Thirty-seven patients with volume-retaining disorders (liver cirrhosis with ascites, n = 8; heart failure NYHA III-IV, n = 12; endstage renal failure, n = 17) and twelve healthy age-matched controls were given a small dose (33 micrograms) of hANF (human atrial natriuretic factor). We tested the resulting hemodynamic and renal effects as well as the effect on plasma cyclic GMP levels and compared them with the properties of platelet ANF receptors. The ANF injection evoked an increase in cyclic GMP plasma levels of 19.3 +/- 2.2 nM in healthy controls. This increase tended to be smaller in the cirrhosis group (15.5 +/- 3.3 nM) and in the heart failure group (16.8 +/- 2.3 nM) than in the dialysis group (20.5 +/- 2.5 nM). The invasion rates of cyclic GMP were comparable in all groups, but the evasion rates increased more in the heart failure and endstage renal failure groups (27.9 +/- 7.7 min and 26.1 +/- 3.4 min, respectively) than in the cirrhosis and control groups (14.9 +/- 1.9 min and 14.2 +/- 1.9 min, respectively). Patients with endstage renal failure and congestive heart failure showed a smaller decrease in diastolic blood pressure than controls and patients with liver cirrhosis. Renal actions of ANF were diminished in cirrhosis and heart failure patients. Binding capacities of platelet ANF receptors were higher in the control group (12.2 +/- 1.5 receptors/cell) than in the patient groups (cirrhosis, 7.8 +/- 1.2; endstage renal failure, 8.0 +/- 0.9; heart insufficiency, 8.0 +/- 1.0 receptors/cell), with no differences among the patient groups. Binding affinities were not significantly different. Correlation analysis showed that the relationship between the actions of ANF and the increases in plasma cyclic GMP levels is loose and cannot predict the hemodynamic or renal effects of exogenous ANF in a given patient. Although the behavior of plasma cyclic GMP levels fails to predict the responsiveness of the body to ANF in a given patient, it does reflect the differences between the patient groups and the control group. In contrast, we found no correlation between the properties of platelet ANF receptors and ANF action.