Endoscopic injection sclerotherapy has been accepted as the procedure of choice for patients with variceal hemorrhage. To evaluate the efficiency of endoscopic injection sclerotherapy in patients with bleeding esophageal varices, we did a retrospective study of 52 patients (non-sclerotherapy group) with bleeding esophageal varices who were admitted to hospitals and did not receive sclerotherapy and of 50 patients (sclerotherapy group) who received sclerotherapy with ethanolamine oleate. The mortality (sclerotherapy group vs. non-sclerotherapy group: 18.0% vs. 32.7%) during index hospitalization, the bleeding risk factor (the number of rebleeds per patient/month; 1.56 +/- 2.76 vs. 4.96 +/- 9.99: mean +/- SEM) and the mortality due to bleeding (14.0% vs. 36.5%) were higher in the non-sclerotherapy group than in the sclerotherapy group. Only those in Child's class C who received sclerotherapy had a significantly better survival rate than the non-sclerotherapy group (p less than 0.05). Although formal comparisons were not made because of the retrospective nature of this study, endoscopic injection sclerotherapy is effective and appears to be superior to conventional medical treatments.