Abstract

One hundred and seven consecutive patients presenting with significant peptic ulcer haemorrhage were randomised to endoscopic injection with 3-10 ml of 1:100,000 adrenaline (55 patients, group 1) or to a combination of adrenaline and 5% ethanolamine (52 patients, group 2). All had major stigmata of haemorrhage and endoscopic injection was undertaken by a single endoscopist. The groups were well matched with regard to risk factors. Rebleeding occurred in eight of the group 1 patients and seven in the group 2 patients; surgical operation rates, median blood transfusion requirements, and hospital stay were similar in both groups. The efficacy of either form of injection was similar whether patients presented with active bleeding or a non-bleeding visible vessel. No complications occurred. In patients presenting with significant peptic ulcer bleeding, the addition of a sclerosant confers no advantage over injection with adrenaline alone.