemorrhages, or stereophotographically documented optic nerve head changes. Endpoints developed in 42 patients: 28 visual field defects, 8 changes in disc appearance, and 6 disc hemorrhages. Of the 42 patients, 20 were treated and 22 were not. Survival analysis showed no statistically significant differences in failure time to any endpoints between the two groups. In the untreated group, the time to failure of disc change was related to the mean IOP during the study and also to the changes in the IOP from baseline. A significant correlation was found between initial cup-to-disc ratio and survival time to visual field defects in the untreated group.