Clinical Question

Does a lower hemoglobin threshold for red cell transfusion in patients with septic shock affect 90-day mortality compared to a higher hemoglobin threshold?

Bottom Line

In patients with septic shock, using a lower hemoglobin threshold (≤7 g per deciliter) for red cell transfusion compared with a higher threshold (≤9 g per deciliter) resulted in fewer transfusions without a significant difference in 90-day mortality, ischemic events, use of life support, or days alive out of the hospital.

Major Points

Guidelines

The current guidelines for transfusion in septic shock patients recommend maintaining a hemoglobin concentration above 7 g/dL unless the patient exhibits symptoms such as myocardial ischemia, severe hypoxemia, or ongoing hemorrhage.

Design

Multicenter, stratified, parallel-group, randomized, partially blinded clinical trial
N=998 adult patients with septic shock

Interventions:

- Lower-threshold group (≤7 g per deciliter) for transfusion (n=502)
- Higher-threshold group (≤9 g per deciliter) for transfusion (n=496)
Primary outcome: Death at 90 days post-randomization
Secondary outcomes: Ischemic events, life support use, adverse reactions to transfusion, percentage of days alive without life support, and days alive out of the hospital within 90 days
Enrollment: December 2011 to December 2013

Population

Inclusion Criteria:
- Adult patients (≥18 years) in the ICU
- Septic shock diagnosis according to criteria
- Hemoglobin concentration ≤9 g/dL
Exclusion Criteria:
- Significant bleeding
- Requirement for ECMO
- Not specified acute myocardial infarction
Baseline Characteristics:
- Similar between both groups

Interventions

- Red cell transfusions at hemoglobin ≤7 g/dL for the lower-threshold group
- Transfusions at hemoglobin ≤9 g/dL for the higher-threshold group

Outcomes

Primary Outcomes:
- At 90 days, mortality was 43.0% in the lower-threshold group versus 45.0% in the higher-threshold group (relative risk, 0.94; P=0.44)
Secondary Outcomes:
- Similar use of life support at days 5, 14, and 28
- Comparable numbers of ischemic events and severe adverse reactions
- No significant difference in the percentages of days alive without life support or out of the hospital

Criticisms

- Lack of blinding for investigators, clinicians, and patients regarding transfusion thresholds could introduce bias.
- Limited power to detect significant differences in secondary outcomes and certain subgroup analyses.
- The choice to use leukoreduced blood, potentially limiting the applicability concerning nonleukoreduced blood.
- Trial did not specify surveillance for myocardial ischemia, possibly missing some events and resulting in detection bias.

Funding

Supported by the Danish Strategic Research Council, Copenhagen University Hospital, Rigshospitalet, the Scandinavian Society of Anaesthesiology and Intensive Care Medicine, ACTA Foundation, and others.

Further Reading