“…Detection bias showed similar results with 50% of trials deemed low risk, 17% unclear and 33% high risk. The large number of trials judged to be unclear or high risk for performance and detection bias were commonly due to difficult blinding participants and personnel to interventions due to the differences in administration [11,16,19,21,22,31,32,35,39,41,42,44], unclear reporting of methods [28,29] and a likelihood that blinding could be broken [18,20]. The Deneux-Tharaux 2013 [32] and Groot 1996 [44] trials were judged to be high risk of performance bias but low and unclear risk of detection bias, respectively, due to the use of objective measurements for the primary outcome.…”