“…Risk of Bias was low in 4 (14%) studies, 41 , 45 , 49 , 50 high in 18 (64%) studies 29 , 30 , 31 , 33 , 34 , 35 , 36 , 38 , 39 , 44 , 46 , 47 , 48 , 51 , 52 , 53 , 54 , 55 and showed some concerns in 6 (21%) studies. 28 , 32 , 37 , 40 , 42 , 43 For one high risk study, only an abstract was available, 35 so downgrading this study based on limited information may be unreasonable. Bias arose mostly from inadequate reporting of results: 5 (19%) studies failed to report on all inflammatory outcomes described in protocols or trial registrations 29 , 39 , 44 , 46 , 48 and 9 (33%) were not prospectively registered.…”