A histomorphometric long-term longitudinal study of trabecular bone loss in glucocorticoid-treated patients: prednisone versus deflazacort.

Administration of a corticosteroid with minor osteopenic effects is considered an effective prevention of glucocorticoid osteoporosis. Deflazacort, an oxazolinic derivative of prednisolone, is reported to be less harmful to cancellous bone mass than other equally effective corticosteroids. However, comparative long-term studies, particularly on trabecular bone, are needed before a smaller detrimental effect on bone of deflazacort can be unequivocally confirmed. We conducted such a prospective long-term study using histomorphometric analysis of iliac bone. For the study, 18 pairs of nonimmobilized patients, matched for age, sex, menopausal state, corticosteroid dose, and type and severity of the disease, were randomly submitted to treatment with therapeutically equivalent doses of prednisone or deflazacort. Bone biopsies from iliac crest were taken before and at various times during treatment. In order to represent the time-related trabecular bone loss and find out possible differences between patients on prednisone or deflazacort, a previously described model of bone loss kinetics was applied. No significant differences in biochemical indices of bone turnover or in histomorphometric variables between prednisone- and deflazacort-treated patients were recorded before treatment. The mean duration of treatment at the final biopsy was similar for prednisone and deflazacort (15.8 and 15.2 months, respectively). Patients showed evident clinical improvement with both treatments. Osteoid and resorption surfaces showed no significant differences throughout the observation period in any of the 18 pairs. On the contrary, both steroids induced a significant decrease in trabecular bone, although the bone loss rate induced by prednisone was significantly higher than that induced by deflazacort (P < 0.05). The kinetics of bone loss and the different osteopenic effects of the two drugs are described by the negative exponential function fitted to data from patients never previously given glucocorticoids; the model yields highly significant difference (P approximately equal to 0.01) between the two drugs and allows estimation of the difference even 3 years after the beginning of treatment (-3.0%/year versus -1.1%/year for prednisone and deflazacort, respectively). This prospective long-term study confirms that an exponential model accurately describes the trabecular bone loss induced by long-term corticosteroid treatment and demonstrates that deflazacort, at therapeutically effective doses, induces less trabecular bone loss than prednisone.