Adrenal function was estimated by synthetic ACTH test in 18 men with AIDS (or AIDS-related complex) and compared with that of 10 HIV positive but otherwise healthy men. According to this test none fulfilled the criteria for adrenal insufficiency defined as plasma cortisol concentration less than 500 nmol/l 30 minutes after ACTH stimulation. Seven of eighteen AIDS patients had baseline cortisol concentrations above 0.5 mumol/l compared to none in the HIV-positive group. Three of eighteen AIDS patients had a limited response to synthetic ACTH injection compared to none in the HIV-positive group. Two of these three AIDS patients had lowered serum Na+ concentration; they survived for two and three weeks, respectively. Otherwise the basal cortisol level and response to synthetic ACTH was uncorrelated with survival time, other signs of adrenal insufficiency, treatment with ketokonazole, CMV-infection, T-helper cell count or Th/Ts-ratio. The pituitary-adrenal axis was estimated by measuring the diurnal rhythm of serum ACTH and cortisol in eight patients and was found intact and normal in all of them. Thus, absolute adrenal insufficiency is uncommon in AIDS-patients. Relative adrenal insufficiency may occur in severely ill preterminal patients as a result of primary target organ failure, but the pituitary-adrenal axis generally appears to be undisturbed.