The benefit of sodium fluoride (NaF) in the therapy of osteoporosis is still controversial. For 3 years we monitored patients with postmenopausal osteoporosis subjected to a continuous treatment with 80 mg NaF/day and patients without fluoride treatment. Every 3 months peripheral total and trabecular bone densities were evaluated with high-precision low-dose quantitative computed tomography, every 6 months biochemical parameters were measured, and every year new crush fractures were determined. The untreated osteoporotics as a group lost bone at a rate of 2.5%/year. In the fluoride-treated group trabecular bone density of the distal tibia remained unchanged in 5 of 15 treated patients; 2 patients experienced a bone loss, 8 patients showed at least a temporary bone gain. After 3 years trabecular bone density of the treated patients was 8% higher than that of the untreated patients. Total bone density was not increased. The fracture rates in a group of untreated and a group of NaF-treated patients matched with regard to age, height, weight, initial fracture rate, and initial trabecular bone density were significantly different in the first year, with 0.3 new fractures in the untreated group and 2.9 new fractures in the treated group. During the second and third year the mean number of new fractures was approximately equal in both groups. In 47% of the treated patients, osteoarticular side effects were observed. In 27%, scintigraphy of the ankle was positive, alkaline phosphatase was increased, and radiologic signs of healing stress fractures were present. Long-term treatment with NaF has a positive effect on trabecular bone density. However, the incidence of osteoarticular side effects and the increased number of new crush fractures of the spine during the first year of treatment raise the possibility of fluoride-induced microfractures.