A stimulation test with 100 micrograms ovine or human corticotropin-releasing factor (CRF) is a useful diagnostic tool in diseases of the hypothalamo-pituitary-adrenal axis. No serious side effects were observed during the test procedure. The results showed that the CRF test is useful in making the differential diagnosis of established Cushing's syndrome (n = 42). The CRF test was also repeated after transsphenoidal surgery in 25 patients with Cushing's disease. Successfully operated patients exhibit no, blunted or normal adrenocorticotropic hormone (ACTH) responses to CRF (n = 15), whereas patients who did not show remission remained hyperresponsive (n = 10). In patients with autonomous adrenal cortisol secretion, the ACTH response to CRF was suppressed (n = 10). After surgery the ACTH response to CRF can already be demonstrated when cortisol levels are still undetectable. Pulsatile administration of CRF in one patient after unilateral adrenalectomy and another patient under corticoid therapy revealed that ACTH responses to CRF normalize rapidly but cannot be sustained if CRF administration is withdrawn, suggesting that the cause of adrenal failure after unilateral adrenalectomy for Cushing's syndrome or long-term corticoid therapy is due to hypothalamic CRF deficiency. The decrease of the ACTH responses to CRF in glucocorticoid-treated patients correlated directly to the daily corticoid dosage. Since the ACTH hyperresponse to CRF in 6 patients with Cushing's disease was also suppressed by short-term dexamethasone treatment, the pituitary level as target site for the acute feedback inhibition is also demonstrated. The evaluation of the CRF-induced ACTH response in patients with secondary adrenal failure without detectable pathology in the sella and suprasellar region (n = 6) enables the differentiation between hypothalamic and pituitary adrenal insufficiency. In patients with hypothalamic lesions the ACTH response to CRF was normal whereas insulin hypoglycemia failed to induce an ACTH rise.