Bone loss is more severe in primary adrenal than in pituitary-dependent Cushing's syndrome.

Either exogenous or endogenous glucocorticoid excess is an established cause of osteoporosis and fractures. Glucocorticoids exert their negative effects on bone through mechanisms that are not yet completely elucidated; however, as many as 50% of patients with Cushing's syndrome suffer from osteoporosis. Bone loss induced by glucocorticoids is potentially reversible after resolution of glucocorticoid excess. It is presently unknown if Cushing's disease (CD) sustained by a pituitary ACTH-producing adenoma and adrenal-dependent Cushing's syndrome (ACS) sustained by an adrenocortical adenoma have a different potential of inducing osteopenia. The aim of the present study was to retrospectively analyze bone mineral density (BMD) in 26 patients with CD (4 men, 22 women, aged 14-79 years), 12 patients with ACS (4 men, 8 women, aged 32-79 years) and 38 healthy subjects carefully matched for sex, age and body mass index (BMI). Measurement of BMD was performed by dual-energy X-ray absorptiometry (DXA) using the Hologic QDR 4500 W instrument. Data were analyzed using absolute BMD values (g/cm2), T-score and Z-score referred to the manufacturer's normative data for the lumbar spine and to the NHANES III dataset for the hip. The patients with CD and ACS were comparable for age, BMI, estimated duration of disease, urinary free cortisol (UFC) levels, midnight serum cortisol and gonadal function. The analysis of variance demonstrated that lumbar bone densitometric parameters were significantly different among the three groups. They were more reduced in patients with ACS (BMD, 0.76+/-0.03 g/cm2; T-score, -2.78+/-0.28; Z-score, -2.25+/-0.30) while patients with CD (BMD, 0.87+/-0.02 g/cm2; T-score, -1.74+/-0.24; Z-score, -0.99+/-0.32) showed DXA values between the first group and controls (BMD, 1.02+/-0.02 g/cm2; T-score, -0.35+/-0.19; Z-score, 0.33+/-0.16). The difference in BMD at the spine remained statistically significant ( P=0.04) after adjustment for the non-significant differences in age, UFC and fat mass between CD and ACS. Conversely, femoral bone densitometric parameters were not significantly different between patients with ACS and CD, even if they were more reduced than in controls. In patients with ACS, we observed a reduction of DHEA-S levels, expressed as standard score ( Z-score) values referred to a group of 180 healthy subjects stratified by sex and different age groups (<40 years, between 40 and 60 years, >60 years) to circumvent the pronounced effect of gender and age on such hormone (ACS DHEA-S Z-score -0.88+/-1.4 versus CD DHEA-S Z-score 2.25+/-2.35, P=0.0001). DHEA-S Z-score values were significantly correlated with lumbar BMD ( r=0.41, P=0.02) and femoral BMD ( r=0.43, P=0.01). DHEA-S Z-score values were also significantly correlated with osteocalcin levels ( r=0.45, P=0.01). Our data suggest that bone loss is greater in ACS than in CD. A plausible explanation comes from the reduced DHEA-S level in ACS since DHEA-S has well known anabolic actions on bone. However, this hypothesis needs to be confirmed in large, prospective series of patients with Cushing's syndrome of different etiology.