Addressing the Crisis in the Treatment of Osteoporosis: A Path Forward

Specific approaches to improve appropriate treatment of patients with osteoporosis

As reviewed in the American Society for Bone and Mineral Research task force reports,27, 28 although the relative risk of AFFs in patients taking bisphosphonates is increased, the absolute risk of AFFs in these patients is low, ranging from 3.2 to 50 cases per 100,000 person-years. Even though this risk may rise to as high as ∼100 per 100,000 person-years with long-term (>5 years) use,27 the overall risk of therapy and benefit/risk ratio with these drugs remains extremely favorable for those patients at high fracture risk of common osteoporotic fractures. Despite these reassuring statistics, primary care physician and patient attitudes have been greatly influenced by the media attention given to AFFs. There is pervasive concern among physicians, and particularly among patients, that these fractures may be vastly underreported and are perceived by many patients as a severe and somewhat paradoxical consequence of treatment intended to reduce fracture risk. As the documented decline in osteoporosis treatment demonstrates,20-22 the current approach of trying to familiarize patients with the favorable benefit/risk ratio is simply not working. As such, the challenge is to demonstrate to patients that we have heard their concerns. Key ways to address patient concerns include developing ways to diagnose AFFs before they occur and, over the longer term, to identify those patients who may be at increased risk of this complication even before starting osteoporosis medications.

Table 1 lists potential short-, intermediate-, and long-term approaches to address patient concerns about the risk of AFFs. Better patient and physician education is certainly a key component, including defining as clearly as possible the favorable benefit/risk ratio of bisphosphonates. Indeed, the best estimates are that with bisphosphonate therapy, one would prevent anywhere from 80 to 5000 fragility fractures for every AFF possibly induced by treatment.29 Additionally, patients should be appropriately counseled to recognize the prodromal symptoms of AFF, such as new groin or hip pain, and to report these to their health care provider. One way to do this might be to develop brief, standardized, and simple-to-administer questionnaires for common prodromal symptoms of AFF that could be linked to prescription renewals by physicians. Also on the provider side, appropriate decision-making tools need to be developed to guide physicians and health care providers regarding identifying high-risk patients and when to order follow-up studies, such as femur X-rays or extended femur dual-energy X-ray absorptiometry (DXA imaging; see below), particularly in patients who have been on therapy for more than 3 to 5 years.

A second short-term approach that could be implemented fairly rapidly is incorporating DXA imaging of the femur in the region where AFFs typically occur to identify potential early signs of an impending AFF. The feasibility of this approach has been demonstrated by McKenna and colleagues,30 who studied 257 patients over the age of 50 years, all of whom had been on bisphosphonate therapy for more than 5 years. Using an extended femur scan at the time of routine DXA (Fig. 1A), they found abnormal DXA images (eg, “flaring” or “beaking” of the cortex) in 19 subjects (7.4%). Follow-up radiographs (Fig. 1B) showed evidence of incomplete AFF in 7 patients (2.7%) (5 with periosteal flare and 2 with a visible fracture line), 7 showed no abnormality, and 5 showed an unrelated radiographic abnormality. It is important to note that although radiographic findings suggestive of AFF were noted in 2.7% of patients on long-term bisphosphonate therapy in this convenience sample of patients, population data on clinical AFFs indicate a far lower prevalence, at worst in the range of 0.13% to 0.22%.31 In addition, the majority of patients with radiographic changes consistent with partial or incomplete AFF may not, in fact, progress to clinical AFFs.32 Nonetheless, monitoring patients for such radiographic changes would clearly identify a potentially high-risk subgroup for more extensive imaging and consideration of drug discontinuation as appropriate.

Importantly, Hologic (Bedford, MA, USA) currently offers the option of performing this extended femur scan on its new scanners, and this approach has been approved by the FDA for monitoring AFF risk; GE Lunar (Madison, WI, USA) is in the process of offering a similar option, also recently approved by the FDA, as an update to existing software. Thus, baseline and follow-up “monitoring” scans performed perhaps annually starting 3 to 5 years after the initiation of therapy makes eminent clinical sense, analogous to monitoring liver enzymes for patients on a statin or serum creatinine for patients starting ACE inhibitors. Although providing some measure of reassurance to patients, it would still be important to guard against false reassurance and to emphasize that patients also notify their health care providers regarding onset of symptoms that might signal an AFF.

In addition to these short-term approaches, over the intermediate term (Table 1), we need to develop new tools to identify patients at increased AFF risk after treatment. As an example, Mahjoub and colleagues33 compared femur geometrical data (Fig. 2) using AP radiographs of the pelvis in 56 AFF patients versus 112 controls with traumatic or fragility hip fractures. In this analysis, patients with AFF had several geometrical features that distinguished them from controls, including an excessive femoral offset, proximal femur neck angle in varus, and greater proximal cortical thickness. If this approach could be validated in larger cohorts and incorporated into a baseline DXA scan, it may serve to identify a subset of patients in whom either shorter duration of therapy and/or closer follow-up on therapy are warranted. Another potential risk factor for AFF is Asian race. Thus, Lo and colleagues34 noted a high proportion of those with atypical fracture had contralateral femur findings and focal cortical thickening, especially among Asians. The race distribution of atypical fractures differed dramatically from hip fractures in the Kaiser database: hip fractures, whites, 85%; blacks, 3%; Hispanics, 6%; and Asians, 5%. In contrast, the race distribution of atypical femur fractures was whites, 45%; blacks, 0%; Hispanics, 5.3%; and Asians, 50%. Asian women tend to have lower bone mineral density (BMD) and shorter hip axis length, suggesting that femoral biomechanical differences may contribute to the risk of atypical fractures.34 Other risk factors for AFF are remarkably similar to risk factors for typical osteoporotic fractures, including rheumatoid arthritis and >6-month use of glucocorticoids.35

