Gastrointestinal bleeding

Less Is More: A Restrictive Transfusion Policy for Acute GI Bleeding

University of South Alabama

Dr Rutecki is professor of medicine at the University of South Alabama College of Medicine in Mobile. He is also a member of the editorial board of CONSULTANT.

Dr Rutecki reports that he has no relevant financial relationships to disclose.

Top Papers Of The Month
Articles You Don't Want To Miss 


What is the optimal hemoglobin threshold for transfusion in patients with acute gastrointestinal bleeding?

Transfusion practices for gastrointestinal (GI) bleeding have come full circle throughout the 20th and early 21st centuries. As soon as transfusions became available early in the 20th century, they were avoided based on fears that transfusions would raise blood pressure and induce rebleeding.1 Then, a 180-degree turn in practice carried through the end of the last century from 1944.1 A threshold hemoglobin level of 10 g/dL for transfusions during GI bleeding episodes replaced earlier caution.1 Most recently, backsliding has occurred, and a 7 g/dL target has emerged.1 There is one problem: there has been a paucity of data to support any of the thresholds. A recent “Top Paper” may remedy the controversy.


Villanueva and colleagues2 compared what they termed a restrictive versus a liberal transfusion strategy for episodes of GI bleeding. A restrictive strategy was defined as transfusion(s) for GI bleeding when the hemoglobin level falls below 7 g/dL. Conversely, a liberal strategy was characterized by transfusion for hemoglobin levels below 9 g/dL. Nine hundred and twenty-one patients with an acute episode of upper GI bleeding were randomized to either the restrictive (n5461) or liberal strategy (n5460). A total of 225 persons representing the restricted hemoglobin targets (51%) did not receive transfusions. In the “liberal” cohort, the corresponding numbers were 65, or 15%. The survival at 6 weeks (95%) in the restricted transfusion group was higher than in the liberal group (91%). Framed another way, the hazard ratio for death with a restrictive threshold for transfusion was 0.55 (95% CI, 0.33 to 0.92; P5.02). There were additional benefits realized with a restrictive transfusion approach. Further bleeding occurred in 10% of persons with a hemoglobin target of 7 g/dL versus 16% for a 9 g/dL target (P5.01). Adverse events were lower as well at 40% compared to 48%, respectively, in the restrictive and liberal groups (P5.02).

The study was stratified for those persons randomized with cirrhosis. In patients with cirrhosis and a Child-Pugh-Turcotte Class A or B score, restrictive transfusion practices during hospitalization for upper GI bleeding led to a 0.30 hazard ratio for death when compared to a liberal transfusion target.


The editorialist emphasized the importance of this study.1 GI bleeding leads to 450,000 annual hospitalizations in the United States. Treating 25 patients with a restrictive transfusion strategy instead of a liberal one will avert one death at 25 days. This scenario is another clinical instance in which the proverbial “less is more.” The editorialist’s closing quote says it all, “The study by Villanueva et. al. provides important evidence to guide clinical practice. Most patients with upper gastrointestinal bleeding, with or without portal hypertension, should have blood transfusions withheld until the hemoglobin drops below 7 g per deciliter.”



1. Laine L. Blood transfusion for gastrointestinal bleeding. N Engl J Med. 2013;368:75-76.

2. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11-21.