The effect of an angiotensin II blockade in lowering the portal pressure in patients with liver cirrhosis and portal hypertension is controversial. This prospective study was undertaken to evaluate the portal hypotensive effect of captopril compared to that of propranolol, and to determine the factors that contribute to a successful reduction in the portal pressure after longterm captopril administration in patients with liver cirrhosis.The hepatic venous pressure gradient (HVPG) and portal venous velocity (PVV) were measured both before and 3 months after initiation of the administration of captopril (n = 29) or propranolol (n = 29) in cirrhotic patients with a variceal bleeding episode. Patients who showed a reduction in the HVPG of more than 20% of the baseline were defined as being responders.At 3 months, the mean reduction in the HVPG after captopril was less than that after propranolol (-3.0 +/- 9.3% vs -28.5% +/- 4.1%; P < 0.05). However, of the 29 patients receiving captopril, 9 were classified as being responders. On multivariate analysis with parameters including age, cause, Child-Pugh score, HVPG, and PVV, only low PVV was found to be a significant independent factor for responders (PVV < 12 cm/s; odds ratio [OR], 12.2; 95% confidence interval [CI], 1.47-102.40) in the captopril group.Longterm captopril administration reduces the portal pressure effectively in cirrhotic patients with a low PVV. This suggests that the reduction in portal pressure after captopril administration is a result of improved portal venous outflow brought about by a decrease in the intrahepatic vascular resistance. When the PVV is below 12 cm/s, a captopril trial might be useful in preventing variceal bleeding in portal hypertensive patients.