In this placebo-controlled double-blind study, psychological and endocrine stress responses were investigated in healthy postmenopausal placebo-treated women (n = 15; 60-75 years; placebo via transdermal patches), healthy postmenopausal estradiol-treated women (n = 13; 60-79 years; 0.1 mg 17beta-estradiol daily via transdermal patches) and young controls (n = 15; 20-31 years; untreated). The aged subjects received estradiol or placebo treatment for 14 days. All subjects were then exposed to the 'Trier Social Stress Test' (TSST) and the dexamethasone (Dex)-human corticotropin-releasing hormone (hCRH) test (100 microgram hCRH after premedication with 1.5 mg Dex). Psychological parameters including perceived stressfulness, mood and subjective well-being were measured by visual analog scales, a mood questionnaire and a mood diary, respectively. Results show that the TSST induced significant increases in adrenocorticotropin (ACTH), free salivary cortisol, total plasma cortisol and heart rates (all p < 0.0001). Regardless of age, comparable hormonal response patterns were observed in the TSST as indicated by similar peak levels and recovery phases. Visual analog scales confirmed that the same amount of stress was experienced by young and elderly subjects. In both age groups, hCRH injection after Dex premedication provoked significant increases in ACTH, free salivary cortisol and total plasma cortisol (all p < 0.0001). In contrast to the psychosocial stressor, elderly women were found to respond with a markedly enhanced cortisol response compared to young controls in the Dex-CRH test (p < 0.025). Additional investigation of morning cortisol profiles could not reveal any age-related differences in basal hypothalamus-pituitary-adrenal (HPA) axis activity. Following estradiol treatment, estradiol levels significantly increased only in substituted postmenopausal women (p < 0.001) reaching concentrations typically found in younger women during the follicular phase of the menstrual cycle. Corticosteroid-binding globulin levels did not differ significantly between groups. When confronted with the TSST, no response differences emerged between the three groups. However, estradiol treatment appeared to blunt the total plasma cortisol response in the Dex-CRH test, resulting in smaller increases in untreated premenopausal women and estradiol-treated postmenopausal women compared to placebo-treated postmenopausal women (p < 0.02). In sum, no response differences were observed after confrontation with a psychosocial stress test in our sample of healthy elderly subjects. As shown with the Dex-CRH test, our data suggest that the negative feedback of the HPA axis in elderly women is altered. Moreover, the current data suggest that estradiol replacement may modulate HPA feedback sensitivity in humans.