An Essential Warning

Over the past two decades, osteoporosis treatment options have expanded greatly and physicians and patients are now able to choose among several classes of medications and numerous specific drugs. As these options have grown, however, the choice of which medication to prescribe and how to balance the risks and benefits in specific populations and individual patients has become increasingly complex. This complexity, in turn, combined with the proliferation of well-publicized reports of rare but serious side effects, is exacerbating the recent trend of declining rates of osteoporosis treatment. Thus, in the absence of rigorous randomized-controlled trial data addressing the length-of-treatment question, it is perhaps not surprising that published guidelines on this issue differ substantially, with some professional organizations and expert panels suggesting that all antiresorptive drugs be stopped after 3 to 5 years and others proposing that high-risk patients can be treated substantially longer.1-3 In the setting of this uncertainty, the timely article by Cummings and colleagues in the current issue of JBMR, in which they present a post hoc analysis of their randomized placebo-controlled FREEDOM Trial, should have an immediate clinical impact.4 In this report, Cummings and colleagues compare the rates of new or worsening vertebral fractures in postmenopausal osteoporotic women who discontinued denosumab with those who discontinued placebo. The fundamental finding of this analysis is that the incidence of vertebral fractures, which was reduced during ongoing denosumab therapy, rapidly returns to rates observed in placebo-treated patients when the drug is stopped. Moreover, the analysis shows that among women who experience a posttreatment vertebral fracture after discontinuing denosumab, 61% experience multiple vertebral fractures and 23% experience ≥4 vertebral fractures, which exceeds the percentage of those discontinuing placebo of whom 39% experience multiple fractures and only 6% experience ≥4 fractures. Importantly, the report by Cummings and colleagues follows several dramatic case reports of patients who experienced multiple vertebral fractures in the 2 to 10 months after stopping denosumab as well as in the context of the long-appreciated and well-described accelerated bone turnover and rapid decline in bone mineral density that occurs when denosumab is discontinued.5, 6 These observations regarding denosumab discontinuation differ from the results of analyses performed in women discontinuing nitrogen-containing bisphosphonates, in which there appears to be continued anti-vertebral fracture efficacy for 3 to 5 years after drug discontinuation, likely because of the long skeletal half-life of these agents.7, 8 They are also in contrast to the pattern observed in women discontinuing teriparatide, in which an anti-fracture effect appears to be sustained for up to 18 months after the drug is stopped, even though bone density is declining.9 So what should a clinician do following this report and the others referenced above? The first important point that this analysis makes clear is that patients treated with denosumab must be encouraged to adhere to a strict “every 6-month” injection schedule and should be counseled that delays of even a few months may have significant negative consequences. It is also reasonable to suggest that physicians reconsider the appropriateness of prescribing denosumab to patients with a history of poor drug adherence, in which intravenous zoledronic acid may be a better choice. Second, these results clearly demonstrate that a standard “drug holiday,” generally associated with bisphosphonates, should not be initiated in patients treated with denosumab. Because denosumab appears to be efficacious for up to 10 years,10 an option that could be considered in patients who remain at very high risk of fragility fracture is to continue denosumab therapy beyond the 5-year time point, despite the fact that the incidence rate of atypical fracture after extended denosumab, while low, is currently difficult to precisely define (of note, similar recommendations were recently suggested in a position statement of the European Calcified Tissue Society).11 Conversely, in patients in whom discontinuing therapy is desired, an intervening course of antiresorptive therapy should probably become the standard of care, despite only limited available evidence of the effectiveness of such an approach at present.12, 13 Although there are no comparative efficacy trials to suggest which antiresorptive is optimal to use after denosumab, given the relative potency of the various agents and their mechanisms of action, bisphosphonates should be strongly considered in patients without a contraindication. Additionally, it is very clear that switching from denosumab to the anabolic agent teriparatide, and presumably abaloparatide as well, should be avoided because this transition is associated with even more extensive skeletal remodeling and even greater bone loss than that observed when denosumab is discontinued without a drug transition.14 Finally, although not addressing the issue directly, this report raises questions concerning the optimal timing of bisphosphonate administration when used after denosumab. Given bisphosphonates’ mechanism of action and preferential deposition at sites of active bone remodeling,15 it is plausible that if exposure occurs while denosumab's antiresorptive effects are still maximal, the drug will be considerably less effective than if given to a patient with ongoing robust bone turnover. This issue may be of minimal importance if an oral bisphosphonate is being repeatedly administered weekly or monthly but may be quite relevant to patients transitioning from denosumab to yearly zoledronic acid. The current analysis by Cummings and colleagues provides an essential warning that must be communicated to patients with osteoporosis and to the physicians and other health professionals who treat them, including primary care providers who may not be exposed to the subspecialty bone literature. Simply put, drug holidays cannot be safely initiated in patients treated with denosumab. Moreover, this analysis, combined with prior work showing that transitioning from denosumab to teriparatide is also contraindicated, points to a rational (albeit unproven) path forward—namely an intervening transition to a limited course of bisphosphonate therapy. Clinical trials are now needed to rigorously assess the efficacy of bisphosphonate use after denosumab, clarify the optimal timing for transitioning from denosumab to bisphosphonates, define an optimal duration of follow-up bisphosphonate therapy when used after denosumab, and determine whether estrogens or SERMs are capable of inhibiting post-denosumab high-turnover bone loss in patients for whom bisphosphonates are contraindicated.