In patients with cirrhosis, vasopressin infusion induces sustained vasoconstriction and elevation of arterial pressure. The vasopressor effect could be caused by impairment of mechanisms normally aimed at buffering increases in arterial pressure (reflex bradycardia and decreases in arteriolar resistance). We studied the acute effects of continuous vasopressin infusion (0.4 IU/min) on systemic hemodynamics in seven patients with cirrhosis and in six patients without cirrhosis (controls). Vasopressin effects on systemic O2 consumption were also studied. In both groups, vasopressin infusion induced similar peak increases in arterial pressure, followed by similar decreases in heart rate and cardiac output. However, cirrhotic patients and controls differed 30 min after the start of vasopressin infusion. At 30 min, mean arterial pressure, diastolic arterial pressure and systemic vascular resistance remained significantly higher than preinfusion values in patients with cirrhosis. No decrease in systemic O2 consumption occurred in cirrhotic patients. In controls, at 30 min, mean arterial pressure and diastolic arterial pressure had returned to baseline. Systemic vascular resistance was not significantly higher than the preinfusion value and systemic O2 consumption had significantly decreased to below preinfusion values. We conclude that the vasopressor effect of vasopressin is abnormally sustained in patients with cirrhosis. This might be caused by insufficient buffering of vasopressin-induced arteriolar constriction rather than by abnormal vagal control of heart rate. In turn, as suggested by the lack of a decrease in systemic O2 consumption, persistent arteriolar constriction might be related to abnormally sustained sympathetic vascular tone in patients with cirrhosis.