Quality assessment

The quality of all selected studies was assessed using an adaptation of the Downs and Black criteria as described in previous systematic reviews [41–43]. From 27 original items in the checklist of the Downs and Black criteria, 17 were employed to accommodate the characteristics of observational studies, while other items specific for interventional randomization studies were removed. As recommended by Wehrmeister and colleagues [44], the total scores range from 0 to 18 points, given that each item scores one point, except for item 4 that can result in 0 (no), 1 (partially) and 2 (yes). Studies could be categorized with a quality score as: high chance of bias (0–5 points), moderate chance of bias (6–11 points) and low chance of bias (12–18 points) [41]. Two reviewers rated each study independently according to the above quality criteria, and discrepancies were discussed and resolved by consensus between referees.