Controlled trials of endoscopic sclerotherapy for the prevention of the first variceal hemorrhage have given controversial results. We continued a previously reported study and randomly assigned 141 patients with esophageal varices and no prior gastrointestinal bleeding to either prophylactic sclerotherapy (n = 70) or no treatment (n = 71). Sclerotherapy was performed until complete eradication of the varices was achieved; recurrent varices were treated with repeat sclerotherapy. The groups were well balanced in terms of demographic and clinical characteristics. Patients in both groups who bled from varices received sclerotherapy whenever possible. During a median follow-up of 56 months, variceal bleeding occurred in 7% in sclerotherapy patients and 44% of control patients (p < 0.01). In the sclerotherapy group 59% died, and in the control group 51% (n.s.). In both groups, the mortality rate increased with the severity of liver function impairment. Sclerotherapy was not found to improve survival in patients, irrespective of the etiology of cirrhosis (alcoholic or nonalcoholic) or variceal size (low-grade or high-grade). We conclude that sclerotherapy is a suitable method to reduce the occurrence of the first variceal hemorrhage, but it does not appear to have an effect on survival.