Background Blood transfusion is a major component in the treatment of acute upper gastrointestinal bleeding. Its safety has been questioned in the treatment of less severe cases. A large randomised controlled trial showed significant reductions in re-bleeding and mortality rates in patients who were randomised to a restrictive policy (transfused when haemoglobin was less than 7 g/dL) compared to a liberal policy (transfused when haemoglobin was less than 9 g/dL).

Aim The aim of this study is to identify whether Muhimbili National Hospital in Tanzania follows the evidence-based medicine that a restrictive policy in blood transfusion is advantageous in the treatment of acute upper gastrointestinal bleeding.

Method I performed a prospective audit of patients presenting to Mulhimbili National Hospital (MNH) with AUGIB to assess the current transfusion practice in this patient group. During a 3 week period I identified patients admitted to MNH who presented with meleana and haematemesis. I collected data about haemoglobin levels and blood transfusion from patient’s notes.

Results 89.5% of patients who presented with haemoglobin <7 g/dL received a blood transfusion. 50% of the patients with a haemoglobin >9 g/dL did receive a blood transfusion.

Conclusions Muhimbili National Hospital in Dar Es Salaam in Tanzania do not follow the evidence-based medicine that restrictive policy in blood transfusion in acute upper gastrointestinal bleeding is beneficial compared to liberal blood transfusion.