BoneKEy-Osteovision | Commentary

How do antiresorptive agents reduce fracture risk?



DOI:10.1138/2001012

Several rigorously designed and executed studies provide compelling evidence that antiresorptive agents reduce the risk of fracture. However, we know little about the structural changes responsible for this risk reduction. For example, there is no evidence that these drugs restore bone lost during aging even though the increase in bone mineral density (BMD) following treatment deceptively suggests otherwise. Nor is there evidence that these drugs reconnect trabeculae, synthesize new trabeculae, restore trabecular thickness or rebuild the thinned and porous cortical shell by increasing periosteal apposition, net endocortical apposition and reducing intracortical porosity. How then do resorption inhibitors increase BMD and reduce fracture risk?

Filling the reversible deficit in the remodeling space

There are two types of deficits in bone mass, an irreversible component which is the result of the negative bone balance produced by an imbalance in the volume of bone removed and replaced in each bone remodeling unit (BMU) during aging. This is the structural basis of bone loss: no negative bone balance, no bone loss. The second deficit is the reversible or transient remodeling deficit produced by the time delay between completion of bone resorption and initiation and completion of bone matrix (osteoid) formation, including its primary and secondary mineralization ().

The size of the reversible remodeling space deficit is determined by the remodeling rate (number of remodeling units created per unit time) and the duration of the resorption and formation phases of the cycle. The presence of this transient remodeling deficit is not discernible in steady state because it is always being created and‘filled’ as remodeling proceeds. (It is not responsible for the progress fall in BMD, it is the irreversible component doing this.). The reversible remodeling space deficit becomes detectable when steady state is perturbed. For example, at menopause, part, perhaps a large part, of the initial accelerated loss of bone observed by densitometry is the expansion of the remodeling space produced by the increased numbers of remodeling sites created by estrogen withdrawal. The loss then continues because of the contribution of the irreversible component of bone loss produced by increased osteoclast survival and reduced osteoblast survival producing the negative bone balance in the BMU.

With antiresorptive drug therapy, the initial rapid gain in BMD observed by densitometry is the result of filling of the remodeling space. Before drug treatment, remodeling is usually high and bone loss proceeds because of the negative bone balance in the many BMUs remodeling bone on its endosteal (trabecular, endocortical, intracortical) surfaces. When the antiresorptive drug is given, the number of new BMUs created on the bone surfaces per unit time decreases, but the high number of BMUs (present at various stages in their remodeling sequence before treatment was started) go to completion filling the reversible remodeling space with osteoid which undergoes primary then secondary mineralization. The mineralization process is also part of the reversible remodeling space.

This filling of the reversible remodeling space occurs on trabecular, endocortical and intracortical surfaces of already existing bone produces the initial rapid increase in BMD so well documented with antiresorptive drugs. The increase in BMD is a ‘step up’ to another level of BMD from which bone loss will proceed at a rate determined by the size of imbalance in the BMU and the new and slower remodeling rate. Thus, the increase in BMD may be perceived to be evidence of the reconstruction of an eroding skeleton, but it is not. There is no evidence of new bone tissue deposited and mineralized, on the subperiosteal surface increasing bone size, on the endocortical surface thickening the cortex, or on trabecular surfaces thickening trabeculae.

Increasing the true mineral density of bone matrix

The slower remodeling rate produced during drug therapy allows more complete secondary mineralization of this newly formed bone, and of bone in other areas in various stages of secondary mineralization. Had the high remodeling rate continued, bone at various stages of mineralization would be removed and replaced with new bone in earlier stages of secondary mineralization. This secondary mineralization plays a major role in increasing in BMD in the first 12-18 months of drug therapy, perhaps longer.

Boivin et al report that the mean degree of mineralization of bone (MDMB) measured by quantitative microradiography in women treated with alendronate was higher than controls by 9.3% (cortical bone) and 7.3% (cancellous bone) after 2 years (n = 9), and 11.6% and 11.4% higher than controls in these respective bone compartments after 3 years (n = 16) (). There was no increase in trabecular bone volume (number or thickness) in the alendronate group. The authors reported that the group differences in MDMB are likely to account for majority of the increase in BMD seen with alendronate.

In a paper published recently by Tonino et al, alendronate given without interruption for 7 years increased lumbar spine BMD by 11.4% (10 mg/day group) and by 8.2% (5 mg/day group). There was a ~1.5% increase in spine BMD in the last 2 years in each of these doses but no increase or decrease in spine BMD in those who took 20 mg/day for 2 years, 5 mg/day for 3 years then placebo during these last 2 years (). This is the longest study done using an antiresorptive drug and it raises important issues.

