Should red cell transfusion be unbundled from the sepsis resuscitation bundle?

The sepsis resuscitation bundle consists of those interventions recommended in the Surviving Sepsis Campaign guidelines which are applicable to the first six hours of treatment. They include timely administration of appropriate antibiotics and early goal directed therapy. The early goal directed therapy protocol calls for a transfusion target of 30% hematocrit if the SVC oxygen saturation is below 70% after the initial CVP and blood pressure goals are met.


Controversy surrounding EGDT is based on the fact that, although it has been well validated as a bundle of interventions, the individual components have not been well studied separately. A recent paper in the Indian Journal of Critical Care Medicine addresses (or, I should say, purports to address) whether RBC transfusion belongs in the bundle:

Background: The optimal hemoglobin level and transfusion threshold in patients with septic shock treated with an early, goal oriented approach to resuscitation remains unknown. Aims: To assess the impact of packed red blood cell (PRBC) transfusion on clinically relevant outcomes in patients with septic shock treated with early goal directed therapy (EGDT).

The study found a non-statistically significant trend toward increased mortality in the transfusion group, and a statistically significant increase in length of ICU stay, ventilator time and hospital stay in the transfusion group.

Conclusions: In this retrospective cohort study, transfusion of PRBCs was associated with worsened clinical outcomes in patients with septic shock treated with EGDT.

What that language implies is that transfusion as part of EGDT is associated with bad outcomes. That's not what the study showed at all, but it takes a close look at the body of the paper to realize that. The authors were not explicit about it, but what the study actually looked at was the effect of transfusion over the first 24 hours of treatment. But EGDT, as anyone familiar with the concept knows, ends in 6 hours. Long before the landmark Rivers trial which established the value of goal directed hemodynamic resuscitation in the first 6 hours, we knew, from several studies in the 1980s, that later implementation of such resuscitation was not helpful, in fact potentially harmful.


What this study did demonstrate, really all it demonstrated, was something else we've known for a long time---for almost 12 years in fact---which is that a liberal transfusion trigger (HCT 30) in critically ill patients is in general (without respect to time) not beneficial.


So at what time points did the patients in the Indian Journal of Critical Care Medicine study get their transfusions? Again, the authors aren't explicit but from the details of the paper you can glean that most of the transfusion was not in the six hours of EGDT. How can we know that? First of all, according to the EGDT protocol, transfusion does not even begin until after fluid resuscitation titrated to a CVP of 8, followed by pressor administration to a MAP of 65 and subsequent measurement of the SVC O2 saturation. This figure from the paper says it all. Note that in the transfusion group it took, on average, over 9 hours to achieve the CVP goal of 8! That means that for most patients, transfusion was given either outside the 6 hour period that defines EGDT or in violation of the EGDT protocol.


Finally, even though baseline characteristics of the two groups were similar the transfusion group is defined by the EGDT protocol as a group of patients doing poorly during the course of resuscitation.


In short, while this study addresses the general issue of transfusion in critically ill patients, it proves nothing about transfusion as a component of EGDT.