The role of atrial natriuretic peptide (ANP) in glucocorticoid hypertension is still controversial, as glucocorticoids enhance the secretion of ANP in vivo, but reduce its biological activity in vitro. In isolated cells, for example, the cGMP response to ANP is suppressed by dexamethasone. We tested the in vivo relevance of this observation by comparing the cGMP, endocrine, and renal responses to exogenous ANP in patients with Cushing's disease (CD; n = 10) and in a matched group of essential hypertensives (EH; n = 8) and normotensive controls (N; n = 4). alpha-human-ANP was infused at 0.01 microg/kg/min for 120 min with a 30-min recovery period; hormonal and arterial pressure measurements were performed at 30-min intervals, and renal parameters were measured at baseline and after infusion. There was a 4-fold increase in plasma ANP in all groups, but the increments in plasma cGMP were about 50% lower in CD than in N and EH; urinary cGMP was similarly lower in CD (247 +/- 61 vs. 529 +/- 146 and 384 +/- 54 nmol/150 min, respectively). This was associated with a reduced peak increase in hematocrit in CD (+2.6 +/- 0.9%) compared with N (+6.6 +/- 0.9%) and EH (+7.1 +/- 0.7%; P < 0.05 CD vs. both); the diuretic and natriuretic effects of ANP were also lower in CD than in EH with very similar systemic pressure levels (382 +/- 63 vs. 848 +/- 130 mL/150 min, and 61 +/- 14 vs. 113 +/- 14 mmol/150 min, respectively; P < 0.05 for both). The increments in plasma and urinary cGMP in response to physiological doses of ANP are thus blunted in CD patients compared with those in N and EH. This biochemical defect is associated with reduced capillary permeability and a lesser diuretic and natriuretic effect. These data are compatible with an impairment of the biological activity of ANP in glucocorticoid hypertension in humans.