Two of the five review authors extracted data independently and in
duplicate, using a customized data extraction form developed in
Microsoft Excel (Microsoft Excel). Disagreements during the review
process were resolved through discussion between the two review authors.
In cases where consensus could not be reached, a third review author was
consulted to help resolve the disagreement and make a final decision.
The Risk of Bias in Non-Randomized Interventions (ROBINS-I) tool was
used to assess the quality of non-randomized studies (table 2). The Risk
of Bias (RoB2) tool was also used to assess the quality of randomized
studies (table 3). We used the same data extraction form to extract the
data. We extracted data including (1) study characteristics (author,
year, place, and study design); (2) patient’s characteristics (sample
size, and sex); (3) intervention and comparison (sample size and
treatment dose); and (4) safety outcomes. All steps stated above were
performed independently by two authors.