Background: There exists considerable practice variation and little evidence to guide red blood cell (RBC) transfusion in patients with nonvariceal upper gastrointestinal bleeding (NVUGIB). Studies in other critically ill cohorts suggest associations between transfusions and adverse patient outcomes.
Aim: To characterise any possible clinically-relevant association between RBC transfusion following NVUGIB with rebleeding and mortality.
Methods: Observational study utilising the Canadian Registry of patients with Upper Gastrointestinal Bleeding and Endoscopy (RUGBE). Multivariable logistic regression models were used to examine and quantify independent associations between RBC transfusion and clinical outcomes.
Results: Overall, 1677 patients were included (66.2 ± 16.8 years, 61.7% male, 2.5 ± 1.7 comorbid conditions, initial haemoglobin, 96.8 ± 27.2 g/L); 53.7% received RBC transfusions (2.9 ± 1.6 units of blood), 31.6% had haemodynamic instability, 5.1% fresh blood on rectal examination and 8.6% in the nasogastric tube aspirate. Endoscopic haemostasis was performed in 35.2%. Overall rebleeding (defined as continuous bleeding, rebleeding or surgery) and mortality rates were 17.9% and 5.4%, respectively. After adjusting for potential confounders, transfusion of RBC within 24 h of presentation was significantly and independently associated with an increased risk of rebleeding (OR: 1.0, 95% CI: 0.6-1.8), but not death (OR: 1.5, 95% CI: 0.94-2.23).
Conclusions: This study suggests an association between RBC transfusion following NVUGIB and subsequent rebleeding, after appropriate and extensive adjustment for confounding. Prospective randomised trial evidence is needed to identify the most efficacious and cost-effective transfusional strategies in these patients.
© 2012 Blackwell Publishing Ltd.