Absolute alcohol is a potentially optimal agent for sclerotherapy of esophageal varices. It is cheap and readily available. We compared the efficacy and safety of alcohol with those of a commonly used sclerosing agent, polidocanol. The study was planned to include patients with previous bleeding from esophageal varices randomly assigned to one of the two treatments. After the inclusion of the first 11 patients (6 in the polidocanol group and 5 in the alcohol group), however, the trial was interrupted because of serious complications in patients treated with alcohol (four major bleeding episodes and one esophageal stenosis). The two agents were of comparable efficacy in the small sample of patients studied. The complications were related to the presence of iatrogenic esophageal ulcers which were more frequent (100% vs 30%) and significantly larger (mean, 1.4 cm vs 0.7 cm, p < 0.05) in patients treated with alcohol.