Twenty-six post-menopausal osteoporosis patients were studied prospectively over a 2-yr period after the cessation of 2 yr of treatment with either nandrolone decanoate (anabolic steroid), 1 alpha hydroxyvitamin D3 or intermittent calcium infusions, the total observation period being 4 yr. Serial evaluations of bone mineral content in the radius, cortical thickness of the metacarpals and spinal roentgenograms were obtained. Nine patients who had been treated with 50 mg nandrolone decanoate every three weeks did not lose bone from the radius or the metacarpals during the 2 yrs following withdrawal of therapy. Eight patients who had received 1 microgram/day of oral 1 alpha hydroxyvitamin D3 retained a positive bone mineral content in the radius but lost cortical thickness from the metacarpals after withdrawal of therapy. Nine patients who had been treated with intermittent calcium infusions over 12 consecutive days annually did not lose further bone mineral content from the radius, but did lose cortical thickness at the expected rate during the 2-yr post-treatment period. The fracture rate at the end of the 4-yr observation period was 40% lower in the nandrolone decanoate group than in the 1-alpha-hydroxyvitamin D3 and calcium infusion groups. The results of this study indicate that, in contrast to oestrogen and calcium therapy, nandrolone decanoate and 1-alpha-hydroxyvitamin D3 have lasting beneficial effects, and that data obtained by photon absorptiometry do not always coincide with those obtained by radiogrammetry. These findings are highly relevant to the design of future therapeutic trials.