Narrowing down the TRICC concept to septic shock patients, the TRISS trial looked at a transfusion trigger of 7 g/dl or 9 g/dl, using leuco-depleted blood, for the first 90 days of admission. Multi-centre, parallel-grouped and sensibly powered (over 1000 patients), they achieved good separation of Hb values.
No mortality difference at 90 days. No greater ischaemic events in the lower group and no increased use of life-support elements. But also no deleterious effects of giving blood were noticed – was this because leuco-depleted blood was used (unlike TRICC)?
Strikingly 36% of the higher threshold group were transfused, 2% of the lower.
Possible issues: not blinded, high baseline mortality rate (45% – cf ARISE, although illness severity measured with SOFA/SAPS v APACHE), difficulty picking up ischaemic events (non-protocolised diagnosis), actively infarcting patients were excluded. In the lower threshold group the patient could be transfused if ischaemic and the doctor felt it was wise but it’s not clear whether this happened.
Aside from the clinical lesson, does the very low need for blood in the ‘lower’ group suggest that there is a concerted physiological resetting of the haemoglobin target by the body itself (7ish), and we meddle with that in our correct-the-numbers tendency?
Subgroup analyses to come no doubt.
