The success of a defined management policy op peptic ulcer haemorrhage which incorporates endoscopic therapeutic intervention depends on the early identification of a high risk group of patients and a high risk group of ulcers. The high risk group of patients consists of those likely to experience further bleeding on the basis of clinical prognostic indicators: shock and severe anaemia on admission and the pattern of bleeding; or tolerate rebleeding and emergency surgery poorly: patients over 60 years and those with associated disease. UGI endoscopy should be performed early (within 6-12 hours) in this group in order to identify the bleeding point and provide prognostic information regarding the risk of further haemorrhage. Peptic ulcers with major stigmata of recent bleeding (spurting or non-bleeding visible vessel) have high risk of rebleeding, the risk is even greater when major stigmata of recent haemorrhage (SRH) are associated with shock on admission. Patients with such ulcers should be monitored intensively and receive endoscopic haemostatic treatment in order to terminate active haemorrhage or prevent rebleeding thereby avoiding the need for emergency surgery with its attendant morbidity and mortality. Patients with ulcers with minor or no SRH have a very low risk of rebleeding and don't require intensive monitoring or endoscopic treatment and can be discharged from hospital early. Ulcers which cannot be completely characterized have an intermediate risk of rebleeding and should be managed as high risk lesions. Secondary to the anatomy of the visible vessel any haemostatic endoscopic treatment should be applied around, but avoiding, the sentinel clot. Well-designed randomized controlled trials of endoscopic haemostatic treatment of peptic ulcer haemorrhage in which stratification of risk was based on the SRH, have demonstrated for non-bleeding vessel a significant reduction in rebleeding and in emergency surgery, for spurting bleeding benefit was found only for the rebleeding risk. No advantage was demonstrated in each group of patients in term of mortality. Such studies also demonstrate the superiority of the Nd:YAG laser over the Argon laser. Perforation is a rare complication of Nd:YAG laser photocoagulation (less than 1%). Precipitation or aggravation of arterial haemorrhage during treatment of a visible vessel, as a result of a direct hit, is a more frequent complication (0-29%). Further laser treatment is successful in terminating 75% of these induced bleeds, the remainder requiring surgery. Preinjection of the ulcer with adrenaline does not appear to prevent this complication.(ABSTRACT TRUNCATED AT 400 WORDS)