A number of studies have shown that an excess of glucocorticoids induces osteoporosis, but the mechanism(s) and the time course of the reduction of bone mass remain uncertain. In order to clarify this issue we carried out a longitudinal clinical and histomorphometric study of patients requiring long-term glucocorticoid treatment. In 23 patients (9 men, 10 post- and 4 premenopausal women) biochemical and bone histomorphometric investigations were carried out before and during treatment with 10-25 mg/day of prednisone. Histomorphometric analysis of bone biopsies of the iliac crest showed that the decrease of TBV (up to -27%, P less than 0.001) occurs predominantly within the first 5-7 months of treatment; during the subsequent stages, which include observations after 12 months of treatment, only minor changes were observed. Therefore trabecular bone loss can be satisfactorily described by a negative exponential function. None of the other histomorphometric parameters (osteoid surfaces, resorption surfaces, etc.) showed significant changes. However, the histological features of the bone biopsies during steroid therapy, showing a virtual lack of osteoblastic activity, ruled out an increase of bone resorption. Moreover, the dynamic study of the bone formation by double tetracycline labelling showed, in a small subgroup of patients, a decrease of the apposition rates (from 0.763 +/- 0.053 to 0.305 +/- 0.074 microns/day (mean +/- SE) after treatment). No significant changes, at any time during steroid treatment, were observed in serum alkaline phosphatase, 25-hydroxyvitamin D, 24,25-dihydroxyvitamin D, 1,25-dihydroxyvitamin D, parathyroid hormone or urinary calcium excretion. Serum calcium increased significantly within the first 1-2 months of therapy and then it returned to baseline. Urinary hydroxyproline excretion decreased significantly within the first 1-2 months and continued to fall throughout the treatment. Thus, both biochemical and histological findings suggest that long-term glucocorticoid therapy causes a reduction of bone turnover, that the bone loss occurs predominantly within the first 6 months of treatment and that patients with lower bone mass have a lower rate of bone loss.