Transfusion Strategies in UGIB

Villanueva C, et al.  Transfusion Strategies for Acute Upper Gastrointestinal Bleeding.  2013 Jan 3;368(1):11-21.
Recently, I presented the landmark article for promoting a restrictive transfusion strategy in upper GI bleeding (UGIB) at our journal club.  I spent a fair amount of time researching the background to this study.  Suffice it to say, there's been a lot of nerds manufacturing non-evidenced based guidelines for decades regarding this issue.  They usually met in Italy, so I can understand the draw to convene.  This study aimed to set the record straight on transfusion strategy for undifferentiated UGIB.


This was a single center non-blinded RCT from 2003-2009 conducted in Barcelona on variceal and non-variceal UGIB.  Patients were enrolled in one of two arms after admission: (1) restrictive transfusion with a  hemoglobin transfusion threshold of 7 g/dL, or (2) liberal transfusion with a transfusion threshold of 9 g/dL.  Patents with massive exsanguination (undefined in the study), and minor bleeding with hemoglobin > 12 were excluded (among other exclusion criteria).  Primary outcome was mortality within the first 45 days.  The study enrolled 961 patients, about half of which were patients with ulcers, a quarter were variceal, and the rest were Mallory-Weiss tears, gastritis, neoplasms, etc...  Mortality was 5% in the restrictive strategy, and 9% in the liberal one.  The hazard ratio was 0.55 (95% CI 0.33 - 0.92).  In a subgroup analysis, there was a mortality benefit for Child-Pugh A/B patients (4% vs 12%).  There was no mortality benefit for Child-Pugh C patients (interpretation: they're going to die anyway).  As far as secondary outcomes: less bleeding in restrictive group, less rescue therapies (i.e. balloon tamponade, TIPS), less length of stay, and less adverse events.

Those are really powerful results!  Now, this is not a perfect study - it's single center, patients got endoscopy within 6 hours (are you kidding me?!), massive exsanguination was not defined, the study was performed after admission (not in the ED) - but it is the only RCT that exists on this issue.  So, in conclusion, restrictive strategy is now the standard (for non-massively exsanguinating patients).

Let's add another weapon to our arsenal of therapies that improve mortality in UGIB.  PPIs do not improve mortality, it only reduce re-bleeding after endoscopy.  Octreotide does not improve mortality, only reduces transfusion needs.  Antibiotics do improved mortality (NNT 22)*!  And now, restrictive transfusion does, too (NNT 25).
*theNNT.com