Quality assessment of studies
Two authors (A.S. and E.C.) assessed study quality and risk of bias independently using a standardised Excel form. Randomised trials were analysed using the Cochrane Collaboration’s tool14 for assessing risk of bias. Case-control and cohort studies were analysed using the Newcastle-Ottawa scale15 for assessing the quality of non-randomised studies. An adaptation of the Murad tool16 was used for case series.