Akyildiz 2018.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, single‐site, non‐inferiority, 'per‐protocol' approach Setting: PICU, tertiary (university‐affiliated hospital), Kayseri, Turkey Recruitment: January 2014 to December 2015 Maximum follow‐up: in‐hospital stay
Participants	180 children in paediatric ICU randomised; 160 completed follow‐up Liberal group: n = 89; M/F 41/48; median (IQR) age = 36 (24 to 48) months Restrictive group: n = 71; M/F 36/35; median (IQR) age = 72 (36 to 84) months
Interventions	Liberal group: Hb < 10 g/dL trigger Restrictive group: Hb < 7 g/dL trigger
Outcomes	Primary outcomes: cardiac output and other haemodynamic measures Secondary outcome: in‐hospital mortality
Notes	Trial registration: not confirmed. Internal (university) documentation: registration TSA‐2014‐5299 Trial funding/Sponsor: Erciyes University Scientific Research Unit Conflict of interest: study authors declare they have none
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"Randomization was provided by a coin toss" (Akyildiz 2018 p 2)
Allocation concealment (selection bias)	High risk	No information was provided beyond the statement above
Blinding of participants and personnel (performance bias) Objective outcomes	Low risk	Blinding of personnel for this intervention is not feasible, but in our view, for objective outcomes such as mortality (the primary outcome used within this review), we graded risk of bias as 'low'
Blinding of outcome assessment (detection bias) Objective measures	Low risk	Blinding of mortality (the primary outcome used within this review) is not relevant, and we will grade risk of bias as 'low'
Blinding of outcome assessment (detection bias) Subjective measures	Low risk	"Cardiologists who assessed the cardiac output measurements were blinded to treatment allocations" (Akyildiz 2018 p 2) No other subjective outcomes were used in this review
Incomplete outcome data (attrition bias) All outcomes	High risk	There was a differential loss to follow‐up, with more losses in the restrictive arm (1 loss in the liberal arm and 19 in the other)
Selective reporting (reporting bias)	Unclear risk	No evidence of prospective registration; trial protocol unavailable; information insufficient to make a judgement
Other bias	Low risk	None apparent