Bone Mass, Bone Fragility, and the Decision to Treat

The explosion of osteoporosis-related diagnostic and therapeutic options in the past 6 to 10 years has created its own challenges. Guidelines have recently been published concerning who should be tested¹ but, in a sense, that issue is moot. Bone mass measurement technology is proliferating rapidly, and physicians increasingly are confronted with ostensibly healthy individuals who bring them a printout showing that they have low bone mass. Should the physician recommend one of the growing array of bisphosphonates, selective estrogen receptor modulators, hormone replacement therapy regimens, or dietary supplements? Normally, the clinician treats the patient, not a test, but when it comes to prevention, the test result may be all the physician has to go on.

Osteoporosis is now defined as a condition of skeletal fragility due to low bone mass, microarchitectural deterioration of bone tissue, or both.² This important revision emphasizes the fragility and relegates bone mass to the status of a risk factor (any of several). The problem is that fragility is not measurable, while bone mass is. Thus, the field is divided on how to approach the issue. An international panel of experts recently assigned a label of "osteoporosis" to bone mineral density (BMD) more than 2.5 SDs below mean values for young adults,³ but recognized, without explicitly dealing with the issue, that patients with fragility fractures must be considered osteoporotic even if their BMD values are higher than that cutoff. The inadequacy of basing a diagnosis solely on bone mass is further emphasized by the fact that, after adjusting for BMD, many other factors, including age, a history of any fracture after age 40 years, and maternal history of hip fracture, independently predict fragility fracture^{4 - 6} —in many cases, more strongly than does low bone mass. History of prior fracture is an especially important prognostic feature because it demonstrates not simply propensity to fracture, but manifest fragility (whatever its basis). In such high-risk individuals, bisphosphonates protect against further bone loss and significantly reduce fracture risk.⁷

In this issue of THE JOURNAL, Cummings et al⁸ report the results of a 4-year intervention using a bisphosphonate, alendronate, in women with below-average bone mass but without prior osteoporotic fractures. An earlier report presented results in women with prevalent fractures.⁷ Despite some similarities, there are important differences in outcomes between the 2 groups.

The treatment-induced change in BMD in these fracture-free women was virtually superimposable on the corresponding changes in the group with prevalent fractures,⁷ as well as on the original alendronate data of Liberman et al.⁹ Moreover, the treatment produced a highly significant reduction in silent but radiographically detectable spine deformities, amounting to 44% in the group as a whole, almost exactly the same as the reduction reported earlier for women with prior fractures.⁷

A positive change in bone mass should mean some reduction in fracture risk. However, that relationship is not a simple one because an increase in BMD does not reverse the loss in strength caused by severed trabecular struts, and no bone active agent is able to reestablish severed connections. Bisphosphonates have been reported to decrease fracture risk substantially in patients with demonstrated fragility,^{7 ,9 - 10} sometimes out of proportion to their effect on BMD. This apparent anomaly is usually attributed to the reduction in overall remodeling because remodeling pits can be loci of weakness.

The important clinical issue is whether bisphosphonates reduce fragility in individuals who have only low bone mass and have not exhibited bony fragility (if they ever will). The answer from the study by Cummings et al⁸ —the largest study of its kind—is maybe. The BMD values in these women would have placed them at or below the 50th percentile for their age, making them logical candidates for fracture prevention. Overall, 584 clinically apparent fractures occurred, for an annual rate of 3 to 4 per 100 women per year. The number of fractures in the alendronate-treated group was 14% less than in the placebo group, but the effect was not statistically significant. There was, however, a significant interaction with baseline BMD. The risk of clinical fracture was significantly reduced in women with the lowest BMD, but not in women with higher BMD.

By subtype, hip fracture risk was reduced by 21%, but wrist fractures in the alendronate group were increased by 19%. Neither difference was statistically significant. In further post hoc subgroupings, fractures other than spine, hip, and wrist were significantly fewer in the alendronate-treated group, but the difference in absolute fracture rates was small (2%). However, when stratified by baseline BMD, wrist fracture was not reduced; in fact, those with higher baseline densities had an increased risk of fracture. The practical significance of such analyses is unclear because performing multiple post hoc analyses inevitably leads to "significant" differences of uncertain clinical importance.

These somewhat disappointing results do not so much represent a failure of the bisphosphonates, which are extremely useful agents overall; rather, they reflect the complexity of osteoporotic fragility. The field of osteoporosis has been overwhelmingly focused on bone mass, both because imagers can measure it and because clinicians can affect it. However, the contemporary definition of osteoporosis stresses the importance of architectural (and other) abnormalities, and the evidence that fragility involves something more than a simple reduction in bone mass is too compelling to continue to be ignored. Among the more than 4000 women who were subjects of this trial, it is likely that many had true osteoporotic fragility but, by chance, had not yet expressed it as a fracture. Presumably, those women received the same benefit from treatment as those in the fracture cohort. But that effect was diluted by the many other women who had only low bone mass and were not at the same risk.

Short of finding some as yet unimagined genetic marker for the propensity to architectural skeletal damage, it is unlikely that any biochemical test will be able to help clinicians distinguish which patients with the same bone mass have greater or lesser inherent fragility. Other testing modes, such as 3-dimensional, fine-structure computed tomography or magnetic resonance imaging, seem promising but are far removed from being cost-effective screening tools.

How is the physician to respond to an individual patient who presents simply with a low bone-mass value? Delaying intervention in women until an osteoporotic fracture occurs seems as unsatisfying as waiting to treat hypertension until a stroke occurs. Clearly, the treatment decision will have to be individualized. Women with BMD T scores of more than 2.5 SDs below the mean, with a history of any fracture after age 40 years, with a maternal history of hip fracture, starting corticosteroid therapy, or about to be subjected to prolonged immobilization are the women most likely to benefit. Lacking such reasonably secure indications, it is certain only that bisphosphonates will protect against development of asymptomatic vertebral deformities and further bone loss.

Fortunately, bisphosphonates are not the only option. Similar benefits to protect bone mass and prevent fracture can be achieved with the selective estrogen receptor modulators,¹¹ hormone replacement therapy,¹² and calcium and cholecalciferol (vitamin D).^{13 - 14} In selecting a prophylactic regimen for individuals at less risk, it makes sense to use an agent that offers multiple benefits. Even if the patient would never have had a fracture, at least for hormone replacement therapy and nutrition, there are also nonskeletal benefits. In any event, the physician needs to ensure that women at risk have a calcium intake of at least 1500 mg/d and cholecalciferol intake of 600 to 1000 IU and participate in weight-bearing exercise. These are the irreducible foundations for whatever pharmacological regimen may be selected. For now, with the exception of asymptomatic spine deformities, the antifracture benefit of bisphosphonates in women with low bone mass but without prevalent fracture must be judged to be small.