Most patients with acute upper gastrointestinal haemorrhage are managed conservatively or with endoscopic intervention but some ultimately require surgery to arrest the haemorrhage. We have conducted a population-based multicentre prospective observational study of management and outcomes. This paper concerns the subgroup of 307 patients who had an operation because of continued or recurrent haemorrhage or high risk of further bleeding.
The principal diagnostic group was those with peptic ulcer. Of 2071 patients with peptic ulcer presenting with acute haemorrhage, 251 (12%) had an operative intervention with a mortality of 24%. In the non-operative group mortality was 10%. The operative intervention rate increased with risk score, ranging from 0% in the lowest risk categories to 38% in the highest. Much of the discrepancy between operative and non-operative mortality was explainable by case mix; however, for high-risk cases mortality was significantly higher in the operated group. In 78% of patients who underwent an operation for bleeding peptic ulcer there had been no previous attempt at endoscopic haemostasis. For patients admitted to surgical units, the operative intervention rate was about four times higher than for those admitted under medical teams.
In patients with acute upper gastrointestinal haemorrhage operative intervention is infrequent and largely confined to the highest-risk patients. The continuing high mortality in surgically treated patients is therefore to be expected. The reasons for the low use of endoscopic treatment before surgery are not revealed by this study, but wider use of such treatments might further reduce the operative intervention rate. Physicians and surgeons have not yet reached consensus on who needs surgery and when.