One hundred seven patients with intraocular pressures (IOPs) between 22 and 28 mmHg with normal visual fields on Goldmann perimetry and without evidence of optic nerve damage were randomly assigned to either a timolol treatment (TT) or a no treatment (NT) arm in a prospective clinical trial. The patients were followed for an average of 56 and 51 months, respectively. Criteria for failure were a confirmed IOP of greater than 32 mmHg, stereophotographically documented optic nerve progression, or development of glaucomatous visual field loss by Goldmann or Octopus perimetry. Nine patients failed in the TT group and 17 in the NT group (P = 0.07). Of the nine TT group failures, six had discontinued timolol before failure (4 for greater than 6 months). In a Cox proportional hazards analysis controlling for confounding variables, timolol was found to be significantly protective with an adjusted risk ratio of 0.38 (95% confidence interval = 0.16-0.89, P = 0.03). When only field and disc failure criteria were considered, timolol treatment was found to be significantly protective in an analysis considering patients who stopped timolol as being lost to follow-up (P = 0.05). A higher tonographic facility of outflow was protective in all analyses. A trend toward a substantial loss of effectiveness of timolol on IOP was not observed. Seasonal fluctuations in IOP were observed (P = 0.0007), with higher IOP occurring in the winter. The results demonstrate a favorable influence of timolol therapy on the clinical course of patients with mildly elevated IOP.(ABSTRACT TRUNCATED AT 250 WORDS)