Primary prophylaxis with nonselective beta-blockers in high-risk subjects has been shown to be effective in reducing both esophageal variceal bleeding and mortality. Recently it has been suggested that band ligation may be a better option for primary prophylaxis. We compared nonselective beta-blockers with band ligation in patients with large varices (F2, F3) and elevated hepatic venous wedge pressure gradient (HVWPG, > or = 12 mm Hg). All patients were prospectively followed for variceal bleeding, mortality, and treatment-related complications. Based on previous published studies, we estimated that 90 patients in each arm would be required to show a difference in bleeding rate. The study was prematurely terminated when we realized that our estimated sample size was inadequate to show a difference based on the observed bleeding rate. At the time of termination, 31 patients (Child A, 11; B, 14; C, 6), with a mean HVWPG of 19 +/- 9.1 mm Hg, were randomized to either band ligation (group A; n = 16) or beta-blockers (group B; n = 15). Baseline demographics of both groups were similar and the mean follow-up period was 27.4 +/- 12.9 months. During the follow-up, two patients in group A and one patient in group B had bleeding. Nine patients (29%; group A, six; group B, three; P = ns) died due to non-bleeding-related causes and five (16%) patients (group A, three; group B, two) underwent liver transplantation. Treatment-related complication were minimal in both groups. Despite the selection of high-risk patients, the observed bleeding rate was much lower than anticipated. Based on our observed bleeding rates, 424 patients would be required in each arm to show a difference between band ligation and beta-blocker therapy.