Bone Mineral Density After Transitioning From Denosumab to Alendronate

David Kendler; Arkadi Chines; Patricia Clark; Peter R. Ebeling; Michael McClung; Yumie Rhee; Shuang Huang; Robert Kees Stad


J Clin Endocrinol Metab. 2020;105(3) 

In This Article


Transitioning to alendronate was generally effective at preserving the gain in BMD and suppression of BTMs achieved with 1 year of denosumab treatment. Among the minority of participants who lost BMD at the LS, TH, or FN after transitioning to alendronate, few fell below their pretreatment baseline BMD value, and only 1 participant lost BMD at all skeletal sites. Larger BMD increases in year 1 on denosumab were observed for participants who lost BMD in year 2 on alendronate, while other participant characteristics showed no numeric trend with the BMD response in year 2 on alendronate. These findings demonstrate that most women receive benefit from oral bisphosphonate therapy following denosumab cessation.

In the pivotal phase 3, randomized FREEDOM trial and open-label extension, treatment with denosumab for up to 10 years was associated with a continued increase in BMD, sustained reduction in BTMs, and low incidence of fractures, and was generally well tolerated.[9] In head-to-head studies, denosumab treatment led to larger increases in BMD and greater reductions in BTMs compared with alendronate,[22,23] which is in agreement with the results presented here. However, because denosumab is a reversible inhibitor of RANKL, denosumab's effects on bone turnover are reversible with discontinuation, and cessation of denosumab has been associated with rapid loss of vertebral fracture protection, including multiple vertebral fracture.[24] Thus, although denosumab treatment can produce large gains in BMD and significant suppression of BTMs, these effects do not protect patients when therapy is discontinued. For this reason, the use of a "drug holiday" in patients receiving denosumab is not recommended.[25] The decision to discontinue denosumab treatment should be accompanied by careful monitoring and use of a follow-on antiremodeling agent.

Limited data are available regarding the optimal post-denosumab bisphosphonate treatment regimen.[26] In a small case series evaluating women followed for up to 2 years after the FRAME trial, zoledronic acid (n = 11) infusion after denosumab discontinuation showed 73% to 87% preservation of the gains in BMD at the TH or LS after 1 year, with minimal further BMD loss at any skeletal site in year 2; participants given risedronate (n = 5) showed only partial preservation of BMD (41%–64%).[13,27] These findings are somewhat inconsistent with a previous, smaller case series showing only minimal efficacy to preserve BMD after denosumab treatment cessation.[28] This difference in outcome may be related to the inclusion of participants with different ages, previous osteoporosis treatments, and durations of denosumab treatment, and it should be noted that intravenous zoledronic acid was administered 6 or 8 months after the last denosumab injection. In the current analysis, which included a larger number of participants , 1 year of alendronate treatment following 1 year of denosumab led to maintained or increased BMD at the TH or LS in 84% to 92% of participants. These findings demonstrate the potential for bisphosphonate treatment to prevent reversible bone loss in patients who discontinue denosumab treatment and may suggest that oral alendronate helps maintain BMD in the period immediately following denosumab discontinuation when the effects of treatment on bone turnover have dissipated. Ongoing randomized clinical trials evaluating bisphosphonate use after denosumab discontinuation should provide clarity on the optimal treatment regimen.[29,30]

Although the majority of participants maintained or gained BMD after transitioning from denosumab to alendronate in the current analysis, 15.9%, 7.6%, and 21.7% lost BMD at the LS, TH, and FN, respectively, with only 1 participant losing BMD at all sites. These differences between sites in the response to treatment may be due to differences in the proportions of cortical versus trabecular bone as well as load-bearing parameters. Also, degenerative artifact is likely to have a greater impact on BMD at the LS compared with TH. In general, cortical bone may respond less rapidly to antiresorptive therapies compared with trabecular bone, as cortical bone has less surface area per unit volume of mineralized bone matrix upon which bisphosphonates can be adsorbed.[31] Specifically, denosumab is likely to have superior effects on cortical bone as compared with alendronate. In a head-to-head study using high-resolution peripheral computed tomography (HR-pQCT), there was greater reduction in cortical porosity at the radius and tibia in patients receiving denosumab compared with alendronate.[32] Supportive data using another technology to measure cortical porosity indicate the efficacy of denosumab in reducing cortical porosity at the hip, a load-bearing site.[33] Although the FN is included in the TH region of interest, the FN comprises a much smaller region and provides a less precise measurement than the TH, which is the preferred region of interest for following individual patients. Together, these differences can lead to variability in the clinical response to osteoporosis therapies.

