In Cushing's syndrome, cortisol causes fat accumulation in specific sites most likely to be associated with insulin resistance, notably in omental adipose and also perhaps in the liver. In idiopathic obesity, cortisol-metabolizing enzymes may play a key role in determining body fat distribution. Increased regeneration of cortisol from cortisone within adipose by 11beta-hydroxysteroid dehydrogenase (HSD) type 1 (11HSD1) has been proposed to cause visceral fat accumulation, whereas decreased hepatic 11HSD1 may protect the liver from glucocorticoid excess. Increased inactivation of cortisol by 5alpha- and 5beta-reductases in the liver may drive compensatory activation of the hypothalamic-pituitary-adrenal axis, hence increasing adrenal androgens and 'android' central obesity. This study aimed to examine relationships between these enzymes and detailed measurements of body fat distribution. Twenty-five healthy men (age, 22-57 yr; body mass index, 20.6-35.6 kg/m(2)) were recruited from occupational health services. Body composition was assessed by anthropometric measurements, bioimpedance, and cross-sectional abdominal magnetic resonance imaging scans. Liver fat content was assessed by magnetic resonance imaging spectroscopy. Insulin sensitivity was measured in a euglycemic hyperinsulinemic clamp. Cortisol metabolites were measured in a 24-h urine sample by gas chromatography-mass spectrometry. In vivo hepatic 11HSD1 activity was measured by generation of plasma cortisol after an oral dose of cortisone. In vitro 11HSD1 activity and mRNA were measured in 18 subjects who consented to provide abdominal sc adipose biopsies. Indices of obesity (body mass index, whole-body percentage fat, waist/hip ratio) were associated with higher urinary excretion of 5alpha- and 5beta-reduced cortisol metabolites (for percentage fat, P < 0.05 and P < 0.01, respectively) and increased adipose 11HSD1 activity (P < 0.05). Liver fat accumulation was associated with a selective increase in urinary excretion of 5beta-reduced cortisol and cortisone metabolites (P < 0.01) and a lower ratio of cortisol/cortisone metabolites in urine (P < 0.001) but no difference in in vivo cortisone-to-cortisol conversion or in vitro adipose 11HSD1. Higher excretion of 5beta-reduced cortisol metabolites was independently associated with insulin resistance and hypertriglyceridemia. Lower conversion of cortisone to cortisol was associated with lower fasting plasma cortisol (P < 0.01). However, visceral adipose fat mass was not associated with indices of cortisol metabolism; indeed, after adjusting for the effects of whole-body and liver fat, increased visceral fat was associated with lower cortisol metabolite excretion. We conclude that alterations in 11HSD1 and hepatic 5alpha-reductase activity are associated with generalized, rather than central, obesity in humans. Activation of 5beta-reductase in men with fat accumulation in the liver may confound the interpretation of cortisol metabolite excretion when liver fat content is unknown, and may contribute to altered bile acid and cholesterol metabolism in nonalcoholic steatohepatitis.