Murphy 2015.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, multicentre, open‐label, non‐inferiority trial Setting: 17 centres (UK) Recruitment: July 2009 to February 2013 Maximum follow‐up: 90 days
Participants	Participants > 16 years of age who were undergoing non‐emergency cardiac surgery with Hb < 9 g/dL 2007 randomised; 4 withdrew Liberal group: n = 1003; median age (IQR) = 70.8 (64.1 to 76.7) years Restrictive group: n = 1000; median age (IQR) = 69.9 (63.1 to 76.0) years
Interventions	Liberal group: received transfusion when Hb was < 9 g/dL Restrictive group: received transfusion when Hb was < 7.5 g/dL
Outcomes	Primary outcome: a composite of serious infection (sepsis or wound infection) or an ischaemic event (permanent stroke, myocardial infarction, infarction of the gut, or acute kidney injury) within 3 months after randomisation Secondary outcomes included: units transfused, infection, ischaemic events, acute kidney injury, hospital stay and ICU stay, general quality of life (using the EQ‐5D), cost
Notes	Trial title: TITRe2 Trial registration (prospective): ISRCTN70923932 Trial funding: "supported by the NIHR Health Technology Assessment program ... Dr. Reeves and the research nurse team in Bristol were supported in part by the NIHR Bristol Biomedical Research Unit in Cardiovascular Disease, and Drs. Murphy, Angelini, and Rogers were supported by the British Heart Foundation" (Murphy 2015 p 1007) COI statement by investigators: "no potential conflict of interest relevant to this article was reported" (Murphy 2015 p 1007)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned to either the liberal transfusion‐threshold group (threshold hemoglobin level, 9 g per deciliter) or the restrictive transfusion‐threshold group (threshold hemoglobin level, 7.5 g per deciliter) by means of a secure Internet‐based system that concealed assignments and used cohort minimization to balance assignments according to center and type of surgery" (Murphy 2015 p 998)
Allocation concealment (selection bias)	Low risk	Trial used an Internet‐based system that concealed assignments and used cohort minimisation to balance assignments according to centre and type of surgery (Murphy 2015 p 998)
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' Physicians and nurses were aware of group assignments
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 graded risk of bias as 'low' Outcomes were adjudicated
Blinding of outcome assessment (detection bias) Subjective measures	Low risk	For the subjective measure of quality of life (EQ‐5D), we rated risk of bias to be low, as participants were unaware of group assignment. Furthermore, investigators reported that they "tested our success in keeping the study groups blinded by asking the patients if they were aware of the group they were in" "At discharge, 15.1% of patients believed they knew treatment and 75.6% were correct" (Murphy 2015 p 1002); however, "3 months after surgery, a greater number of patients (27.5%) thought that they knew which treatment they had received, but fewer (56.6%) were correct"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data were missing for the primary outcome analysis (composite as above) for 4.8% of the sample; overall data including for mortality were missing for only 1.2% at 3 months
Selective reporting (reporting bias)	Low risk	Trial was prospectively registered, and all prespecified outcomes appear to have been reported in full
Other bias	Low risk	No other biases were apparent