The current concept that blunted adrenocorticotropic hormone (ACTH) response to human corticotropin-releasing-hormone (h-CRH) in depression is primarily determined by elevated circulating plasma cortisol levels is still unproven. We tested this hypothesis by comparing ACTH release following intravenous administration of 100 micrograms h-CRH in 10 normal controls and in 21 inpatients with a major depressive episode. Eleven of these depressed patients were pretreated with an oral dose of 2 g metyrapone, which inhibits cortisol biosynthesis by blocking C-11 beta-steroid-hydroxylase. This intervention deprives the entire system of cortisol, which is the major feedback signal for the regulation of ACTH secretion at various pituitary and limbic sites. ACTH responses, assessed as areas-under-time-course-curves, were: in normal controls, 6.8 +/- 2.4 (SD) pg/ml/min x 10(3); in unmedicated patients, 2.6 +/- 1.1 pg/ml/min x 10(3); and in metyrapone pretreated patients, 9.0 +/- 6.7 pg/ml/min x 10(3). Thus, ACTH release in unmedicated depressed patients was significantly (p less than 0.001, Mann-Whitney U-test) blunted when compared with normal controls. In contrast, this blunting was completely avoided after metyrapone pretreatment, which resulted in net ACTH responses that were indistinguishable from those of the controls.