Recent trials have shown that somatostatin (SMT) is as effective as sclerotherapy in the treatment of acute variceal bleeding and that the combination of both treatments is more effective than sclerotherapy alone. To assess whether the addition of sclerotherapy improves the efficacy of SMT alone, all patients admitted to our unit with gastrointestinal bleeding and with suspected cirrhosis received a continuous infusion of SMT (250 micrograms/h). Endoscopy was performed between 1 and 5 hours later, and patients with esophageal variceal bleeding were randomized to receive or not to receive sclerotherapy. In both groups, SMT infusion was continued for 5 days. Fifty patient admissions were allocated to each group. Therapeutic failure occurred in 21 cases of the SMT group and in 7 cases of the combined-therapy group (P =.002). Failure to control the acute episode occurred in 24% vs. 8% (P =.03) and early rebleeding in 24% vs. 7% (P =.03), respectively. Transfusional requirements were significantly higher in the SMT group, while the incidence of complications was lower (8% vs. 24%; P =.029). In the multivariate analysis, the presence of shock at admission and active bleeding during endoscopy were the variables that better predicted the failure of therapy with SMT alone. Mortality at 6 weeks was similar. These data demonstrate that the addition of sclerotherapy significantly improves the efficacy of SMT alone for the treatment of acute variceal bleeding, although it also increases the rate of complications. Patients with shock and those with active bleeding are more likely to benefit from this combined therapy.