We studied the effects of tranexamic acid (an antifibrinolytic agent) and cimetidine on acute upper-gastrointestinal-tract bleeding in a double-blind randomized placebo-controlled trial in 775 patients with hematemesis or melena or both. Mortality was significantly reduced in patients receiving either tranexamic acid (mortality, 6.3 per cent) or cimetidine (7.7 per cent), as compared with patients receiving placebo (13.5 per cent) (P = 0.0092 for tranexamic acid vs. placebo, P = 0.045 for cimetidine vs. placebo). Ninety-nine patients were withdrawn before the code was broken, mainly because their primary illness was considered not to be due to acute upper-gastrointestinal-tract bleeding. Mortality among those withdrawn was high (22 per cent), and their exclusion reduced death rates to 4 per cent in those given tranexamic acid, 8 per cent in those given cimetidine, and 11 per cent in those given placebo (P = 0.0072 for tranexamic acid vs. placebo, P greater than 0.50 for cimetidine vs. placebo). The reduced mortality associated with tranexamic acid was detectable at both participating hospitals and in most of the main subgroups of patients classified according to site of bleeding. However, treatment with this agent was not associated with any decrease in the rate of rebleeding or the need for operation.