In alcoholics, disturbances of the autonomic nervous system as well as of the hypothalamic-pituitary-adrenal axis (HPA) are known. However, these two systems have never been analyzed, under stimulated conditions, in parallel in the same patients. Moreover, studies using intravenous (i.v.) corticotropin releasing factor (CRF) to assess neuroendocrine function bypass the hypothalamic component of the HPA axis. Therefore, i.v. human (h) CRF (pituitary stimulation/exogenous CRF) and a multifaceted stress test (hypothalamic activation/endogenous CRF) were compared with respect to their effects on hemodynamics as well as plasma norepinephrine (NE), epinephrine (E), ACTH, and cortisol in abstinent alcoholics (n = 11) versus healthy men (n = 10). Each stimulus was tested twice, 12 weeks apart, in two separate experimental blocks (I and II). Alcoholics entered block 18 days after the last ethanol ingestion and were controlled for abstinence up to block II. hCRF caused a fall in mean arterial pressure (MAP), most pronounced in alcoholics, particularly in block II. In contrast, stress testing raised MAP in both groups and blocks. A sustained increase in ACTH, cortisol, and NE occurred after hCRF, although the ACTH response in alcoholics was blunted in both blocks. Stress testing elevated NE in both groups and blocks, while raising plasma ACTH and cortisol during block I only in controls. However, unlike the persistently blunted ACTH response to i.v. CRF, a normalization of the stress-induced ACTH output occurred in alcoholics after 12 weeks of abstinence. During block I, basal E levels were elevated in alcoholics whereas NE levels tended to be lower than in controls, resulting in a significantly decreased NE/E ratio that returned to near control values in block II. Neither CRF nor stress had any effect on circulating E in either group or block. To conclude: (1) Normalization of the ACTH response to stress, but not to i.v. CRF, after 12 weeks of abstinence, suggests that other ACTH secretagogues may be compensating for CRF dysfunction in alcoholics. (2) Despite the dramatically lowered plasma NE/E ratio in alcoholics, the NE response to stimuli was unaffected. (3) The exaggerated hypotensive reaction and blunted ACTH response to i.v. CRF may reveal a long-term dissociative dysregulation of CRF actions in alcoholics.