A prospective, randomized, placebo-controlled, double-blind, multicenter clinical trial of intravenous somatostatin (Stilamin; Serono Laboratories, Inc., Randolph, MA) was performed in 102 patients with actively bleeding esophageal varices from August, 1985, to November, 1986. Patients had major hemorrhage indicated by hematemesis or melena and evidence of significant blood loss. For entry, patients had to have endoscopic demonstration of active bleeding from esophageal varices or stigmata of recent hemorrhage and bright red blood in the gastric aspirate with no other source of bleeding found. Randomized patients received identical-appearing somatostatin or placebo for a 30-hr study period. Those given somatostatin received a 250-micrograms bolus and a 250-micrograms per hr infusion with repeat bolus and doubling of the infusion if the bleeding was not controlled. In retrospect, 18 patients could not be evaluated. Of the 84 evaluable patients, 48 received somatostatin and 36 placebo. They were comparable in age, gender, severity of liver disease and history of variceal bleeding. Transfusion requirements were similar in both groups. Bleeding stopped for 12 consecutive hr during 30 hr of the study period in 31 (65%) of the somatostatin group vs. 30 (83%) of the placebo group (p = 0.06). The median time to cessation of bleeding was 2 hr in the placebo group and 3 hr in the somatostatin group. Deaths following the study period were nine (25%) in the placebo group and 15 (31%) in the somatostatin group. Within the limitations of the present study, we conclude that somatostatin was ineffective in the management of active bleeding of esophageal varices.