Burden of anaemia in upper GI bleeding in a secondary care hospital



Objective Upper GI bleeding (UGIB) remains a considerable burden causing premature mortality throughout the world. Uncertainty exists about the best haemoglobin (Hb) measure to aim for on discharge. Methods The secondary care patient sample was identified from University Hospital Birmingham informatics electronic databases, which has a catchment population of more than 500,000. Demographics, clinical and follow-up information were collected on first admission per patients having: (i). upper gastrointestinal endoscopic procedure; (ii). a Hb measurement; and (iii). an UGIB admission between 1st January 2010 and 31st December 2014. One year survival post discharge was collected and analysed using log rank statistics in Kaplan Meier survival curves. A naïve analysis of all data was compared against an analysis where patients were matched on propensity to anaemia using risk factors in a generalised logistic model. Results There were 1304 cases of upper GI bleeding identified in the 5 year period. The median range of patients was 67 years (interquartile range 54-80 years). 38.6% (504) of patients were female. Ethnicity included: Caucasian 81.7% (1065), South Asian 9.3% (121), and Other 9.0% (118). Prevalence coding of past and current medical history was predominantly Peptic ulcer (50.3%) but also included: Liver disease (18.8%); Diabetes Mellitus (16.9%); Cancer (16.6%). Independent risk factors for overall mortality included: admission method (emergency or elective; p<0.0001), age (p=0.0005), gender (p<0.001), body mass index (p<0.001) and comorbidities (liver disease, renal disease, peptic ulcer all p<0.001). One year survival between those discharged with a Haemoglobin >=10 and <10 when adjusted for age, gender, admission method, demographics, body mass index was 85.1% versus 76.7% (p=0.0034). However, when comorbidities were also adjusted for, there was limited evidence to suggest overall mortality differences (p=0.31). Key messages: 1. This large-scale informatics study in a single centre hospital population has identified associations post-UGIB for a 1 year mortality difference dependent on Hb on discharge. 2. However, this burden appears to be associated with comorbidities rather than anaemia per se. 3. Further studies to assess the long-term effects of post-UGIB anaemia and the benefits of intervention are required.


Neeraj Bhala

Queen Elizabeth Hospital Birmingham

Consultant Physician and Gastroenterologist, University Hospital Birmingham
Honorary Senior Research Fellow, University of Birmingham

Clinical and research background from West Midlands, Oxford, London and Wellington, New Zealand. General and specialist treatment of common gastrointestinal and liver diseases. Special interest in population health science, preventing GI cancers and GI bleeding. Now working on collaborative projects including Global Burden of Disease, regional and national public health initiatives and big data informatics projects.