The long-term vision (Table 1) for tackling the AFF problem includes a number of research strategies and stimulation of the new drug pipeline, which has been inexorably declining. On the research side, there has been a suggestion that genetic factors may play a role in conferring risk for AFF, such as the observation noted above that AFFs may occur more in Asian women.34 In fact, a small pilot study using an exome array in 13 patients with AFF and 268 controls identified a greater number of rare variants in the cases compared with the controls. The analyses were restricted to the variants with minor allele frequencies less than 3%, which limited the ability to make meaningful inferences about individual variants. There was no evidence of enrichment of any specific pathways using all of the results.36 At the 2016 annual meeting of the American Society for Bone and Mineral Research, Roca-Ayats and colleagues performed whole exome sequencing in three sisters who developed AFFs while taking bisphosphonates and three unrelated AFF cases.37 They found that all three affected sisters harbored a p.Asp188Tyr mutation in the GGPS1 (geranylgeranyl pyrophosphate synthase) gene, which is critical to osteoclast function, and can also be inhibited by bisphosphonates. This study was far too limited in size to have the power to identify genetic risks for AFF, yet provided early evidence that there may be a genetic predisposition to AFF. Ongoing studies are needed that can leverage well-phenotyped cases and controls in sufficient numbers to have the power to detect rare variants associated with AFFs, while not overlooking the possibility that common variants in multiple genes may contribute as well. Also noteworthy is a recent study of DNA methylation changes in bisphosphonate users that might implicate differential gene expression conferred by use of osteoporosis medications. This represents another promising line of research to identify the underlying epigenetic drivers of AFF risk in bisphosphonate users.38

Looking ahead to closing the treatment gap in osteoporosis, one of the most troubling concerns is the lack of new drug development. There is a need to bring new drugs to this growing market. As the average life expectancy increases worldwide, there will be a steady rise in the proportion of the population who are older than 65 years and even 75 years, and thus at risk for osteoporotic fractures, just by virtue of advancing age alone. There will be a need for new osteoporosis drugs that are effective in reducing common fractures, while circumventing the AFF side effect, which has been observed in most drugs with antiresorptive mechanisms. Recent activities provide encouragement for the possibility that drug approval requirements may be changing. The Biomarkers Consortium-Bone Quality Project is attempting to qualify a surrogate marker for fracture prediction that could be used in clinical trials,39 obviating the need for multiple large randomized trials with fracture as an endpoint. If such a surrogate marker is approved for osteoporosis drug development, this will provide a financial incentive to bring new drugs to market.

An additional long-term strategy to close the treatment gap is to better align recommendations and guidelines regarding screening and treatment of patients at risk for fracture. In contrast to guideline development in areas such as hyperlipidemia40 and hypertension,41 which produce regular unified and updated guidelines, osteoporosis guidelines come from a number of different organizations such as the National Osteoporosis Foundation42 and the American Association of Clinical Endocrinologists.43 At least in the US, a single guideline for fracture prevention would help providers follow a well-accepted strategy. Finally, and perhaps most importantly, we need much better patient engagement strategies to help build trust with our patients; this is a prerequisite for any approach that would increase treatment rates for those patients at high fracture risk.

An external assessment of osteoporosis management and research

The American Society for Bone and Mineral Research and its members have been very fortunate to have decades of strong support for basic and clinical research efforts by NIH from multiple Institutes. At a fiscal level, this has led to support for randomized trials, larger observational cohort studies, and novel high-risk, high-impact proposals, as well as bench studies that in some cases have the potential to translate to the bedside. For the most part, these efforts have paid substantial dividends with respect to osteoporosis medications. More recently, attention at the federal level has focused on patient-specific outcomes as primary endpoints for diseases of the musculoskeletal system. For example, numerous tools have been developed to measure individual aspects of disease progression as well as the impact of treatment. But surprisingly, we have few measures that provide insight into how patients view current treatment modalities for osteoporosis, nor how much risk an individual is willing to tolerate for adherence to these agents. Fortunately, NIH has been quick to recognize the enormity of the problem and is committed to providing several possible long-term approaches in collaboration with clinician-scientists in our field. The most compelling is the potential investment into a systematic review of the scientific basis for current osteoporosis treatments, both in terms of benefits and risks, through the Office of Disease Prevention. This effort could ultimately lead to a greater scientific consensus while at the same time incorporating the fundamental importance of patient perspectives into future algorithms. NIH can also promote the development of tools that clinicians can use to gauge patient knowledge and willingness to adhere to anti-osteoporosis therapies. Undoubtedly, both short- and long-term efforts in collaboration with NIH will be necessary at several levels to address this urgent problem.

Although this perspective is focused principally on patient and physician concerns over AFFs, fear of osteonecrosis of the jaw (ONJ) also represents a significant barrier to treatment. Although rare in people treated for osteoporosis with bisphosphonates, with an estimated incidence of only 0.001% to 0.01%,44 reports of its occurrence and the associated media exposure have created significant concerns among patients, physicians, and dental practitioners. As with AFFs, there is a need to communicate more clearly to patients and their physicians and dentists the rarity of ONJ and the positive benefit/risk balance of bisphosphonate therapy in people at high risk of fracture. In addition, because dental disease is a major risk factor for the development of ONJ, taking measures to maintain good oral hygiene and dental health can further minimize the chances of developing this rare adverse effect.