Differentiating Cushing's disease (CD) from pseudo-Cushing (PC) states may still be difficult in current practice. Because desmopressin (1-deamino-8D-arginine vasopressin, DDAVP), a vasopressin analogue, stimulates ACTH release in patients with CD but not in the majority of normal, obese, and depressed subjects, we investigated its ability to discriminate CD from PC states. One hundred seventy-three subjects (76 with active CD, 30 with PC, 36 with simple obesity, and 31 healthy volunteers) were tested with an iv bolus of 10 microg DDAVP. Sixty-one of these subjects also underwent a control study with saline. DDAVP induced marked ACTH and cortisol rises in CD (P < 0.005 vs. saline, for both ACTH and cortisol) but not in PC. A significant ACTH elevation occurred upon DDAVP administration also in normal and obese subjects, but it was much smaller than that observed in patients with CD (P < 0.0001). A peak absolute ACTH increase (> or =6 pmol/L), after DDAVP, allowed us to recognize 66 of 76 patients with CD and 88 of 97 subjects of the other groups. The same criterion correctly identified 18 of 20 patients with mild CD (24-h urinary free cortisol < or = 690 nmol/day) and 29 of 30 PC, resulting in a diagnostic accuracy of 94%, which was definitely higher than that displayed by urinary free cortisol, overnight 1-mg dexamethasone suppression test, and midnight plasma cortisol. In conclusion, the DDAVP test seems to be a useful adjunctive tool for the evaluation of hypercortisolemic patients chiefly because of its ability to differentiate mild CD from PC states.