OBJECTIVE To test whether slow-release sodium fluoride inhibits spinal fractures and is safe to use. METHODS Placebo-controlled randomized trial. METHODS Slow-release sodium fluoride, 25 mg twice daily, in four 14-month cycles (12 months receiving sodium fluoride followed by 2 months not receiving it) compared with placebo. Calcium citrate, 400 mg calcium twice daily, continuously in both groups. METHODS 48 of 54 patients who received sodium fluoride and 51 of 56 patients who received placebo completed at least 1 year of the study. All patients had postmenopausal osteoporosis. RESULTS Compared with the placebo group, the fluoride group had a lower individual vertebral fracture rate (0.064 +/- 0.182 per patient-year compared with 0.205 +/- 0.297 per patient-year; P = 0.002), a higher unadjusted fracture-free rate (85.4% compared with 56.9%; P = 0.001), and a greater survival estimate (relative risk, 0.3 [95% CI, 0.12 to 0.76]) for new fractures. The recurrent spinal fracture rate did not differ between the two groups. The fluoride group had a substantial increase in L2-L4 bone mass of 4% to 5% per year for 4 years, a mean increase in femoral neck bone density of 2.38% +/- 3.33% per year, and no change in radial shaft bone density. The frequency with which minor side effects and appendicular fractures occurred was similar in the two groups; no patients developed microfractures or gastric ulcers. CONCLUSIONS Slow-release sodium fluoride and calcium citrate administered for 4 years inhibits new vertebral fractures (but not recurrent fractures), augments spinal and femoral neck bone mass, and is safe to use.