Risk of bias assessment
The two authors (Y.O. and F.M.) independently assessed the risk of bias of the included studies for each outcome using the Revised Cochrane risk of bias tool for RCT (RoB 2.0) 9 and the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool for non-randomized studies 10. The following five domains of the risk of bias were assessed for RCTs: randomization process, deviations from the intended intervention, missing outcome data, measurement of the outcome, and selection of the reported result. Conflicts of assessment between the two reviewers were resolved through discussion, with the third reviewer (T.I.) adjudicated it, if needed.
Assessment of certainty of evidence and summary tables The Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) approach was used to assess the certainty of evidence for each outcome across the studies. Certainty of evidence was categorized into four levels (high, moderate, low, or very low) and summarized in an evidence profile table. 11,12 As recommended by the GRADE approach, 11,13 certainty of evidence was preliminarily started as “high” for not only RCTs but also non-randomized studies because the risk of bias in non-randomized studies was assessed according to ROBINS-I on the same metrics as those of RCTs. 10 The certainty of evidence was downgraded for risk of bias, imprecision, inconsistency, indirectness, and publication bias. For assessing imprecision, the sample size required to detect a 20% risk ratio [RR] reduction, called “optimal information sizes,” was calculated using total event rates in the control groups of the included studies. 11,13