The atrial natriuretic peptide hormonal system is altered to a variable degree in patients with cirrhosis. Portal pressure and portal-systemic shunting are also varied in cirrhosis. We used a portal vein–ligated rat model with predictable portal hypertension to study the effects of portal hypertension alone on the atrial natriuretic peptide hormonal system. Sham-operated rats were used as controls. Mean portal pressure was significantly increased in portal vein–ligated rats (portal vein–ligated rats, 21.7 ± 0.74 cm H2O; sham-operated rats, 13.7 ± 0.47 cm H2O; p < 0.0001). Plasma atrial natriuretic peptide decreased 50% in the portal vein–ligated rats (p < 0.0001). Atrial natriuretic peptide messenger RNA level was decreased by 40% to 60% in the left and right atria and in the ventricles of portal vein–ligated rats (p < 0.05 for each chamber). Only one class of glomerular binding site was identified by competitive binding studies. The atrial natriuretic peptide glomerular receptor density increased in the portal vein–ligated rats (portal vein–ligated rats, 1,660 ± 393; sham-operated 725 ± 147 fmol/mg protein, p < 0.02), whereas affinity decreased (portal vein–ligated, 1.69 ± 0.49; sham-operated, 0.55 ± 0.12 nmol/L, p < 0.02). No difference was seen in the amount of cyclic GMP generated by atrial natriuretic peptide stimulation in isolated glomeruli from portal vein–ligated and sham-operated rats. Thus, in the rat model of portal hypertension without liver disease, decreased plasma atrial natriuretic peptide level is associated with decreased cardiac atrial natriuretic peptide messenger RNA and increased density (but decreased affinity) of glomerular atrial natriuretic peptide binding sites. Portal hypertension alone may account for the increased density of glomerular atrial natriuretic peptide binding sites reported in cirrhosis, but it does not account for the changes in cardiac atrial natriuretic peptide and plasma atrial natriuretic peptide reported by most investigators. (Hepatology 1992;15:696–701).