Android obesity is associated with increased cortisol secretion. Direct effects of cortisol on gluconeogenesis and other parameters of insulin resistance were determined in normal subjects. Gluconeogenesis was determined using the reciprocal pool model of Haymond and Sunehag (HS method), and by the Cori cycle/lactate dilution method of Tayek and Katz (TK method). Glucose production (GP) and gluconeogenesis were measured after a 3 h baseline infusion and after a 4-8 h pituitary-pancreatic infusion of somatostatin, replacement insulin, growth hormone (GH), glucagon and a high dose of cortisol (hydrocortisone). The pituitary-pancreatic infusion maintains insulin, GH and glucagon concentrations within the fasting range, while increasing the concentration of only one hormone, cortisol. Two groups of five subjects were each given high-dose cortisol administration, and results were compared with those from a group of six 'fasting alone' subjects (no infusion) at 16 and 20 h of fasting. Fasting GP (12 h fasting) was similar in all groups, averaging 12.5+/-0.2 micromol x min(-1) x kg(-1). Gluconeogenesis, as a percentage of GP, was 35+/-2% using the HS method and 40+/-2% using the TK method. After 16 h of fasting, GP had fallen (11.5+/-0.6 micromol x min(-1) x kg(-1)) and gluconeogenesis had increased (55+/-5% and 57+/-5% of GP by the HS and TK methods respectively; P<0.05). High-dose cortisol infusion for 4 h increased serum cortisol (660+/-30 nmol/l; P<0.05), blood glucose (7.9+/-0.5 mmol/l; P<0.05) and GP (14.8+/-0.8 micromol x min(-1) x kg(-1); P<0.05). The increase in GP was due entirely to an increase in gluconeogenesis, determined by either the HS or the TK method (66+/-6% and 65+/-5% of GP respectively; P<0.05). Thus cortisol administration in humans increases GP by stimulating gluconeogenesis. Smaller increases in serum cortisol may contribute to the abnormal glucose metabolism known to occur in the metabolic syndrome.