How can BMD continue to increase during 5-7 years of treatment with antiresorptive agents? MDMB did increase between years 2 to 3 so the slower remodeling may allowing more complete mineralization of more and more bone, even in the 6th and 7th year of treatment. Is this continued increase desirable? The secondary mineralization reported by Boivin et al was not hypermineralization and it is possible that the reduction in fracture risk observed with alendronate during three years is partly due to this change.

However, there is an inverted U shaped relationship between bone strength and ash density of bone (). Strength increases as density increases but then falls as further increase in true matrix density makes the bone more brittle (less ‘tough’). Bisphosphonates have been reported to increase microcrack density and may reduce bone toughness in dogs (). The doses of alendronate and risedronate used were high, the study was done in ribs and bending strength was maintained, but the observations should not be ignored. Tonino et al reported spine fractures in ~6% and nonspine fractures of ~7% during the last 2 years of the 7 year study in those taking placebo, 5 mg or 10 mg (~3% per year). These fracture rates were no different in the three groups, but different compared to fracture rates reported in the 3 year study by Liberman et al from which these individuals originate (). These fracture rates were 3% (spine) and 3.7% (nonspine) in the treatment arm (~1% per year). Inferences from these data cannot be made because the patients were older than those in the study by Liberman et al and there is no parallel control group, but they do raise the question of the long term antifracture efficacy of antiresorptive drugs.

The FIT study was one of the best designed and executed studies in the field and it provided compelling evidence for antifracture of alendronate during 3-4 years. However, the evidence that bisphosphonates, or any antiresorptive agent, reduces fracture risk beyond this time is not available so we have no rational basis for giving treatment beyond 3-4 years. Randomized trials are needed to address this hypothesis given the possibility that the contrary may be true. The hypothesis is two sided. Does continuing an antiresorptive agent beyond 3-4 years reduce the risk of fracture or increase it? Does stopping an antiresorptive agent after 3-4 years increase fracture risk during the next 1-2 years? If drugs are to be recommended for the long term use or stopped then evidence is needed justifying each approach.

Reducing the imbalance in the volume of bone removed and replaced in the BMU

BMD may increase if there is a positive bone balance produced by the antiresorptive drugs in each BMU. If the imbalance in the volume of bone removed and replaced in the BMU is lessened bone loss will slow down, if the BMU imbalance is abolished bone loss will stop; but neither of these changes increase BMD. A positive BMU balance will increase BMD if each completed remodeling sequence deposits new bone, thickening trabeculae and the cortex by net endocortical apposition. There is evidence for the production of a less negative or positive BMU balance with antiresorptive agents. Risedronate has been reported to improve BMU balance when given to beagle dogs, and there are studies in humans with estrogen and etidronate supporting the view that these drugs may reduce the BMU imbalance (). There is evidence that alendronate reduces osteoclast survival, which may reduce the depth of bone resorption, while increased osteoblast survival may increase bone formation (). In this way a more shallow resorption pit can be filled or overfilled with bone. This is a plausible way that antiresorptive agents can increase trabecular width or cortical width but there is no histomorphometric evidence to support this change. It is difficult to envisage how this can be achieved if remodeling is reduced. If a positive BMU balance could be achieved, it would then be advisable to increase remodeling to take advantage of the positive bone balance to rebuild the skeleton.

Early reduction in fracture rates

Black et al report the results of a subgroup analysis in which the fracture arm of the Fracture Intervention Trial (FIT) and the patients with osteoporosis (T > 2.5 SD below the mean at the hip) in the nonfracture arm of FIT were combined (). Combining these arms did not produce any surprises given the reduction in spine and nonspine fracture risk was observed in each of these two arms in the original studies, but what emerges is that the fracture risk reduction occurred within 12-18 months of treatment. This is also reported with risedronate (), raloxifene (), and calcium and vitamin D (). In other words, within this time period, the proportion of subjects sustaining fractures is less in the treated group than controls.

Several issues arise. Is the early risk reduction observed early maintained? The cumulative incidence curves seen in all the above mentioned studies separate early but it is difficult to determine whether they continue to separate - if so, the relative risk reduction is maintained, if the curves are parallel after the first 12-18 months then the early risk reduction is lost.

What is the structural basis of the risk reduction? Is the risk reduction due to continued progression of bone fragility in controls with, stabilization, slower progression, or a decrease in fragility (restoration of strength) in the treated group? Of these processes, what is the role of filling of the remodeling space and increased true density of the mineralized matrix, reduced remodeling, reduced progression of architectural disruption, or changes in osteocyte survival in the risk reduction? Answers to these questions are not available because the relative contributions of these and other structural abnormalities to bone fragility are not known. We have effective treatments for delaying the progression of bone fragility in high risk individuals but zero fracture growth is a distant horizon that moves further and further away as we run towards it because the solution to the population burden of fractures may not be in drug therapy, but that's just another one of the 6 million stories in the Naked City.


Creative Commons License This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.