Bracey 1999.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, single‐site trial Setting: non‐profit research/treatment centre, Houston, TX, USA Recruitment: February to November 1997 Maximum follow‐up: in‐hospital stay
Participants	428 consecutive participants undergoing elective primary coronary artery bypass graft surgery, randomly assigned to 1 of 2 groups: Liberal group: n = 212; M/F 82/18; mean (SD) age = 61 (11) years Restrictive group: n = 216; M/F 83/17; mean (SD) age = 62 (11) years
Interventions	Liberal group: received transfusions on the instructions of their individual physicians, who considered clinical assessment of the participant and institutional guidelines, which proposed a Hb < 9.0 g/dL as the postoperative threshold for RBC transfusion Restrictive group: received an RBC transfusion in the postoperative period at Hb < 8.0 g/dL
Outcomes	Mortality, length of hospital stay, blood usage (units), blood loss, complications, infection rates, cardiac events
Notes	Trial registration: not confirmed Trial funding: not reported COI statement by investigators: none provided
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants "were randomly assigned on the basis of the last digit of their medical record number" (Bracey 1999 p 1071) Review authors' judgement: this is not a genuinely random method of allocation. High risk of bias
Allocation concealment (selection bias)	High risk	Sequence generation (record number) obviated concealment of any kind
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 grade risk of bias as 'low'
Blinding of outcome assessment (detection bias) Subjective measures	High risk	Investigators provided no information regarding the survey questionnaire "When analyzed by Hb content, FACT‐F and FACT‐An scores were significantly different on postoperative Day 5, when patients with a Hb <9 g per dL were compared with patients with a Hb level >10 g per dL (p = 0.004). However, the same groups had similar survey scores on postoperative Day 3. Because the goal of our study was to maintain Hb content at >8 g per dL, we reanalyzed the subgroup and excluded patients with lowest Hb <8 g per dL. Then, survey scores did not differ significantly between low and high Hb groups...." (Bracey 1999 p 1075)
Incomplete outcome data (attrition bias) All outcomes	Low risk	The trial used intention‐to‐treat analysis. Mortality data were reported for all participants, but for analysis of transfusion, participants who died during hospitalisation were excluded from analysis, as "early death precludes observation of transfusion and morbidity rate" (Bracey 1999 p 1075)
Selective reporting (reporting bias)	Unclear risk	No reporting bias was apparent, but in the absence of prospective registration or a trial protocol, assessment must remain 'unclear'
Other bias	Low risk	None identified