The most important determinant of fracture risk is bone mineral density. Fracture risk approximately doubles for each standard deviation by which this density is less than peak adult bone mass at age 30 to 35 years [3]. Accurate, precise, noninvasive methods for measuring bone mass have been available for almost three decades. Yet, osteoporosis is still most often diagnosed only after the first fragility fracture has occurred. Most authorities regard dual-energy x-ray absorptiometry as the "gold standard" for bone mass measurement. There are fewer than 1500 dual-energy x-ray absorptiometry instruments in the United States, and many of them are clustered in major academic centers. Even at full capacity, each instrument can accommodate no more than 10 to 15 persons per day. As a relatively expensive ($75 000 to $100 000) single-use instrument with limited reimbursement, prospects for wider deployment of dual-energy x-ray absorptiometry are not strong. Several alternative methods for bone mass measurement, including single-energy x-ray absorptiometry and broadband ultrasound attenuation of the calcaneus [4] and radiographic absorptiometry of the phalanges [5], are being tested extensively. One or more of these may emerge for wider use before too long.

There is no joy in establishing an individual woman's risk for an osteoporotic fracture if nothing can be done to prevent that fracture. It is equally frustrating to detect a new osteoporotic vertebral fracture when limited therapeutic options are available for preventing the next one. This first fracture is unfortunately an excellent predictor of subsequent fractures, independent of bone mass [6]. For the treatment of osteoporosis, the Food and Drug Administration (FDA) has only approved estrogen—which many physicians are still reluctant to prescribe for older women, even those with osteoporotic fractures—and calcitonin, which is expensive and can only be administered by subcutaneous injection. An FDA advisory panel has recently recommended the approval of nasal spray calcitonin, and alendronate, a third-generation aminobisphosphonate. These could be available to clinicians by the end of the year. Etidronate has been widely prescribed to treat osteoporosis since two articles reported the antifracture efficacy of this drug [7, 8]. The wide-spread use of etidronate continues, even though an FDA advisory panel has twice (most recently in November 1994) recommended that the drug not be approved for the treatment of osteoporosis because its efficacy has not been well established. This clearly reflects the desire of physicians and patients for a safe, oral therapy for osteoporosis.

In this issue, Pak and colleagues [9] provide the final report from their clinical trial of a slow-release oral sodium fluoride preparation. A preliminary report with essentially similar findings was published last year [10]. Ninety-nine women with postmenopausal osteoporosis received slow-release sodium fluoride or placebo for an average of 3.5 to 4 years. New vertebral fractures occurred significantly less frequently in the women treated with sodium fluoride—a pleasing result, particularly in light of two previous trials sponsored by the National Institutes of Health (NIH), which found that sodium fluoride was no more effective than placebo [11, 12]. There are several plausible explanations for the difference between this most recent study and the earlier two, including the lower dose, the different formulation, and the intermittent regimen of sodium fluoride used by Pak and colleagues. But there may be a more important reason for the difference. Patients in the earlier trials almost certainly had more advanced osteoporosis and more fractures at the start of therapy. Careful inspection of the new data, all of which is clearly documented in the report, indicates that slow-release sodium fluoride was most effective in those persons who had the least disease at the start of therapy. Sodium fluoride was, in fact, no more effective than placebo in preventing recurrence of fracture in an already fractured vertebra. It was also only marginally better than placebo in preventing fractures in those persons with the lowest bone mass (< 65% of bone mass in a healthy 30-year-old woman). When given to women whose bone mass was at least 65% of that of a healthy 30-year-old woman, new fractures were almost abolished.

It has been difficult for the scientific community to accept that sodium fluoride, which is so potent in stimulating new bone formation and increasing bone mass, has had such a lackluster performance in reducing the vertebral fracture rate in osteoporosis. The explanation usually put forward is that the dose of fluoride in these studies was so high that it was somehow toxic to the skeleton. Some patients treated with fluoride had histologic changes on bone biopsy specimens that were consistent with a diagnosis of osteomalacia [13]. This probably does not fully explain the limited therapeutic efficacy of the drug, because the vertebral changes seen radiographically in osteomalacia in no way resemble those seen in osteoporosis, even after sodium fluoride therapy. At least one study [14] has shown that fluoride-treated bone is more resistant to resorption than normal bone. This may partially explain the increased bone mass seen with sodium fluoride therapy but would also diminish the plasticity of the skeleton, which is believed to be the major rationale for the phenomenon of bone remodeling or turnover. Finally, animal and human data show that, per unit of bone mass or volume, fluoride-treated bone is less biomechanically competent than normal bone [15, 16].

Nonetheless, it is now possible to draw some fairly firm conclusions about the potential role of sodium fluoride therapy in osteoporosis. Sodium fluoride increases bone mass in what seems to be a dose-dependent manner, to a greater extent than is currently possible with other approved or experimental therapies. Sodium fluoride is potentially toxic to the skeleton: It inhibits mineralization, interferes with normal bone remodeling, and decreases inherent biomechanical properties in a dose-dependent manner. Finally, even when used in doses sufficient to maximize the balance between these beneficial and deleterious effects, the best therapeutic results are seen when the drug is administered as early as possible in the course of spinal osteoporosis. As has been suggested [17, 18], sodium fluoride may have its most useful role in preventing even the first fracture.

The emphasis is again placed on the early detection of low bone mass using any means currently available. Plain radiographs of the spine are inferior to all techniques for direct quantitation of bone mass, and it has been estimated that 30% of bone mass must be lost before loss can be reliably detected on plain radiographs. Despite this severe limitation, ordering a spine radiograph for a patient and specifically requesting that the radiologist comment on bone mass starts the physician thinking about osteoporosis in that patient, perhaps for the first time. That is the time to advise the patient about adequate calcium intake, adequate exercise, and a safe home environment. The risk:benefit ratio of prescribing one of the two FDA-approved osteoporosis therapies should be carefully considered and discussed with the patient. It is not yet time to consider prescribing experimental medications not yet approved by the FDA, including slow-release sodium fluoride or one of the several newer bisphosphonates that are receiving extensive clinical testing. But that time is not far off, and one or more additional therapeutic options will almost certainly be approved by the FDA within the next 2 years. Now is the time for the primary care physician to start boning up on osteoporosis.