Seven RCTs (2,083 eyes, range 32 to 517) were included. The studies were generally of high quality: all were double blind; six reported observer masking; six reported a computer-generated randomisation schedule; and all used calibrated Goldmann tonometry with intraocular pressure measured at least twice at each time point.
At all the assessment periods, the pooled results fell below the non-inferiority level of 1.5mmHG. This suggested that fixed therapy was non inferior (had similar efficacy) to non-fixed therapy for reducing intraocular pressure.
There was a statistically significant difference between fixed and non-fixed therapies that favoured non-fixed therapies at two hours (mean difference 0.39, 95% CI: 0.04 to 0.75; six trials) and at eight hours (mean difference 0.50, 95% CI: 0.16 to 0.85; four trials), but not at baseline (mean difference 0.20, 95% CI: -0.11 to 0.51; six trials). Sensitivity analysis that compared intention to treat and per protocol data did not alter the results. There was no evidence of heterogeneity at any time point.
The main adverse effects were hyperaemia, ocular irritation and keratitis. One study reported significantly more cases of hyperaemia in the fixed compared to the non-fixed therapy group. In five studies there was a higher number of discontinuations due to intolerance in the fixed therapy group (1% to 12.5%).
The authors stated that the funnel plots showed no evidence of publication bias.