Sixty-three patients with degree III or IV esophageal varices and the so-called red color sign, but without previous bleeding were randomly assigned to either prophylactic sclerotherapy (PST) (n = 30) or to a control group (n = 33). In 58 cases the portal hypertension was caused by liver cirrhosis (40% alcoholics). The two groups were comparable with respect to demographic data and endoscopic appearance, causes and severity of liver damage. Sclerotherapy was performed as combined intra-and paravariceal injections of 2 or 3% polidocanol. All patients, both in the treatment and in the control groups, who bled from varices after randomization, received sclerotherapy until the varices were eradicated, and remained in their groups. After a mean follow-up of 44.5 months, the bleeding rate in the PST group was significantly lower (30% vs 75%, p less than 0.01). The difference became significant from the second year onward. Fourteen patients of the PST group and 19 of the controls died (4 and 14, respectively, p less than 0.05 as a result of the bleeding). Life table analysis (Kaplan-Meier) revealed no differences in survival between the two groups. At the present time PST cannot yet be recommended as a method for clinical routine use.