Is a change in bone mineral density a sensitive and specific surrogate of anti-fracture efficacy?

Anti-resorptive agents perturb steady state remodeling; they suppress, but do not abolish, the birth rate of new basic multicellular units (BMUs). In doing so, remodeling goes to completion with bone formation in the many BMUs created before treatment but now with fewer resorption cavities appearing concurrently. As a result, cortical porosity and trabecular stress concentrators decrease reducing bone fragility. From this improved bone strength, steady state is re-established at a slower remodeling rate that again produces bone fragility but more slowly as fewer new BMUs, each with a less negative BMU balance, produce cortical thinning and porosity, trabecular thinning and loss of connectivity while bone fragility progresses rapidly in controls. Thus, the fracture risk reduction--the incidence of fractures in patients treated with an anti-resorptive agent relative to the incidence in controls--is the net effect of the slowing or partial reversal of fragility and then reduced progression of structural abnormalities in treated patients and continued structural decay in controls. Although some morphological features in treated patients and controls may be captured in the bone mineral density (BMD) measurement, many are not. The early increase in BMD is largely determined by the pre-treatment remodeling rate whereas the later and more modest BMD increase is a function of the degree of suppression of remodeling and secondary mineralization. When pre-treatment remodeling rate is low, the increase in BMD is small but the fracture risk reduction (relative to controls with comparable baseline characteristics) is no different to that in patients with high baseline remodeling (relative to their controls) and a greater BMD increase. Therefore, a small increase in BMD does not mean treatment has failed and a large increase in BMD is not indicative of a greater fracture risk reduction.