Sengstaken-Blakemore tamponade is used for the initial control of bleeding esophageal varices (BEV), although it is known to be potentially dangerous. Sclerotherapy has been shown to be effective in the treatment of BEV. This trial has been designed to evaluate comparatively the effectiveness of both procedures in the initial control of the hemorrhage. Forty-three patients with BEV were included in the trial. Twenty patients (group SB) were treated by tamponade. Twenty-three patients (group ST) were treated by sclerotherapy by means of a simple technique. During the first 24 hours, hemostasis was obtained in 16 of the SB patients and in all 23 of the ST patients (p less than 0.05). At seven days, nine SB patients and 19 ST patients were free of hemorrhagic relapse (p less than 0.05). By stratifying in relation to hepatic failure, the difference was greater (p less than 0.005) if patients with Child's A classification were excluded. It is concluded that sclerotherapy should be undertaken in almost every instance at the same moment that diagnosis is made, bypassing the intermediate step of tamponade.