Four hundred and ninety-seven postmenopausal women with hormone-receptor-positive breast cancer were randomized to receive adjuvant exemestane 25 mg or anastrozole 1 mg daily for 5 years. Two subgroups were identified: those with baseline T-scores of -2.0 or greater at the spine and hip (subgroup 1), and those with baseline T-scores less than -2.0 at the spine and hip and taking bisphosphonates (subgroup 2). Results at 2 years revealed no significant differences in change in bone mineral density between exemestane and anastrozole in either subgroup. Furthermore, regardless of the subgroup, there were no significant differences between groups with respect to the incidence of fracture and changes in serum bone turnover markers.