Endogenous hypercortisolism and high-dose and long-term glucocorticoid (GC) therapy reduce bone mass. Patients in remission after successful treatment of Cushing's syndrome (CS) often present hypoadrenalism and require long-term GC replacement. The aim of our study was to evaluate whether this GC "replacement" had any further effect on bone in women after long-term remission of CS. Thirty-seven women (mean age: 50 +/- 14 yr; 27 of pituitary and 10 of adrenal origin) with cured CS (mean time of cure: 11 +/- 6 yr), 14 with active CS, and 85 sex-, body mass index (BMI)-, and age-matched controls were enrolled. BMD and BMC were measured by DXA scanning. Bone biochemical markers were also measured. Duration and dose of GC replacement and duration of endogenous hypercortisolism were calculated. Cured and active CS patients had less BMC, BMD, and osteocalcin than controls (p < 0.01). These differences were observed in estrogen-sufficient women but not in those with estrogen deficiency. Duration of GC treatment (mean: 42 mo; range, 2-420 mo) and endogenous hypercortisolism (mean: 70 mo; range, 13-241 mo) negatively correlated with BMC and lumbar spine BMD. After regression analysis, the main predictor of abnormal BMC and BMD was the duration of GC replacement (p < 0.01). Patients treated for CS persistently have less bone mass despite long-term cure. Both duration of endogenous hypercortisolism and mainly exogenous "replacement" therapy with GC negatively affect bone mass. Thus, the additional deleterious effect of GC for the treatment of adrenal axis suppression should be considered.