The present double-blind study was aimed at investigating the hemodynamic and humoral effects of somatostatin or placebo in patients with cirrhosis. Patients were randomly assigned to receive either an injection of 250 micrograms of somatostatin followed by a constant infusion of somatostatin at 250 micrograms/h (n = 13), an injection of 250 micrograms of somatostatin followed by a 500 micrograms/h continuous infusion (n = 10), or an injection of placebo followed by a placebo infusion (n = 9). Placebo had no effect. Somatostatin bolus markedly decreased the hepatic venous pressure gradient: by 52% at 1 minute; P < .001; 19% at 3 minutes, P < .01; and by 13% at 5 minutes, P < .04. Azygos blood flow decreased similarly by 45% at 1 minute, P < .001; 16% at 3 minutes, P < .02; and 9.5% at 5 minutes, P = .05. Mean arterial pressure increased by 25% (P < .001). Continuous somatostatin infusions (250 or 500 micrograms/h) had no systemic effects, but significantly reduced hepatic venous pressure gradient (250 micrograms/h: -6.1%, P < .05 and 500 micrograms/h: -15%, P < .01) and hepatic blood flow (250 micrograms/h: -10%, 500 micrograms/h: -18%, P < .05). Azygos blood flow was not changed after 250 micrograms/h infusion but was reduced after 500 micrograms/h (-23%, P < .02). Somatostatin but not placebo, suppressed glucagon to normal levels. This study shows that a bolus injection of somatostatin caused an immediate and marked decrease of hepatic venous pressure gradient and azygos blood flow. Continuous infusion of somatostatin had a mild but sustained effect on splanchnic hemodynamics; this effect was more pronounced with the higher dose.