Estimating the long-term visual field consequences of average daily intraocular pressure and variance : a clinical trial comparing timolol, latanoprost and travoprost

Abstract

Objective: To estimate the long-term consequences of average daily intraocular pressure (IOP) and variance in patients treated with travoprost, latanoprost or timolol.

Design and setting: IOP measures were derived from a 12-month randomised, double-masked, double-dummy, phase III clinical trial conducted in patients with primary open angle glaucoma (POAG) or ocular hypertension. IOP was measured at weeks 2, 12, 24 and 48, at 8am, 10am and 4pm. The visual field defect (VFD) risk functions estimated by Stewart were applied, the statistical unit being the eye. A second-order analysis of variance was performed including eye, time, treatment and investigator. Analyses according to baseline IOP were performed to identify patients at risk.

Interventions: IOP-lowering drugs: travoprost 0.004% once daily, latanoprost 0.005% once daily and timolol 0.5% twice daily.

Outcome measure: Daily IOP average and variance, and daily minimum and maximum IOP and VFD.

Results: Travoprost-treated patients (n = 200) had a daily IOP average significantly lower than timolol-treated patients (n = 200; p < 0.0001) and latanoprost-treated patients (n = 196; p < 0.001). Similar results were found for daily minimal and maximal IOP values. No difference was found for IOP variance between the prostaglandin analogues (p < 0.25), while timolol-treated patients had a higher value than travoprost-treated patients (p < 0.004). Consequently, patients treated with timolol would have a 1.54 greater chance of developing a new VFD over 5 years than those using latanoprost. Patients treated with latanoprost would have a 1.11 greater chance of developing a new VFD than travoprost-treated patients.

Conclusions: Better control of daily IOP average and variance should avoid new VFDs. Prostaglandins controlled the variance better than timolol, and travoprost was found to be the drug having the most effect on IOP average. Longitudinal prospective data collection should be performed to confirm these findings.