February 11, 1999
A MULTICENTER, RANDOMIZED, CONTROLLED CLINICAL TRIAL OF TRANSFUSION REQUIREMENTS IN CRITICAL CARE.
What was the research question?
Does restrictive blood transfusion improve mortality compared to liberal blood transfusion strategy in patients who were admitted to the ICU?
How did they do it?
Randomized, non-blinded, control trial in 25 ICUs in Canada.
6541 adult patients were assessed of whom 838 randomized to receive restrictive strategy targeting Hb level of 7-9 g/dl, or liberal strategy with Hg target of 10-12 g/dl.
What did they find?
Primary outcome of 30-day mortality was similar in restrictive strategy compared to liberal strategy (18.7% vs. 23.3%; P=0.11). However, in sub-group analysis, mortality was significantly lower in less acutely ill patients with APACHE II score ≤ 20 (8.7% vs. 16.1%, p=0.03), and in patients younger than 55 years (5.7% vs. 13%, P=0.02). In patients with cardiac disease, the morality rate was similar (20.5% vs. 22.9%, p=0.69).
Cardiac events (MI, pulmonary oedema, angina, cardiac arrest) were less common in restrictive strategy compared to liberal strategy (13.3% vs. 21%; P<0.01).
Secondary outcome of hospital mortality was lower in the restrictive group compared to the liberal group (22.2% vs 28.1%; P=0.05).
What are the limitations of the study?
Possible selection bias as only 13% patients were included in the trial.
Subgroup APACHE score was changed from <15 to <20 post hoc, likely to get the significant results.
Possibility of type II error as the sample was only 838 compared to 1620 needed for the power of the study.
No subgroup analysis for patients with traumatic brain.
What does it mean?
A restrictive strategy of red-cell transfusion is at least as effective as and possibly superior to a liberal transfusion strategy in critically ill patients, with the possible exception of patients with acute myocardial infarction and unstable angina.
The transfusion threshold in critically ill patients is changed based on this study from <9 g/dL to <7 g/dL except in patients with evidence of active cardiac ischemia.