Risk of bias for analysis 2.1 Mean IOP changes from baseline.
| 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 | |
| Kahook 2019 (ROCKET‐2) | Low risk of bias | Randomized by a computer‐generated method. No statistical baseline differences between groups (table 1, Serle et al 2018). | High risk of bias | Double‐masked:Patients and designated study site personnel were fully masked to treatment assignments. A vehicle bottle was provided for AM dosing in the netarsudil q.d. PM treatment groups to maintain masking. In order to include the results from the most patients, we used the results from the NCT (PP with no IOP‐restrictions; Netarsudil (QD) 206/251, timolol (BID) 217/251 = All subjects who did not have a major protocol violation likely to seriously affect the primary outcome of the study). | High risk of bias | Article reports results from PP (PP population with maximum baseline IOP < 25 mm Hg, 3 months). In order to include the results from the most patients, we used the results from the NCT (PP with IOP‐restrictions; Netarsudil (QD) 206/251, timolol (BID) 217/251 = All subjects who did not have a major protocol violation likely to seriously affect the primary outcome of the study). In the original report at 3 months, 82% (205/251), 60% (153/254), and 94% (237/251) of patients in the netarsudil q.d., netarsudil b.i.d., and timolol b.i.d. groups, respectively, completed the first 3 months. | Low risk of bias | IOP was measured using a calibrated Goldmann applanation tonometer. Two consecutive IOP measurements of each eye were obtained. If the 2 measurements differed by more than 2 mm Hg, a third measurement was obtained. IOP was to be recorded as the mean of 2 measurements or as the median of 3 measurements. | Low risk of bias | Statistical and trial plan published before the study initiation and followed. Primary efficacy endpoint (IOP) reportet as a number ‐ only way to report that outcome. | High risk of bias | Judged as high risk of bias in Domain 2 and 3. Thus, the overall risk of bias was judged as high. |
| Khouri 2019 (ROCKET‐4) | Low risk of bias | Patients were randomized (1:1) by a computer‐generated method. Randomization was stratified by study site and maximum baseline IOP (< 25 mm Hg vs ≥ 25 mmHg). No description of allocation concealment. No significant differences in baseline values (table 1). | High risk of bias | Double‐masked. 708 patients were randomized, but only 186 in each treatment arm were included in the per‐protocol analysis ( = 1) no serious protocol violations 2) baseline IOP < 25 mmHg). Data is extracted from NCT (PP for overall study population = 306 (netarsudil) and 316 (timolol)). According to the NCT, 243 netarsudil and 314 timolol patients completed the study, respectively. Thus, a substantial amount of patients that per protocol definition didn't complete the study. |
High risk of bias | Data is extracted from NCT ( PP for overall study population = . 306 (netarsudil) and 316 (timolol)). According to the NCT 243 (Netarsudil) and 314 timolol patients completed the study, respectively. Thus, a substantial amount of patients that per protocol definition didn't complete the study have been used in the analysis of the netarsudil‐group. | Low risk of bias | IOP measured with a calibrated Goldmann applanation tonometer (the most clinically accurate and the standard tonometer used in the diagnosis and treatment of glaucoma [3]). |
Low risk of bias | A priori analysis was described and followed. IOP is reported as a number, which is the only way to report that outcome. Detailed description of measurements e.g. timeslots and number of measurements. |
High risk of bias | Based on the judgment of high risk of bias in domains 2 and 3, the overall risk of bias is judged as high. |
| Serle 2018 (ROCKET‐1) | Low risk of bias | Eligible patients were randomized by a computer‐generated method (...). No information on allocation concelament. Significant differences in "iris colour" between intervention groups (P=0,0085). However, no clinical influence. | High risk of bias | Double‐masked: Patients and study site personnel were fully masked to treatment assignments. A vehicle bottle was provided for AM dosing in the netarsudil to maintain masking. In ROCKET‐1, the primary efficacy population was the per‐protocol population with maximum baseline IOP < 27 mm Hg. PP is defined as subjects without major protocol violations likely to seriously affect the primary outcome of the study. Netarsudil‐group: 182/202 analyzed (netarsudil 171 completed the study according to NCT), timolol‐group: 188/209 (timolol 196 completed the study according to NCT). Thus, per study definition several non‐completing subjects have been added to the analysis in the netarsudil‐group. |
High risk of bias | In ROCKET‐1, the primary efficacy population was the per‐protocol population with maximum baseline IOP < 27 mm Hg. PP is defined as subjects without major protocol violations likely to seriously affect the primary outcome of the study. Netarsudil‐group: 182/202 analyzed (netarsudil 171 completed the study according to NCT), timolol‐group: 188/209 (timolol 196 completed the study according to NCT). Thus, per study definition several non‐completing subjects have been added to the analysis in the netarsudil‐group. | Low risk of bias | IOP was measured using a calibrated Goldmann applanation tonometer. Two consecutive IOP measurements of each eye were obtained. If the 2 measurements differed by more than 2 mm Hg, a third measurement was to be obtained. IOP was to be recorded as the mean of 2 measurements or as the median of 3 measurements Double‐masked: Patients and study site personnel were fully masked to treatment assignments. | Low risk of bias | Statistical plan are published and followed. Primary efficacy endpoint (IOP) reportet as a number ‐ only way to report that outcome. | High risk of bias | Judged as high risk of bias in Domain 2 and 3. Thus, the overall risk of bias was judged as high. |