Hébert 1999.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, multicentre trial Setting: 22 tertiary‐level and 3 community ICUs in Canada Recruitment: November 1994 to November 1997 Maximum follow‐up: 30 days
Participants	838 critically ill participants with euvolaemia after initial treatment and with Hb < 9.0 g/dL within 72 hours after admission to the ICU were randomly assigned to 1 of 2 groups: Liberal group: n = 420; M/F = 255/165; mean (SD) age = 58.1 (18.3) years Restrictive group: n = 418; M/F = 269/149; mean (SD) age = 57.1 (18.1) years
Interventions	Liberal group: transfused when Hb fell to < 10.0 g/dL. Hb maintained between 10.0 g/dL and 12.0 g/dL Restrictive group: transfused if Hb dropped to < 7.0 g/dL. Hb was maintained between 7.0 g/dL and 9.0 g/dL
Outcomes	Mortality, length of hospital stay, length of ICU stay, blood usage (units), complications, infection rates, cardiac events, pulmonary oedema, pneumonia
Notes	Trial registration: none ascertainable Trial funding: supported by the Medical Research Council of Canada and by an unrestricted grant from Bayer COI statement by investigators: published paper included no information
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The random order was computer generated (Hébert 1999 p 410)
Allocation concealment (selection bias)	Low risk	"The data coordinating centre prepared sealed opaque envelopes, which they distributed to each participating institution where they were opened up sequentially to determine the participants treatment assignment. The envelopes were returned periodically to the coordinating centre for auditing" (Hébert 1999 p 410)
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 graded risk of bias as 'low' "It was not feasible to mask the assigned transfusion strategy from health care providers" Participants were ICU patients. Most outcomes were based on laboratory measures
Blinding of outcome assessment (detection bias) Subjective measures	Low risk	No data from subjective outcomes (e.g. function)
Incomplete outcome data (attrition bias) All outcomes	Low risk	Of 838 participants randomised, there were 4 withdrawals from the liberal transfusion arm and 5 from the restrictive arm
Selective reporting (reporting bias)	Unclear risk	Evidence of prospective registration/trial protocol unavailable; insufficient information available to make a judgement
Other bias	Low risk	No other biases were apparent