The paper to be discussed compared a liberal transfusion trigger (of 9g/dl) to a restrictive strategy of 7g/dl in patients with UGI bleeding.
Less than half of the cirrhotics were alcohol induced - whether that relates to the frequency of bleeding in non-alcoholic cirrhosis I don't know, but that's not the Sunderland experience.
FFP and platelet use were similar (low) in both groups - now that we have early use of FFP and platelets established, certainly in trauma, I wonder if we will eventually be putting more emphasis on correcting the coagulopathy of haemorrhage than replacing red cells - not of relevance to this study but a thought I had none the less!
There was a 10% violation rate in the restrictive group. Even in the context of a study, physicians are clearly uncomfortable with 7 (it was always 10 when I did medicine!) I've also heard it on more than one occasion that the justification for a higher Hb is in case of a rebleed - I do love the irony that the liberal group had more rebleeding!
The crystalloid infusion was the same in both groups over 72hrs surprisingly. I wonder therefore if the results could be explained by the degree of relative hypervolaemia in the liberal group - the fact they were red cells being less relevant. This would fit with splanchnic expansion, increased portal pressure etc.
Those are a few of my thoughts, I look forward to reading yours....