Davenport 2007.

Methods	Double‐blind trial comparing glucocorticosteroid administration versus placebo, in two UK centres
Participants	73 infants from 2 UK centres 34 male, 39 female infants were included. Age in days at surgery, median (interquartile range) Treatment: 60 (50 to 71) Placebo: 54 (45 to 70) Preoperative bilirubin (μmol/L), median (interquartile range (IQR)) Treatment: 132 (112 to 166) Placebo: 158 (125 to 183) Preoperative AST (IU/L), median (IQR) Treatment: 163 (118 to 202) Placebo: 54 (99 to 220) Preoperative ΓGT (IU/L), median (IQR) Treatment: 420 (275 to 898) Placebo: 54 (304 to 992)
Interventions	Participants received either oral prednisolone 2 mg/kg/day on days 7 to 21 following Kasai portoenterostomy, then 1 mg/kg/day on days 22 to 28 following Kasai portoenterostomy (n = 34) or placebo (n = 37).
Outcomes	Primary outcomes: clearance of jaundice (defined as the achievement of a plasma total bilirubin level of ≤ 20 μmol/L at 6 and 12 months); the proportion of infants surviving with their native liver at 6 and 12 months. Secondary outcomes: biochemical liver function tests 1, 6, and 12 months after the operation: total bilirubin, aspartate aminotransferase (AST), γ‐glutamyl‐transpeptidase (ΓGT), alkaline phosphatase (ALP), and albumin.
Notes	Two‐year native liver survival was reported in a graph and the exact values were confirmed with the author. Subgroup of infants who were operated on by day 70 of life was reported.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participating infants were randomized to receive either oral prednisolone or placebo" Comment: sequence generation was performed by a centralised agency within the pharmacy and was performed independently of the study investigators.
Allocation concealment (selection bias)	Low risk	Comment: medications or placebo were prepared and concealed but the respective pharmacies of the hospitals included in the study. Investigators were unable to identify if glucocorticosteroid or placebo was being given to any particular patient.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Double blinded" Comment: methodology was clarified upon contacting author and investigators and clinical personnel were blinded as the administered medication was not identifiable as placebo or glucocorticosteroids by clinical staff or investigators.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: the authors confirmed that outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no significant incomplete reporting was identified. Two infants excluded from the trial due to the finding that they had factors which excluded them from eligibility after already being enrolled.
Selective reporting (reporting bias)	Low risk	Comment: all predefined outcomes from the protocol were reported on.
Other bias	Low risk	None identified

AST: aspartate aminotransferase
 IU/L: international units per litre
 ΓGT: γ‐glutamyl‐transpeptidase