Methods. In a randomized, double-blind, placebo-controlled clinical trial, we evaluated the effect of a 2-year treatment with nandrolone decanoate (ND) on bone mineral density (BMD) of lumbar spine, femoral neck, and trochanter and on vertebral fracture rate, muscle mass, and hemoglobin levels. Sixty-five osteoporotic women older than 70 years were studied. Thirty-two patients received injections of 50 mg ND, and 33 received placebos every 3 weeks. All patients received 500 mg calcium tablets daily. Results. Compared to baseline, ND increased the BMD of the lumbar spine (3.4% ± 6.0 and 3.7% ± 7.4; p <.05) and femoral neck (4.1% ± 7.3 and 4.7% ± 8.0; p <.05) after 1 and 2 years, respectively. The BMD of trochanter increased significantly only after the first year (4.8% ± 9.3, p <.05). Compared to the placebo group, the ND group presented with significantly increased BMD of the trochanter and neck. ND significantly reduced incidence of new vertebral fractures (21% vs 43% in the placebo group; p <.05). ND showed a significant statistical increase in lean body mass after the first (6.2% ± 5.8; p <.01) and second years (11.9% ± 29.2; p <.01). In addition, a 2-year treatment with ND significantly increased hemoglobin levels compared to baseline (14.3%; p <.01) and placebo (p <.01). Conclusions. ND increased BMD, hemoglobin levels, and muscle mass, and reduced the vertebral fracture rate of elderly osteoporotic women.