Article Text


* Red cell transfusion practice in patients presenting with acute upper gastrointestinal bleeding- a survey of 815 clinicians
  1. V Jairath *1,
  2. B Kahan2,
  3. R Logan3,
  4. K Palmer4,
  5. S Travis5,
  6. M Murphy1
  1. 1NHS Blood and Transplant, Oxford, UK
  2. 2MRC Clinical Trials Unit, London, UK
  3. 3Division of Epidemiology and Public Health, Nottingham, UK
  4. 4Western General Infirmary, Edinburgh
  5. 5John Radcliffe Hospital, Oxford, UK


Introduction Acute Upper Gastrointestinal Bleeding (AUGIB) accounts for 14% of all red blood cell (RBC) transfusions in England. Studies of RBC transfusion practice in critical care and cardiac surgery have shown widespread practice variation. Since there are no such data in AUGIB and, as a consequence of recent evidence that early transfusion is associated with adverse outcomes in AUGIB,1 the current study was designed to assess attitudes towards RBC transfusion practice in patients presenting with AUGIB.

Methods A survey depicting six vignettes representing a range of clinical scenarios in AUGIB were piloted for content and emailed to practicing consultants and registrars in gastroenterology, acute medicine and upper gastrointestinal surgery through their member organisations. Respondents were asked to select a haemoglobin (Hb) trigger at which they would ordinarily transfuse red blood cells. Results were analysed using proportional odds models and robust standard errors were used to account for clustering.

Results The overall response rate was 48% (815/1709). Transfusion triggers differed significantly between all six scenarios (p<0.001). A Hb trigger >9g/dl was chosen by only 13.3% of respondents to scenario 2 (young patient with features of shock). In comparison, a Hb trigger of >9 g/dl was chosen for 45.2% of respondents in scenario 4 representing an older patient with ischemic heart disease. <4% of clinicians reported that they would ever transfuse at a Hb trigger of >10 g/dl in all scenarios. The selected Hb trigger differed between clinical specialities. Surgeons selected a lower Hb transfusion trigger than physicians across all 6 scenarios (p<0.005), as did clinicians who had graduated within the past 5 years(p<0.05 compared to graduants >5 years). Only 70% of respondents reported familiarity with national guidelines for AUGIB. Most clinicians favoured a restrictive approach to transfusion, which is in part discordant with actual transfusion practice observed in the UK National Audit of AUGIB, where 15% of all early RBC transfusions were administered to patients with a Hb >10g/dL (38% of whom were haemodynamically stable) and to 50% of admissions with a presenting Hb 8.1–10 g/dl and haemodynamic stability.

Conclusion There is widespread variation in attitudes towards RBC transfusion for AUGIB. Both clinician and patient characteristics influence RBC transfusion decisions. Clinical studies are needed to evaluate the safety and efficacy of differing blood transfusion strategies in AUGIB.

Statistics from


  • Competing interests None.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.