Holst 2014.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, multicentre trial Setting: 32 general ICUs in Denmark, Sweden, Norway, and Finland Recruitment: December 2011 to December 2013 Maximum follow‐up: 12 months
Participants	1005 participants with septic shock and Hb < 9 g/dL were randomised Liberal group: n = 496; age (interquartile range) = 67 (58 to 75) years Restrictive group: n = 502; age (interquartile range) = 67 (57 to 73) years
Interventions	Liberal group: transfused with single units of cross‐matched, prestorage leuco‐reduced RBCs when Hb decreased to ≤ 9 g/dL Restrictive group: transfused with single units of cross‐matched, prestorage leuco‐reduced RBCs when Hb decreased to ≤ 7 g/dL The intervention period was the entire ICU stay, to a maximum of 90 days after randomisation
Outcomes	Primary outcome: 90‐day mortality
Notes	Trial registration: NCT01485315 Trial funding: supported by a grant (09‐066938) from the Danish Strategic Research Council and by Copenhagen University Hospital, Rigshospitalet, the Scandinavian Society of Anaesthesiology and Intensive Care Medicine (ACTA Foundation), and Ehrenreich’s Foundation COI statement by investigators: "Dr. Johansson reports receiving grant support from Pharmacosmos; and Dr. Perner, receiving grant support from CSL Behring, Fresenius Kabi, Cosmed, and Bioporto Diagnostics, and lecture fees from LFB. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org" (Holst 2014 p 10)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"A centralised computer generated the assignment sequence. Randomization was performed with the use of a centralized computer generated assignment sequence, with stratification according to study site and the presence or absence of active hematologic cancer ... Patients with septic shock were randomly assigned in a 1:1 ratio, with the use of permuted blocks of varying sizes of 6, 8, or 10, to blood transfusion at the higher haemoglobin threshold or the lower haemoglobin threshold" (Holst 2014 p 10)
Allocation concealment (selection bias)	Low risk	Use of a centralised computer ensured allocation concealment. See 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 graded risk of bias as 'low' "Treatment assignments were concealed from the investigators assessing mortality, the data and safety monitoring committee, and the trial statistician" (Holst 2014 p 10)
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	99.8% (998) included in all analyses of mortality; (977) 97.7% included in analyses of all outcomes
Selective reporting (reporting bias)	Low risk	Trial was prospectively registered (NCT01485315). Reporting was comprehensive for 90‐day outcomes in Holst 2014; longer‐term data appear to have been fully reported in Rygård 2016
Other bias	Low risk	No other sources of bias were apparent