“…Overall, 94% (17/18) of the studies had a high or unclear risk of bias in at least one of the seven bias domains [ 18 , 23 , 26 , 30 , 46 – 58 ]. Major sources of bias were a lack of allocation concealment (13/18 (72%)) [ 18 , 23 , 26 , 30 , 46 , 47 , 49 , 51 – 54 , 56 , 57 ], no blinding of participants and personnel (12/18 (67%)) [ 18 , 23 , 26 , 30 , 46 – 50 , 56 – 58 ], and incomplete outcome data (12/18 (67%)) [ 18 , 23 , 26 , 30 , 46 , 48 , 49 , 53 – 57 ]. In addition, only 44% (8/18) of studies clearly described the method of how the random sequence was generated [ 24 , 47 , 48 , 50 – 52 , 54 , 55 ].…”