A total of 33 women with postmenopausal osteoporosis were matched pairwise by age, years since menopause, and body mass index and randomized to receive either cyclic estrogen-progestagen replacement treatment (group 1) or the same treatment plus nandrolone decanoate (ND; group 2). Both groups were treated during 3 years and subsequently followed for another year off treatment. A year after cessation of the treatment the distal forearm bone mineral content in group 2 was significantly higher than that in group 1. Bone mass measurements in the axial skeleton already showed a significant difference in favor of group 2 after 3 years treatment, which persisted during the year off treatment. The decline in lumbar bone mineral mass and density in the 1 year off treatment was similar in both groups. Correction for body mass did not change these results. Bone turnover parameters did not show significant differences between the two groups after cessation of treatment. A higher muscle mass, induced by ND, could partly explain the differences between the groups since even 1 year after treatment was stopped an increased serum creatinine level was still observed in group 2.