To compare the effects of sodium fluoride and etidronate in severe postmenopausal osteoporosis, we conducted a 3 year, prospective, trial in 118 postmenopausal osteoporotic women with at least one vertebral fracture, who were randomly assigned to receive sodium fluoride (25 mg twice daily, as enteric-coated tablets) plus calcium (1000 mg/day) or intermittent etidronate (400 mg/day for 14 days) followed by calcium (1000 mg/day for 76 days). Lateral spine X-ray films and dual-energy X-ray absorptiometry (DXA) measurements of the lumbar spine and proximal femur were performed at enrollment and yearly. Nonvertebral fractures were recorded every 6 months. Thirty-one women in the fluoride group and 47 in the etidronate group completed the trial. At 36 months, the mean change from baseline of the lumbar bone density in the fluoride group was 8.5 +/- 2.04% (p = 0.001) and in the etidronate group was of 3.6 +/- 0. 84% (p < 0.001). The changes in the fluoride group were significantly higher than in the etidronate group (p = 0.01). Both groups showed nonsignificant changes in femoral neck bone density. There was no significant difference between groups in the cumulative proportion of women with new vertebral fractures, with an incidence in the fluoride group of 16% vs. 17% in the etidronate group. However, the number of new vertebral fractures was significantly lower in the fluoride group (6 fractures) than in the etidronate group (19 fractures) (p = 0.05). The number of patients with nonvertebral fractures was similar in both groups. A high incidence of side effects, mainly gastrointestinal symptoms and lower extremity pain syndrome, was observed in the fluoride group. Etidronate was well tolerated. We conclude that, in women with severe osteoporosis, although sodium fluoride is more favorable than cyclical etidronate for increasing lumbar bone mass, no differences were observed in the incidence of fractures.