The role of endoscopic thrombin injections in the treatment of gastroduodenal bleeding.
One hundred and ninety-one patients who had emergency endoscopy performed for acute gastroduodenal haemorrhage and were found to be bleeding actively (N = 141) or to have an ulcer containing a non-bleeding visible vessel (N = 50) were studied. Patients with active bleeding were divided into two groups: group I included 47 patients in whom diagnostic endoscopy was performed and conventional management applied; and group II contained 94 patients in whom, apart from conventional treatment, endoscopic thrombin injections could also be offered. Fifty patients with non-bleeding visible vessels were admitted to a prospective randomised trial and assigned to either endoscopic injections (group A, N = 25), or no primary endoscopic treatment (group B, N = 25). The introduction of the injection therapy in patients who were actively bleeding resulted in a significant reduction of the emergency operation rate (70.2% v. 12.8%; P < 0.001), mortality rate (21.3% v. 8.5%; P < 0.05 and blood transfusion requirements (4,220 +/- 1,840 ml v. 1,980 +/- 980 ml; P < 0.05). Endoscopic thrombin injections of non-bleeding visible vessels reduced the rebleeding rate (40% v. 4%; P < 0.01) and blood transfusions (2,939 +/- 1070 ml v. 1,390 +/- 552 ml; P < 0.05).