The search for the optimal anabolic osteoporosis therapy


  • Dennis M Black,

    Corresponding author
    1. Department of Epidemiology and Biostatistics, University of California, San Francisco, CA, USA
    • Address correspondence to: Dennis M Black, PhD, University of California, San Francisco, Department of Epidemiology and Biostatistics, 185 Berry Street, Lobby 5, Suite 5700, San Francisco, CA 94107. E-mail:

    Search for more papers by this author
  • Anne L Schafer

    1. Department of Medicine, University of California, San Francisco, CA, USA
    2. Endocrine Research Unit, San Francisco Veterans Affairs Medical Center, San Francisco, CA, USA
    Search for more papers by this author

  • This is a Commentary on Horwitz et al. (J Bone Miner Res. 2013;28:2266-2276. DOI: 10.1002/jbmr.1978).

Large increases in bone mineral density (BMD) and reductions in fracture risk, particularly vertebral fractures, have generated much enthusiasm for anabolic therapy in the osteoporosis community. Two forms of parathyroid hormone (PTH) (PTH[1-34] or teriparatide, and PTH[1-84]) have been used clinically for almost a decade (PTH[1-84] has not been approved for use in the United States). They have been shown to be safe and effective in increasing BMD and reducing fracture risk. However, PTH increases bone resorption as well as formation, and an ideal anabolic treatment would be one that could optimize the impact on formation while producing lesser changes in resorption. The search for such a treatment has been a theme in many interesting clinical studies over the last 10 years. The report by Horwitz and colleagues[1] in this issue of the Journal of Bone and Mineral Research details a 3-month trial of one such approach: PTH related protein (PTHrP[1-36]). Earlier, preliminary data collected in short-term and small human studies,[2-4] as well as preclinical studies,[5] had suggested that PTHrP might increase formation without increasing resorption and provided the rationale for the current randomized trial comparing two doses of PTHrP to teriparatide over 3 months.[1] The absence or near absence of hypercalcemia with comparable doses in those prior studies[2-4] also suggested that PTHrP might avoid that known potential side effect of PTH therapy.

Horwitz and colleagues[1] show that, indeed, PTHrP has a net anabolic effect with a significant bone formation marker increase over the 3 months of the study. The net effects of PTHrP on BMD over 3 months were indistinguishable from those of teriparatide, although inferences about BMD are limited by the small study size and relatively short duration. However, contrary to the investigators' prior hypotheses, a resorption marker increase was also seen. Furthermore, while the resorption marker increase was modest, the increase in formation with PTHrP was also much more modest than that for teriparatide. Figure 2 in Horowitz and colleagues[1] shows this dramatically: at 3 months, procallagen type 1 amino-terminal propeptide (P1NP) (the marker of formation) had increased by 171% with teriparatide compared to only 46% for the 400-μg dose of PTHrP. Cross-linked C-telopeptide (CTX) (the measure of resorption) also increased much more with teriparatide than with PTHrP (92% versus 30%).[1]

In terms of safety, the investigators had hypothesized that PTHrP would be less likely to induce hypercalcemia than teriparatide, based on the absence or near absence of hypercalcemia in their prior studies with the doses used here. Surprisingly, the results showed the opposite: PTHrP led to more instances of elevated serum calcium (defined as serum Ca >10.5 mg/dL) than teriparatide. The difference was quite striking, with zero cases in 35 participants randomized to teriparatide versus 18 of 70 (26%) among those randomized to (either dose of) PTHrP. In their discussion, the investigators speculate about the possible reasons for increased hypercalcemia compared to their earlier studies, but they could not definitively explain these results. The number of adverse events, although small, did not differ by study group. However, the number of participants who discontinued treatment due to adverse events was higher in the PTHrP groups (7/70, 10%) than in the teriparatide group (1/35, 3%), suggesting that teriparatide might be better tolerated overall than PTHrP.

Is there hope for future development of PTHrP(1-36)? In order for PTHrP to become an attractive new anabolic option, it would need to be either more effective or safer than PTH therapy, or more acceptable in its administration (eg, not a daily injection). In terms of efficacy, these results may seem disappointing in that increases were seen in the bone resorption marker, and increases in both resorption and formation markers were more modest than those with teriparatide. However, this pattern of changes in bone turnover markers—with smaller increases in formation markers than teriparatide but also smaller increases in resorption—could potentially translate into increases in BMD and bone strength that win the BMD race with longer-term treatment. We really do not know how to interpret changes in bone turnover markers with anabolic therapy nor to infer their impact on bone mass and strength. Indeed, our limited ability to infer bone mass changes from changes in bone markers was demonstrated in a recent study of teriparatide in combination with alendronate.[6] In that study, the addition of teriparatide to alendronate resulted in smaller changes in bone turnover markers but larger increases in BMD compared to the discontinuation of alendronate and initiation of teriparatide. Although the BMD changes with PTHrP(1-36) were not different from those of teriparatide, inferences are limited by small sample sizes and short duration. If PTHrP(1-36) is to be pursued as a future therapy, an investment in a longer-term and possibly larger trial with BMD endpoints—and eventually fracture endpoints—would be necessary. Regarding safety, the results of this trial suggest that a dose of 400 μg may be the maximal safe dosage for clinical use. The increased hypercalcemia would need to be addressed, possibly through lower calcium intake and/or the reevaluation of the synthesis of the drug as the authors suggest in their discussion.

