Efficacy and tolerability of prostaglandin analogs: a meta-analysis of randomized controlled clinical trials

Abstract

Aim: This systematic meta-analysis was performed to evaluate the intraocular pressure (IOP) lowering effects and tolerability of latanoprost, bimatoprost, and travoprost.

Methods: Clinical trials published up to July 2006 were thoroughly searched using all available databases and resources. The inclusion criteria were prospective randomized controlled clinical trials; patients with primary open-angle glaucoma or ocular hypertension; and prostaglandin monotherapy, without systemic/ocular medications or laser/surgery that could affect IOP within the past 3 months. Study quality was assessed with the Jadad scoring system, and potential bias was eliminated by robust statistical and independent reviews of publications. The main outcome measures were efficacy assessed by IOP (taken at 8 AM, noon, 4 PM, and 8 PM) change at 3 months from baseline and tolerability assessed by the incidence of conjunctival hyperemia.

Results: Eight trials were identified (n=1610 patients). IOP change from baseline was statistically significantly greatest with bimatoprost, compared with latanoprost at all time points [weighted mean (WM) 8 AM: WM=0.50 mm Hg; P=0.05; 95% confidence intervals (CIs) 0.01-0.99; noon: WM=1.17 mm Hg; P<0.001; 95% CI 0.68-1.66; 4 PM: WM=0.78 mm Hg; P=0.003; 95% CI 0.26-1.29; 8 PM: WM=0.67 mm Hg; P=0.04; 95% CI 0.02-1.32], and with travoprost during the daytime (8 AM: WM=1.02 mm Hg; P=0.004; 95% CI 0.32-1.72; noon: WM=0.86 mm Hg; P=0.02; 95% CI 0.12-1.59). Latanoprost and travoprost were comparable in their ability to reduce IOP at all time points (P<or=0.82). The incidence of hyperemia was less with latanoprost and travoprost [latanoprost vs. bimatoprost: relative risk=0.59; P<0.001; 95% CI 0.50-0.69; travoprost vs. bimatoprost: relative risk=0.84; P=0.05; 95% CI 0.70-1.00].

Conclusions: The findings suggest a greater efficacy of bimatoprost compared with latanoprost and travoprost, although the incidence of hyperemia was lower with the latter 2 agents.