To study the effect of fluoride on bone mineral density (BMD) in patients treated chronically with glucocorticosteroids, 15 subjects (renal grafted, n = 12; skin disease, n = 1; broncho pulmonary disorder, n = 1; Crohn's disease, n = 1) were prospectively studied in a double-blinded manner and randomly allocated either to group 1 (n = 8) receiving 13.2 mg/day fluoride given as disodium monofluorophosphate (MFP) supplemented with calcium (1,000 mg/day) and 25-hydroxyvitamin D (calcifediol) (50 micrograms/day), or to group 2 (n = 7) receiving Cas+ calcifediol alone. An additional group of 14 renal transplant patients treated chronically with glucocorticosteroids but exempt of specific therapeutic intervention for bone disease was set up as historical controls. BMD was measured by dual-energy X-ray absorptiometry (DXA, Hologic QDR 1000) performed at months 0, 6 and 12 for groups 1 and 2 (lumbar spine, total upper femur, diaphysis and epiphysis of distal tibia), or 11-31 months apart with calculation of linear yearly changes for the historical cohort. Lumbar BMD tended to rise in groups 1 and 2, and to fall in group 3, the change reaching statistical significance (p < 0.05) in group 1, thus leading to a significant difference between groups 1 and 3 (p < 0.05). At upper femur, tibial diaphysis and tibial epiphysis, no significant change in BMD occurred in any of the groups. In conclusion, lumbar BMD rises more after a mild dosis of fluoride given as MFP and combined to calcium and calcifediol than on Ca+ calcifediol alone, without changes in BMD at the upper femur or distal tibia.