Haematemesis patients should be managed in special units.

This report outlines the results of 568 episodes of acute upper gastrointestinal haemorrhage managed in the Gastroenterology Unit of The Royal Newcastle Hospital during 1964-1974. In this Unit a conservative regimen of blood transfusion and surgery was used with the aim of operating immediately on the recurrence of bleeding for patients with chronic gastric ulcers. In comparison with the 523 patients who were bleeding who were treated in other medical units in the same hospital during 1964-1969, the mortality rate was lower for all peptic ulcers (4.9% compared with 10.6%; P less than 0.025) and for gastric ulcers (8.9% compared with 23.1%; P less than 0.01), but differences for duodenal ulcers (3.6% compared with 5.8%) and in the other diagnostic sub-groups were not significant. The results are also compared with those from units that manage all admissions for upper gastrointestinal bleeding at Prince Henry's Hospital, Melbourne (which has an aggressive transfusion and surgical policy), and The Royal North Shore Hospital of Sydney (which has a "standard" approach). In spite of radically different policies, all three gastrointestinal units obtained rather similar results, with the Newcastle mortality rate from bleeding ulcers of 4.9% being the lowest of all. It is suggested that all large hospitals should have haematemesis and melaena units, as they do coronary care units, but that there is no advantage to be gained by a policy of aggressive treatment.