The aim of this randomised study was to compare the effects of progestins and aromatase inactivators on bone remodelling markers and the components of insulin-like growth factor in patients with metastatic breast cancer. Within the framework of a large (769 patients), randomised double-blind clinical trial comparing exemestane (EXE) with megestrol acetate (MA), serum 17 beta-estradiol (E2), estrone (E1), estrone sulphate (E1S), bone alkaline phosphatase (BAP), carboxy-terminal cross-linking telopeptide of type I collagen (ICTP) and the components of insulin-like growth factor (IGF) family (IGF-1, IGF-2 and IGFBP-3) were determined in 53 patients (24 randomised to EXE and 29 ramdomised to MA). After eight weeks of treatment, both ICTP and BAP increased (p < 0.01) in the EXE group, but only ICTP in the MA group (p < 0.03). The 8-week suppression of E2 and E1S was more pronounced in the EXE group (to, respectively, 11.2% and 9.9% of baseline values) than in the MA group (33.1% and 29.7%). IGF-1 increased (p < 0.01) in both groups, but more so in the patients treated with MA. Estrogen levels negatively correlated with ICTP in both groups, but were not related to BAP in either. IGF-1 negatively correlated with estrogens in both groups. The results of this study indicate that anti-aromatase therapy is associated with increased osteoclast activity, and suggest the existence of possible differential effects of different hormonal therapies on bone remodelling markers regardless of the estrogen suppression induced by EXE.