ObjectiveLack of clear evidence in red blood cell (RBC) transfusion during gastrointestinal bleeding has led to varied recommendations over the years. However, studies in broad areas of medicine have provided evidence about appropriate RBC transfusion thresholds, and a ‘landmark’ study published in 2013 provided evidence in patients with upper gastrointestinal (UGI) bleeding. We hypothesized that the response to the evidence would lead to improved RBC transfusion practice. Our aim was to determine the response in RBC transfusion practices at our institution.DesignWe examined RBC transfusion practices in patients with UGI bleeding who presented to the Medical University of South Carolina from January 2010 through December 2013. We abstracted extensive clinical data including demographic, medical history (comorbidities), medications, physical examination findings, laboratory data, endoscopic data, and RBC transfusion practices. We considered appropriate RBC transfusion to have occurred when performed for a hemoglobin level of <70 g/L.Results270 patients hospitalized with UGI bleeding had 606 RBC transfusions; 355 transfusions in 107 patients were appropriate, and 251 transfusions in 163 were inappropriate. In 2010, 2011, and 2012, the rates of appropriate RBC transfusions were 61/124 (49%), 92/172 (53%), and 84/142 (59%), respectively. There was a statistically significant difference in appropriate transfusions in 2013 (118/168 (70%)) compared with 2012 (84/142 (59%), p=0.003), as well as during 2010–2012 (237/438 (54%), p≤0.003).ConclusionsThe data suggest that there was an improvement in RBC transfusion practices after a landmark study. However, the data also highlight that RBC transfusion practices in UGI bleeding remain imperfect.