prospective, randomised controlled trial

We conducted a prospective randomised controlled trial of 137 patients with massive peptic ulcer haemorrhage over a period of 12 months to compare the haemostatic effects of endoscopic heat probe thermocoagulation and pure alcohol injection. Seventy eight patients (56 9%) were in shock at the time ofrandomisation to the trial. The age, sex, number of patients in shock, haemoglobin value at the time of entry to the trial, number of patients with severe medical iliness, location of bleeders, and stigmata of recent haemorrhage were comparable among the heat probe, pure alcohol, and control groups. The initial haemostatic effect of the heat probe was better than that of the pure alcohol injection (44 of 45 v 31 of 46, p=00004). The ultimate haemostasis achieved by the heat probe group (41 of 45) was better than that of the pure alcohol group (31 of 46, p=0012) and ofcontrols (24 of 46, p=00001). The duration of hospital stay was shorter for patients in the heat probe group than for the control group (6-2 days v 13-8 days, p<005). The incidence of emergency surgery was less for the heat probe than the control group (three of 45 v 12 of 46, p=0027).' The mortality rate was less in the heat probe than in the control group (one of 45 v seven of 46, p=0031). We suggest that heat probe thermocoagulation should be the first treatment of choice for arrest of massive peptic ulcer haemorrhage.