Since 1989, 45 patients 26 females and 19 males, aged 19-79 years (median 58) bearing incidentally discovered adrenal masses were studied. Endocrine work-up included determination of urinary catecholamines and their metabolites, measurement of plasma renin activity and aldosterone levels in clino- and orthostatic posture, basal and dynamic (dexamethasone-suppression, o-CRH stimulation) evaluation of hypothalamic-pituitary-adrenal (HPA) axis. The most frequent finding was the reduction of DHEA-S levels below the 3rd percentile of controls in 19 (42%) patients. As a whole group, DHEA-S levels were significantly lower in patients than in controls: 68 (5-1000) micrograms/dL vs 208 (34-326) micrograms/dL; p < 0.001. Three patients (7%) had high 24-h mean serum cortisol levels and 6 (14%) had blunted day-night amplitude of cortisol rhythm. Defective dexamethasone suppressibility was found in 15% of patients vs 8% of controls (p < 0.05). ACTH and cortisol responses after o-CRH did not significantly differ between patients and controls although blunted ACTH responses were found in 22% of cases. The above mentioned endocrine alterations could be accounted for by autonomous cortisol secretion by the adrenal nodule at a rate not sufficient to give clinical expression but able to inhibit to some extent the HPA axis. These results indicate that silent cortisol hypersecretion is frequently observed in patients with adrenal incidentaloma even if progression to overt Cushing's syndrome seems unlikely, at least in a short-term follow-up. From a mere cost-benefit ratio, the evaluation of DHEA-S levels and dex-suppression has sufficient sensitivity to identify the occurrence of silent hypercortisolism.