Certainty of evidence
Risk of bias assessment as performed by study authors revealed 20
studies at high risk of bias, 10 at unclear risk of bias and 20 at low
risk of bias. Most of the studies lacked placebo groups thus were at
high risk of performance bias furthermore allocation to treatment groups
was not blinded in many studies. None of the studies were at risk of
attrition or reporting bias. Supplementary figure 3 depicts bar charts
showing contribution of each study to the network estimate. Overall
there were 27 mixed evidence and 78 indirect evidences and quality of
evidence was ranked from very low to high and has been provided in
supplementary table for primary outcome i.e. SER. Funnel plots showed
mild asymmetry for the outcome SER however on regression analysis it was
not significant furthermore funnel plot for SET was symmetrical
(Supplementary figure 5 and 6)