To determine the potential usefulness of atrial natriuretic peptide (ANP) in patients with cirrhosis, we examined the effects of the infusion of a low dose of alpha-human ANP (alpha hANP, 25 ng.kg-1.min-1 for 30 min) on renal, splanchnic, systemic hemodynamics and sympathetic outflow in eight patients. Pulmonary arterial plasma ANP concentrations increased from 59 +/- 9 to 328 +/- 41 pg/ml (mean +/- S.E., p less than 0.05). Mean values of glomerular filtration rate and renal plasma flow were not significantly changed. Individual renal plasma flow responses differed from one patient to another. Renal plasma flow increased in two patients, decreased in three and did not change in the other patients. Renal plasma flow changes were correlated with basal renal plasma flow values (r = -0.938, p less than 0.05) but not with arterial pressure changes or renal vein plasma norepinephrine concentration changes. Azygos blood flow increased from 0.43 +/- 0.10 to 0.63 +/- 0.13 l/min (p less than 0.05) and the hepatic-venous pressure gradient decreased from 19.9 +/- 1.5 to 17.5 +/- 2.9 mmHg in post-infusion (p less than 0.05). Mean arterial pressure decreased significantly by 18% and cardiac output by 12%. Systemic vascular resistance and pulmonary arterial plasma norepinephrine concentrations were not significantly modified. Thus, in patients with cirrhosis, alpha hANP appears to have a direct vasodilating action on renal arterioles when basal renal vascular tone is high. In addition, although alpha hANP might exert a portal hypotensive action, alpha hANP induced arterial hypotension as a result of both low cardiac output and a lack of increased sympathetic vascular tone. The arterial hypotensive action may, thus, limit the therapeutic use of low doses of alpha hANP in cirrhotic patients.