Sanne de Bruin

12 Chapter 1 Transfusion in critically ill patients A substantial number of critically ill patients develops anaemia and/ or coagulopathy on the ICU 15–17 , which are independently associated with increased mortality and mor- bidity 16,18–20 . At themoment of admittance about two third of the patients have a Hb level < 12 g/dL 21 and about a quarter have a Hb level < 10 g/dL 15,22 . Furthermore, incidences of thrombocytopenia (platelet count <150 x10 9 cells/L) of 40% are reported 19 and a prolonged PT is reported in 25% in critically ill patients 17 . During the last decades a substantial decrease in RBC transfusion is observed. In 2008 approximately 25% of all patients received one or more red blood cell transfusions during ICU stay while a decade earlier about 40%of all ICU patients were transfused 21,22 . This shift towards a more restrictive transfusion strategy is the result of multiple large RCTs, which are performed during the last two decades, in which restrictive (transfu- sion at Hb levels < 7 g/dL) and liberal (transfusion at Hb levels < 9-10 g/dL) transfusion strategies were compared. In 1999 the first trial showed the safety of a restrictive trans- fusion strategy in critically ill patients and a beneficial effect in the patients who are less acutely ill 23 . This result was repeatedly confirmed in other large RCTs in and outside the ICU 8,13,24–26 . However, uncertainty still exists for different subpopulations, especially for patients with acute coronary syndrome, since the oxygen delivery from the coronary arteries is already compromised. In addition, patients with brain injury which might be more sensitive to anaemia induced cerebral hypoxia, and therefore it may be dangerous for this patient group as well to apply a more restrictive transfusion regime. Even though the number of RBC transfusions on the ICU are decreasing, still a large proportion of patients is receiving RBC transfusions at Hb levels higher than 7 g/dL, with large differences between world regions 15 . It is unclear why physicians are tempt- ed to transfuse patients at Hb levels higher than 7 g/dl. One of the explanations is the ongoing debate on transfusion triggers as they are not a read out of impaired oxygen delivery. In line with the definition of anaemia, which is defined by decreased Hb levels, treatment for anaemia is in guidelines also based on Hb levels 27,28 . Nevertheless, the aim of blood transfusion is not increasing a Hb value, but ensuring a sufficient oxygen supply. Therefore, several other triggers have been mentioned in literature which are based on global or regional oxygenation status. These triggers include lactate levels 29 , central venous oxygen saturation 29 and regional tissue oxygenation 30,31 . These triggers might reflect the impaired oxygen delivery better than Hb levels. One small RCT com- pared haematocrit (Ht) levels with a personalized algorithmbased on brain oxygenation during cardiac surgery. The use of this algorithm did not result in decreased number of