A cohort of 101 patients were treated with enteric-coated sodium fluoride tablets and calcium supplements. Vitamin D was also given in supra-physiologic doses in 70% of the cases. Lumbar bone mineral density (BMD), as measured by dual-photon absorptiometry, increased in a linear fashion up to four years, irrespective of the value of initial BMD and of the underlying condition, be it involutional osteoporosis (the vast majority), glucocorticoid osteoporosis, or even osteogenesis imperfecta. Estrogen replacement therapy (ERT) seemed to promote the fluoride-induced increase in lumbar BMD, as did the vitamin D supplements. Of these patients, 17% proved "resistant" to the therapy. There was no way of predicting who would be in this category. Compared with an age- and sex-matched control group, women showed significantly different behavior of their bone mass. In the control group, the losses were highly significant at the lumbar spine and at all three scanning sites of the forearm, as measured by single-photon absorptiometry. In contrast, the fluoride group had a significant gain of BMD at the lumbar spine and changes of BMC at the forearm were not significant. Fluoride thus preserved bone mass at the appendicular skeleton, while increasing it at the axial skeleton. When comparing the patients who received vitamin D supplements and those who did not, there was a significant difference in the appendicular skeleton. The distal forearm in the vitamin D-supplemented group tended to gain, whereas the midforearm lost significant bone mass. The trend was reversed in the group without vitamin D-supplementation, a more favorable pattern. Therefore, vitamin D supplements should not, as a rule, be provided to such patients. The biochemical hallmark of the fluoride-induced changes is a slight rise of the alkaline phosphatase within the normal range. Alkaline phosphatase levels that exceed the upper limit of normal signal a warning that too much fluoride and/or too little calcium supplements are being administered, or that a fluoride-related complication is impending or has occurred (e.g., a stress fracture). Osteosclerosis was achieved in 69% of the cases who had a radiological followup of at least four years (average period of appearance: 1.8 years). Stress fractures in the lower limbs occurred in 17 patients, almost exclusively in females, and appeared on average 2.2 years after initiation of therapy. In this group of stress fractures there was significant cortical bone loss at midforearm.(ABSTRACT TRUNCATED AT 400 WORDS)