Ninety-two patients over the age of 60 with acute upper GI-haemorrhage were included in a prospective randomized doubleblind three-centre study of the effect of a five day treatment with cimetidine. Twenty patients had to be excluded because of different reasons. The remaining 72 patients bled from either erosive gastritis, gastric ulcer or duodenal ulcer. Thirty-two patients received cimetidine and 39 placebo. There was no difference in the number of transfusions, rebleedings or operative interventions. Mortality was 1/33 (3%) in the cimetidine group compared to 5/39 (13%) in the placebo group (NS). In the gastric ulcer group there was no mortality among 10 cimetidine patients compared to a mortality of 4 of 12 (33%) patients receiving placebo (p less than 0.05). It is concluded with caution that cimetidine might be effective in haemorrhage from gastric ulcer in patients 60 years and older. For convincing conclusions a larger study of patients with bleeding ulcers is desirable.