Determinants of serum leptin levels in Cushing's syndrome

Corticosteroids and insulin increase leptin expression in vivo and in vitro. To investigate whether increased serum cortisol influences serum leptin concentrations in humans, we analyzed fasting serum leptin and insulin levels in 50 patients with Cushing's syndrome [34 female patients: 27 with the pituitary form and 7 with the adrenal form; age, 41.6 +/- 2.7 yr; body mass index (BMI), 29.6 +/- 1.2 kg/m(2); 16 male patients all with the pituitary form; age, 39.2 +/- 3.1 yr; BMI, 26.3 +/- 2.3 kg/m(2)] and in controls matched for BMI, age, and gender. Serum leptin levels were higher in female than in male patients in both the Cushing (P < 0.01) and control (P < 0.001) groups. Disease-specific differences in serum leptin levels were only detected in male (106 vs. 67 pmol/L; Cushing's syndrome vs control, P < 0.05), not female, patients. Multiple stepwise regression analysis of both patient groups revealed insulin as the best predictor of serum leptin concentrations, accounting for 37% of the variance in serum leptin levels, in contrast to BMI or mean serum cortisol (as measured by sampling in 10-min intervals over 24 h). In the subgroup of patients (n = 9) with pituitary adenoma, serum leptin levels were reduced after tumor resection, with concurrent decreases in serum cortisol, insulin, and BMI. In conclusion, chronic hypercortisolemia in Cushing's syndrome appears not to directly affect serum leptin concentrations, but to have an indirect effect via the associated hyperinsulinemia and/or impaired insulin sensitivity.