Fourteen trials involving 1,824 participants were included.
It was often unclear whether the studies met the quality criteria. Adequate allocation concealment could be determined in 7 trials, double-blinding in 4 trials, and intention-to-treat analysis in 5 trials. The withdrawal rates ranged from 0 to 19%.
For IOP reduction, there was a statistically significant effect in favour of latanoprost (post-treatment WMD 1.10, 95% CI: 0.57, 1.63). Significant heterogeneity was present (chi-squared 38.29, p=0.001, I-squared 66%), but only 2 trials did not show a greater IOP reduction with latanoprost than with brimonidine.
In the subgroup analyses, treatment durations of 6 months or more showed greater IOP reduction (WMD 1.64, 95% CI: 0.92, 2.36) than shorter treatment durations (WMD 0.76, 95% CI: 0.12, 1.39). There were no significant differences between patients with different diagnoses.
Superiority of latanoprost was greater (WMD 1.56, 95% CI: 0.90, 2.23) when the drugs were used as monotherapy than when they were used in combination with other glaucoma treatments (WMD 0.58, 95% CI: -0.04, 1.19).
In the 3 trials that reported measures of ocular haemodynamics, latanoprost significantly increased ocular blood flow, peak systolic velocity of the ophthalmic artery, and ocular perfusion pressure, while brimonidine had no significant effect.
A funnel plot did not show evidence of publication bias.
Risk for fatigue was statistically significantly lower with latanoprost than with brimonidine (RR 0.27, 95% CI: 0.08, 0.88). In the trials that reported fatigue, 5% of brimonidine patients and 1% of latanoprost patients experienced fatigue. Other adverse events did not show significant differences between the groups.