3.2. Risk of bias assessment
The details for the risk of bias assessment are displayed inAppendix2 . The random sequence generation was specified in 24 trials (72.7%). 18 trials (54.5%) reported the allocation concealment but one trial has high risk of bias for it. Only one trial lack of blinding methods. Selective reporting was identified in 8 trials (24.2%). No trials were found in high risk of bias for incomplete outcome data addressed and other bias. Overall, the quality of the included studies was considered relatively high.