Risk of bias for analysis 1.2 Mean IOP changes from baseline: sensitivity analysis.
| Study | Bias | |||||||||||
| Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | |||||||
| Authors' judgement | Support for judgement | Authors' judgement | Support for judgement | Authors' judgement | Support for judgement | Authors' judgement | Support for judgement | Authors' judgement | Support for judgement | Authors' judgement | Support for judgement | |
| Subgroup 1.2.1 Netarsudil 0.02% once per day | ||||||||||||
| Araie 2021 | Low risk of bias | All the eligible patients were randomized (1:1:1:1) by a computer‐generated randomization list using an interactive web response system (...). There was no statistically significant difference in demographics and other baseline characteristics across the treatment groups (Table 1). | Low risk of bias | Patients and designated study site personnel (investigators, sponsor, monitor, data manager, statistician, and personnel involved in study management) were fully masked to treatment assignments. | Low risk of bias | All randomized persons were included in the ITT. No lost to follow up. | Low risk of bias | The IOP was measured at screening visit, qualification visit and all study visits using a calibrated Goldmann applanation tonometer (the most clinically accurate and the standard tonometer used in the diagnosis and treatment of glaucoma). | Low risk of bias | Protocol published on clinicaltrials.gov. Primary efficacy endpoint (IOP) reported as a number, which is the only way to report that outcome. All IOP measurements were listed in NCT whereas IOP measurements in the published paper were modified ITT results although well described. The mITT population may affect the results in the following manners: The ANCOVA model explains additional variability (i.e., reduces standard error) through inclusion of baseline covariates and therefore improves statistical power. The LS means calculated from this model are the means adjusted for the covariate. LS means are better estimates of true population means compared to arithmetic means. |
Low risk of bias | Based on the low risk of bias judgement in all domains, the overall judgement is low risk of bias. |
| Peace 2021 | Low risk of bias | Randomization schedule prepared by the sponsor representative. The study centre was provided with a predefined order sequence (i.e., kit number) for the dispensation of the study drug to patients. Investigator, subjects, sponsor, sponsor representative, and clinical monitors were not allowed to know the treatment assigned to each subject. Active and vehicle study medications packaged in identical bottles with identical labels. . No obvious differences in appearance(e.g., colour or viscosity) between the netarsudil and vehicle solutions. No apparent differences in baseline characteristics between treatment groups. | Low risk of bias | Double‐masked. The intent to treat (ITT) population served as the primary efficacy population. All randomized subjects who received at least one dose of the study drug were included in the ITT population. Subjects in the ITT population were analyzed in accordance with their assigned randomized treatment, even if the actual treatment the subject received was different from the planned treatment (...). Of the 12 patients randomized, 100% received their assigned study medication and therefore comprised the ITT population. | Low risk of bias | All 12 patients enrolled completed the study (8:4). | Low risk of bias | IOP was measured at each time point by study staff masked to treatment using a Perkins tonometer. The method of IOP measurements are well described, e.g. head position, and number of measurements. | Low risk of bias | Study registered at ClinicalTrials.gov before study initiation with a priori analysis described and outcome reported in the publication as planned. Primary efficacy endpoint (IOP) reported as a number which is the only way to report that outcome. | Low risk of bias | Judged as low risk of bias in all domains. Thus, the overall risk of bias is judged as low. |
| Subgroup 1.2.2 Ripasudil 0.4% twice per day | ||||||||||||
| Tanihara 2013 | Low risk of bias | Permuted blocks method with allocation ratio of 1:1:1:1. Site investigators were not informed about the block size throughout the study period. Allocation concealment not described. Even though statistical comparisons of baseline characteristics are not performed, the intervention groups seem alike (table 1). | Some concerns | Double‐masked. Site investigators not informed about block size throughout the study period. A total of 7 patients were excluded from analyses because of adverse events in 6 and protocol deviations in 1. In the remaining 203 patients, 52 in the placebo group, 50 in the 0.1% group, 52 in the 0.2% group, and 49 in the 0.4% group, subsequent statistical analyses were conducted for the estimation of efficacy in IOP reduction. | Low risk of bias | A total of 7 patients were excluded from analyses because of adverse events in 6 and protocol deviations in 1. In the remaining 203 patients, 52 in the placebo group, 50 in the 0.1% group, 52 in the 0.2% group, and 49 in the 0.4% group, subsequent statistical analyses were conducted for the estimation of efficacy in IOP reduction. | Low risk of bias | A calibrated Goldmann applanation tonometer (the most clinically accurate and the standard tonometer used in the diagnosis and treatment of glaucoma [3]) was used for IOP measurements. Double‐masked. Site investigators were not informed about the block site throughout the study period. |
Low risk of bias | Protocol JAPIC 101015 and a priori analysis plan were conducted. Primary efficacy endpoint (IOP) reported as a number, which is the only way to report that outcome. Reports covariance adjusted timepoint‐specific means pre‐and post‐administration. Pre‐administration values extracted. |
Some concerns | Based on the judgment of some concerns in Domain 2, the overall risk of bias was judged as having some concerns. |