Ten RCTs (458 eyes in 397 patients) were included in the review. Allocation concealment was adequate in six studies. Withdrawals ranged from 2% to 17% (where reported). The same surgeon performed both surgeries in six studies. Follow-up ranged from six months to 3.3 years.
Efficacy: Trabeculectomy was associated with a significantly lower mean intraocular pressure at six, 12 and 24 months when compared to viscocanalostomy (six months MD 2.25mmHg, 95% CI 1.38 to 3.12; eight studies, 12 months MD 3.64mmHg, 95% CI 2.74 to 4.54; six studies and 24 months MD 3.42mmHg, 95% CI 1.80 to 5.03; three studies).
Trabeculectomy was consistently superior to viscocanalostomy in all subgroup analyses. Trabeculectomy was associated with significantly fewer antiglaucomatous medications postoperatively than viscocanalostomy (MD 0.93, 95% CI 0.81 to 1.04; four studies). Substantial heterogeneity was identified for the six-month overall analysis of intraocular pressure (and some subgroup analyses) and number of medications.
Safety: Viscocanalostomy had a significantly higher risk of perforation of Descemet membrane than trabeculectomy (RR 7.72, 95% CI 2.37 to 25.12; six studies). Viscocanalostomy had a significantly lower risk of other adverse events (hypotony RR 0.29, 95% CI 0.15 to 0.58; nine studies, hyphema RR 0.50, 95% CI 0.30 to 0.84; nine studies, shallow anterior chamber RR 0.19, 95% CI 0.08 to 0.45; nine studies and cataract formation RR 0.31, 95% CI 0.15 to 0.64; eight studies). No heterogeneity was identified in any of the safety analyses.
A funnel plot of the difference in intraocular pressure did not suggest potential for publication bias.