Outcomes following restrictive or liberal red blood cell transfusion in patients with lower gastrointestinal bleeding

Background Restrictive red blood cell (RBC) transfusion reduces mortality and rebleeding after upper gastrointestinal bleeding (UGIB). However, there is no evidence to guide transfusion strategies in lower gastrointestinal bleeding (LGIB). Aim To assess the association between RBC transfusion strategies and outcomes in patients with LGIB Methods This was a post hoc analysis of the UK National Comparative Audit of LGIB and the Use of Blood. The relationships between liberal RBC transfusion and clinical outcomes of rebleeding, mortality and a composite outcome for safe discharge were examined. Transfusion strategy was dichotomised and defined as liberal when transfusion was administered for haemoglobin (Hb) >= 80 g/L (or >= 90 g/L in patients with acute coronary syndrome) or major haemorrhage, and restrictive otherwise. Multivariable logistic regression models were used to assess the independent association between liberal RBC transfusion and outcomes. Results Of 2528 consecutive patients enrolled from 143 hospitals in the original study, 666 (26.3%) received RBC transfusion (mean age 73.3 +/- 16 years, 49% female, initial mean haemoglobin 90 +/- 24 g/L, 2.3% had haemodynamic instability). The rebleeding rate in transfused patients was 42.3%. After adjusting for potential confounders, there was no difference between liberal and restrictive RBC transfusion strategies for the odds of rebleeding (OR 0.89, 95% CI 0.6-1.22), in-hospital mortality (OR 0.54, 95% CI 0.3-1.1) or of achieving the composite outcome (OR 0.72, 95% CI 0.5-1.1). Conclusion Although these results could be due to residual confounding, they provide an important foundation for the design of randomised trials to evaluate transfusion strategies for LGIB.