Placebo Effect and Its Determinants in Ocular Hypotensive Therapy: Meta-analysis and Multiple Meta-regression Analysis.

ABSTRACT

TOPIC The placebo effect and its potential determinants in ocular hypotensive therapy. CLINICAL RELEVANCE The placebo effect has been studied and documented within a wide clinical context. It remains unclear whether placebo is effective in glaucoma treatment or, if so, which factors are determinative of effect size (ES). METHODS: Randomized controlled trials (RCTs) of topical ocular hypotensive therapy for patients with open-angle glaucoma or ocular hypertension, conducted through June 2, 2022, were included. First, a perceived placebo effect was measured as the overall intraocular pressure (IOP) change from the baseline. It was evaluated in terms of the ES (mean difference between the baseline and the end point) and then was compared with the ES, as obtained from the untreated control participant to obtain a true placebo effect. The primary outcome was ES based on 4 weeks of treatment. Meta-analysis-based statistical pooling was performed where appropriate, and 95% confidence intervals (CIs) were used for comparison. Potential placebo effect determinants were scrutinized using a multiple meta-regression model (PROSPERO identifier, CRD42022348098). RESULTS: A total of 40 RCTs (7829 eyes) with 33 placebo groups (2055 eyes) along with 7 untreated groups (1184 eyes) were included. Among placebo-controlled trials, placebo was determined to be effective in lowering IOP (ES, -1.30 mmHg; 95% CI, -1.75 to -0.84 mmHg). Using NMA, the ES for placebo was -2.27 mmHg (95% CI, -3.52 to -1.01 mmHg) greater than ES for untreated control participants.. According to the multiple meta-regression model, the active treatment ES was a significant factor to predict the amount of placebo effect. Placebo additionally lowered IOP by -0.45 mmHg per -1 mmHg of active treatment effect. Add-on study design and larger sample size also were associated with greater amount of placebo effect. No publication bias was evident in either a funnel plot or the Begg and Mazumdar adjusted rank correlation test results (P = 0.24). DISCUSSION: This meta-analysis indicated that placebo is effective in lowering IOP and is superior to the effect observed for the untreated control participants. However, caution is required in interpreting the results because of the small number of untreated controlled trials and potential bias from the lack of direct comparison between the placebo and untreated arms. FINANCIAL DISCLOSURE(S) The author(s) have no proprietary or commercial interest in any materials discussed in this article.