Red blood cell transfusion strategies in critically ill patients: lessons from recent randomized clinical studies.

A randomized, multicenter trial conducted in 32 northern European general intensive care units (ICUs) enrolled some patients with septic shock randomly assigned to receive a red blood cell transfusion when the hemoglobin (Hb) level was ≤7 g/dL ("lower threshold"; N.=502) or ≤9 g/dL ("higher threshold"; N.=496) throughout the ICU stay. Patients were excluded if they had an acute coronary syndrome, life-threatening bleeding, acute burn injury, had already been transfused or had previously experienced transfusion-related reactions. The two groups of patients had comparable severity of disease scores and chronic cardiovascular conditions. Median Hb values were 7.7 g/dL in the lower and 9.3 g/dL in the higher threshold groups and these values remained stable during the study period. There was no significant difference in 90-day mortality (primary end-point) between the two groups (216/502, 43.0% in the lower vs. 223/496, 45.0% in the higher group, RR 0.94 [95% CI: 0.78-1.09; P=0.44]), even after adjustment for several confounders. In the higher threshold group, approximately twice as many transfusions were given (3088 vs. 1545 units transfused, P<0.001) as in the lower threshold group. In the lower threshold group, more patients received no RBC transfusion (36% vs. 1.2%, P<0.001) than in the higher threshold group, but there were also more temporary protocol suspensions (5.9 % vs. 2.2%, P=0.004), in particular because of myocardial ischemia (6/488, 1.2% vs. 0/489), life-threatening bleeding (18/488, 3.7% vs. 9/489, 1.8%) and need for higher Hb levels during extracorporeal membrane oxygenation therapy. We discuss how anemia should be managed in patients with sepsis or other critical illness, especially in the context of the potential risks associated with RBC transfusion and data from other recent large randomized trials.