When to Transfuse?

How often do you all hear one of us ask, “Well, at what Hg do you want to transfuse?”. It always becomes a game of “Guess What I am Thinking”. It often feels so arbitrary, that we could be playing the game of how many fingers I am holding up behind my back. What is the correct answer to this question? Is there a correct answer, or is it all ‘gut instinct’? Lets review some evidence.
Lets start off with the best evidence we have. In 1999, a large multi-centered trial was published called the TRICC Trial (Transfusion Requirements In Critical Care). The study enrolled nearly 900 ICU patients and assigned them to either a liberal transfusion cut-off (hg of 10) or a restrictive cut-off (hg of 7). The goal Hg for the liberal group was to maintain them around 10-12, while the restrictive was 7-9. The study excluded younger patients (<16: I will not discuss this, but there is also a study on peds that found the same results, came out in 2007 called TRIPICU), patients with an active bleed defined as a recent drop of 3 grams in Hg over a 12 hour period, pregnancy, brain dead, or refusal of blood products. They also excluded patients with chronic anemia under 9. So, what did they find?
I think to the surprise of many, there was not a difference in end points (30 and 60 day death, ICU stay) between the groups. There was a trend of decreased mortality in the restrictive group, reaching statistical significance in patients <55y/o with APACHE II scores of under 20. Looking even deeper into it, patients with cardiac disease did not fare any different between the two groups, and again there was a trend towards better outcomes in the restrictive group. As would be expected, the restrictive group received far fewer blood transfusions than the liberal group. Wow! In the sickest of the sick, a cut-off of 7 did as well or better as a cut-off of 10, even in patients with underlying cardiac disease! Stunning.
This is by far the best evidence there is. So, why don’t we always use 7? I think it falls into a few areas:
1. Fear: We all start getting anxious when see that number under 8.0. Does it make sense? Probably not, but there it is.
2. Other data: This is the best study, and does answer quite a few questions, but it didn’t answer all of the questions. What didn’t it answer? Well, what about those patients with acute bleeds? What about patients with ACS? There are studies on these, but they are all small and for the most part not great studies. As such, they seem to have mixed results. We do know that patients who are having a heart attack do worse if they have hg <10, but would those patients do better if we transfused them? Not clear. What about GI bleed? There is an older study which showed that perhaps a more restrictive approach is superior (remember that the TRICC study excluded acute bleeds defined as Hg drop of 3 in 12 hours, many of our patients with those slow bleeds fall out of that category and were included in the TRICC). Lastly, in patients with ARF, there seems to be worse outcomes in severe anemia (<9.5 or so: too late and too tired to pull the article), but again, does transfusing these patients help? Not clear.
Where do we go next? 
A study is underway now to answer many of the above questions (FOCUS). It is a large trial, and specifically is tackling the whole cardiac question. But, I think we have more than enough evidence to stop some practices. We should be using 7.0 as our cut-off, unless the patient falls in one of the following categories: Active ACS or Active GI bleed. Transfusing above 7 should be the exception, not the rule! 
In those situations that fall outside of TRICC, well, we will need to again use our ‘gut’, which by the way has been wrong this entire time!