The current analysis sought to better understand patient characteristics linked to loss of BMD after transitioning from denosumab to alendronate treatment, and our results identified the change in BMD with denosumab treatment from baseline to month 12. Specifically, the BMD gain in year 1 on denosumab was numerically higher in individuals who lost BMD in year 2 on alendronate than in those who maintained or gained BMD in year 2. This result may be attributable to "regression to the mean," a characteristic of imprecision of measurement. Bone remodeling differs between individuals and is influenced by a variety of factors.[34,35] Closure of the remodeling space with antiresorptive treatment in high remodelers might produce greater gains in BMD with treatment. In these individuals, subsequent discontinuation of reversible treatment may result in resumption of the same level of remodeling, which may not be fully inhibited with a bisphosphonate. However, alendronate treatment was able to maintain BMD above the pretreatment baseline level in the majority of women who lost BMD in year 2.

No other baseline or year 1 participant characteristics could consistently identify participants who lost BMD after transitioning to alendronate treatment in this analysis. This finding suggests that all patients, regardless of baseline characteristics and fracture history, may benefit from follow-on therapy with bisphosphonates after discontinuation of denosumab. We did not observe a meaningful effect of adherence to oral alendronate treatment on the change in BMD in year 2 after denosumab discontinuation. However, this may be the result of the overall good adherence in our clinical trial patient population, as compared to clinical practice where adherence can be low. Thus, it is likely that the low rate of nonadherence did not allow us to detect an effect of nonadherence on BMD decline in our patient population. Our findings also stress the importance of BMD monitoring while on treatment to identify those at greatest risk for fracture, as those patients with the largest increase in BMD with denosumab might also be the most vulnerable to BMD loss while receiving alendronate. Unfortunately, in clinical practice, most patients who discontinue denosumab treatment do not receive any osteoporosis treatment in the year following discontinuation; moreover, approximately half of those who begin a prescription medication for osteoporosis after denosumab cessation stop the therapy in the subsequent year.[11]

Several limitations of this analysis should be considered. First, denosumab was only administered for 1 year before transition to alendronate. With longer-term denosumab treatment, there will be continued gains in BMD and the potential for greater bone loss after treatment is stopped, and it could be more difficult to preserve bone mass when transitioning to bisphosphonates. Determining the optimal timing, dose, and bisphosphonate medication to administer after denosumab cessation warrants further study. Second, this post hoc analysis had a modest sample size, particularly for the analysis of participants divided into BMD response groups (ie, lost, maintained, and gained). However, the current sample size is larger than that reported in other ad-hoc case series investigating denosumab discontinuation and bisphosphonate follow-on therapy. Third, the analysis was not powered to detect statistical relationships between participant characteristics and their BMD response in year 2, and all analyses were descriptive in nature. Fourth, the effectiveness of alendronate to maintain BMD in this study may have been unrealistically high, given the difficulties with compliance and dosing faced in clinical practice.[36] Finally, DXA assessment for this study was not centralized; therefore, BMD results may vary by center.

Denosumab treatment increased BMD at all skeletal sites examined, and the gains in BMD achieved with denosumab were maintained in the majority of participants after transitioning to alendronate. Among participants who lost BMD in year 2 with alendronate, the majority remained above their pretreatment baseline value. Those with larger BMD increases in year 1 often showed greater BMD losses in year 2, with other participant characteristics not related to the response in year 2. These data highlight the need for oral bisphosphonate therapy following denosumab cessation and BMD monitoring of patients transitioning from denosumab to bisphosphonates.