Suggesting optimism for PTHrP therapy in general, an analog of hPTHrP was tested in a Phase II trial and at the highest dose resulted in 12-month spine BMD increases similar to or higher than those for teriparatide.[7] Over 6 months, there were increases in bone turnover markers, and there was a similar or lesser incidence of hypercalcemia compared to teriparatide.[8] This hPTHrP analog is currently being tested for fracture efficacy in a 2400-person Phase III trial.[9]

This work on PTHrP represents but one crusade in the overall hunt for the “holy grail” of anabolic therapy. That search has led in numerous interesting directions, including development of new drugs such as an anti-sclerostin antibody, for which results over 12 months have shown a strong increase in formation with a decrease in resorption,[10, 11] and larger Phase III studies are in progress. Another exciting area of exploration is combination therapy with PTH (currently our only anabolic therapy) and antiresorptive agents. Whereas the concurrent use of PTH(1-84) and daily alendronate seemed to result in the blunting of the anabolic effect of PTH on spine BMD and on markers of bone formation,[12] concurrent combination with other antiresorptive agents might hold more promise. In particular, a study of concurrent use of teriparatide with raloxifene (a milder antiresorptive than alendronate) showed intriguing results.[13] Also, concurrent use of less frequently dosed antiresorptives might be better than concurrent use of daily alendronate, and two studies provide some evidence for this. In a trial of concurrent zoledronic acid and teriparatide versus monotherapy,[14] BMD rose more rapidly in the early months with combination therapy, although by 12 months the combination group did not show a clear advantage. Recently, combination twice-yearly denosumab and teriparatide[15] produced a distinct balance of change in formation and resorption markers and greater increases in BMD over 1 year compared to denosumab or teriparatide alone (with a second year underway). Different approaches to anabolic therapy administration also warrant further exploration. In one recent trial, weekly teriparatide therapy resulted in an impressive reduction in vertebral fracture incidence compared to placebo.[16] A transdermal PTH(1-34) patch[17] showed promising results over 6 months in terms of bone marker changes and trends toward greater increases in BMD compared to injected teriparatide. However, all of these trials had BMD and/or bone turnover marker endpoints, and larger trials are sorely needed to determine possible fracture benefits of combination therapy and alternative routes of administration.

Taking advantage of the dynamism of changes in bone formation and resorption over time with different forms of anabolic (and antiresorptive) therapy could also hold promise to maximize long-term increases in bone strength. Daily PTH therapy dramatically increases formation very soon after initiation of therapy, but formation marker levels plateau and even begin to decline by 12 to 18 months, despite continuation of therapy.[14, 18, 19] We do not know whether this presages the end of the anabolic effect and might have implications for the optimal period of anabolic treatment. Furthermore, markers of bone resorption initially increase more slowly than formation. Whether this delay in resorption compared to formation (the “anabolic window”) is the primary engine of anabolic bone building is unclear. We have little understanding of the effects of these varying temporal patterns of changes in remodeling on changes in bone strength. Understanding these dynamic processes more fully may lead us down other roads such as multiple shorter courses of PTH therapy, which have had some limited but promising studies.[20, 21] Yet another interesting course to explore is the sequential use of anabolic and antiresorptive therapy. The use of 1 year of PTH followed by 1 year of alendronate showed strong increases in BMD and femoral strength,[22, 23] and some additional studies are exploring other sequential combinations.[24]

Odanacatib, although not an anabolic agent, was recently shown to have its own interesting pattern of bone turnover marker changes over 12 months.[25] Formation marker P1NP initially decreased (along with resorption marker CTX), but by 1 year P1NP had returned to pretreatment levels, whereas resorption remained suppressed, suggesting a possible long-term favorable change in formation/resorption balance.[25] It has been suggested that the initial decrease in bone formation may result from reduced growth factor release from the bone matrix (eg, transforming growth factor β [TGFβ], bone morphogenic proteins [BMPs]) as osteoclastic bone resorption is inhibited, but that the subsequent recovery of bone formation may reflect the persistence of relatively normal osteoclasts on bone surfaces during cathepsin K inhibition, and with it the persistence of direct cell-cell stimulation of osteoblasts by osteoclasts.[26] More research into the implications of dynamic changes in formation/resorption balance on bone strength, over the short-term and long-term, could help to identify fruitful areas for new drug development or exploration of new combinations of therapies.

The quest will continue for the “holy grail” of anabolic osteoporosis therapies, which will optimize the impact on bone formation relative to resorption. One hopes that current avenues of research, including new molecules such as PTHrP, anti-sclerostin antibody, and odanacatib, as well as combinations of existing therapies, will lead to fracture endpoint trials and eventually to powerful new tools for the treatment of osteoporosis.


DMB has received research grants from Novartis, Merck, and Roche and has consulted with Nycomed, Amgen and Eli Lilly. ALS has nothing to disclose.


This work was supported by the Department of Veterans Affairs under grant 5 IK2 CX000549-03 (to ALS).

Authors' roles: Both authors performed all author roles.