Seventy-eight patients with cytologically and/or histologically confirmed prostatic cancer were randomly allocated to orchidectomy (ORX, n = 37) or combined intramuscular and oral estrogen treatment (ESTR, n = 41). Serum levels of testosterone (T), 17 alpha-hydroxyprogesterone, dehydroepiandrosterone, dehydroepiandrosterone sulfate, total estrone (tE1; sum of unconjugated and conjugated estrone, greater than or equal to 85% estrone sulfate), cortisol, luteinizing hormone, follicle-stimulating hormone, prolactin, growth hormone, sex hormone-binding globulin (SHBG), and albumin were determined prior to treatment and 12, 24, and 36 months after initiation of treatment. Fifty patients responded to treatment or had stable disease, and 28 did not respond (12 in the ORX and 16 in the ESTR group). There was no association between pretreatment hormone or protein levels and outcome of the treatment, neither in the total material nor within either of the two treatment subgroups. Significantly higher pretreatment levels of cortisol and prolactin and significantly lower levels of T, tE1, and albumin and a significantly lower T/SHBG-ratio (index on biologically active T) were found in patients with metastatic disease, compared with the patients without metastases. There was no association between testicular or adrenal androgens, SHBG, T/SHBG, and albumin values during treatment and the clinical outcome. The differences found between metastatic and nonmetastatic disease probably simply reflect the more stressful and catabolic condition and generally poorer health in patients with disseminated malignant disease. Furthermore, the study does not lend any support to the hypothesis that indicates an important role of adrenal "rest androgen" in prostatic cancer tumor growth.