The purpose of this study is to verify whether acute pre-treatment with alprazolam (ALP), a benzodiazepine that inhibits HPA secretion in normal subjects, could better characterize patients with subclinical Cushing's syndrome (SCS) than the 1-mg dexamethasone test (DST). In 22 patients with SCS, 10 with overt Cushing's syndrome (CS), 11 with non-functioning adrenal incidentalomas (NF) and 14 normal subjects (NS) we studied the effect of ALP (1 mg, p.o. at 2300 hours) on cortisol levels after 1-mg DST. Cortisol levels (mean ± SEM) after DST were lower (P = 0.012) in SCS (3.9 ± 0.3 μg/dl) than in overt CS (10.4 ± 1.9 μg/dl), while they were higher (P = 0.0005) than in NF (1.1 ± 0.1 μg/dl) and NS (1.5 ± 0.1 μg/dl). After ALP pre-treatment, cortisol levels further decreased (P = 0.004) in SCS (3.0 ± 0.3 μg/dl), but neither in CS (9.3 ± 1.3 μg/dl) nor in NF (1.3 ± 0.1 μg/dl) and in NS (1.3 ± 0.1 μg/dl). In SCS, cortisol levels after ALP + 1-mg DST persisted lower (P = 0.0005) than those in CS, but higher (P = 0.0005) than those in NF and NS. Considering individual cases, ALP pre-treatment reduced cortisol levels < 3 and < 1.8 μg/dl in 50 and 23 % of SCS patients, respectively. ALP amplifies the cortisol inhibition exerted by 1-mg DST in patients with SCS but not in those with CS. The clinical usefulness of ALP to increase the sensitivity of 1-mg DST to identify true autonomous cortisol release in patients with adrenal incidentalomas as well as to predict different clinical outcomes remains to be clarified.