Physical exercise increases portal pressure (hepatic venous pressure gradient [HVPG]) in patients with cirrhosis. It is unknown if this deleterious effect is associated with changes in gastroesophageal collateral blood flow and if these can be prevented by propranolol administration. The aim of this study was to characterize the effects of propranolol on the splanchnic hemodynamic response to exercise in patients with cirrhosis. Twenty-three patients with cirrhosis and portal hypertension had hemodynamic measurements in baseline conditions, and during moderate cycling exercise (40 W) under double-blind propranolol or placebo administration. In patients receiving placebo, HVPG significantly increased during exercise (from 16.7 +/- 0.9 to 19.0 +/- 1.0 mm Hg; P < .01), hepatic blood flow (HBF) decreased (-18% +/- 4%; P < .01), while azygos blood flow (AzBF) was unchanged (4% +/- 12%; ns). In patients receiving propranolol, portal pressure did not increase during exercise, but decreased from 16.3 +/- 1.0 to 12.9 +/- 1.1 mm Hg (P < .01). The lack of increase in HVPG in response to exercise in patients receiving propranolol may be related to a more pronounced decrease in HBF, as compared with patients receiving placebo, and to a blunted increase in cardiac output (CO). Moderate physical exercise adversely influences the hepatic hemodynamics in patients with cirrhosis, causing a significant increase in portal pressure. This is effectively prevented by propranolol pretreatment.