Bisphosphonate Use and Atypical Femoral Fractures: Getting Down to Brass Tacks

No doubt about it—times are rough for clinicians and their patients trying to make rational decisions about when and how to use (or stop) effective therapies for osteoporosis. On the one hand, we have increasingly clear evidence that bisphosphonates given to older women and men with osteoporosis [defined as an existing vertebral fracture or a bone mineral density (BMD) T-score less than −2.5] prevent future hip, vertebral, and non-spine fractures (1–3). Conversely, with help from the Internet, high-profile newspaper articles, and even the Food & Drug Administration (4), we seem to be reminded on a daily basis that bisphosphonates may cause serious harm, such as osteonecrosis of the jaw, atrial fibrillation, esophageal cancer, and spontaneous femur fractures. It is no wonder that patients are worried and openly share their angst about osteoporosis treatment with their doctors, while clinicians are equally worried about causing more harm than good. The numerous reports of the “unusual femur fractures” in particular have been disquieting; clinicians are not happy to hear that they may be responsible for causing the problem they thought they were treating to prevent. The worry that excessive bisphosphonate dosing or prolonged use might cause microdamage accumulation from oversuppression of bone remodeling and increase rather than decrease skeletal fragility is not new (5). Such concerns were voiced a decade ago (6) but seemed unlikely after the above-mentioned large fracture endpoint studies showed beneficial effects on fracture risk with few complications. The catch? Placebo-controlled fracture endpoint trials have been at most 3–5 yr in duration, so inferences about any effects of treatment beyond 5 yr, both beneficial and harmful, are lacking solid evidence. Numerous case reports of unusual fractures of the femur among bisphosphonate-treated women have appeared over the last 5 yr (7–9). These “atypical” fractures were subtrochanteric (below the lesser trochanter) or diaphyseal (above the distal metaphysis), occurred with little trauma, and were characterized by transverse or nearly transverse “chalk-like” fracture lines as opposed to the more typical spiral or comminuted femoral shaft fractures. Subsequent reports highlighted the presence of prodromal thigh pain, bilateral involvement, the presence of other diseases such as rheumatoid arthritis and diabetes, and the concomitant use of other medications such as corticosteroids or proton pump inhibitors. Some reports, but certainly not all, suggest marked suppression of bone turnover as assessed by bone turnover markers and iliac crest histomorphometry (7). Even if causally related, these atypical fractures must be quite rare among women receiving bisphosphonate therapy for 3- to 5-yr duration because the rates of all types of subtrochanteric and diaphyseal fractures were low and similar between placebo and bisphosphonate-treated women in a reanalysis of three large clinical trials (FIT, FLEX, and HORIZON) (10). Although the study is reassuring, an important limitation is that few women received more than 5 yr of bisphosphonate treatment, information on atypical features was not specifically collected, and for the most part only radiographic reports were reviewed. Large observational studies have also examined the relationship between bisphosphonate use and subtrochanteric or diaphyseal fractures (11). In this issue of JCEM, Abrahamsen et al. (12) extend their previous work using Danish administrative data to further examine the relationship between bisphosphonate use and subtrochanteric or diaphyseal femur fractures among 39,567 alendronate users and 158,268 nonusers. As would be expected because of their higher pretreatment risk, alendronate users were more likely to suffer classic hip fractures than nonusers (relative hazard = 1.45; 95% confidence interval, 1.37–1.53), and a similar increase was observed for subtrochanteric and diaphyseal femur fractures (relative hazard = 2.02; 95% confidence interval, 1.77–2.32). The observation by Abrahamsen et al. (12) that subtrochanteric and diaphyseal fracture risks were similar among individuals receiving short-term (several months) and long-term (5–10 yr) alendronate treatment is somewhat reassuring, but the study was unable to specifically identify which, if any, of the fractures included were “atypical.” Contrary to the Abrahamsen results, a large retrospective cohort study from a California HMO linking pharmacy and radiographic data (13) found that among 15,000 femur fractures identified between 2007 and 2009, radiographic review identified 135 subtrochanteric and diaphyseal fractures with atypical features. Nearly all of the individuals with atypical femoral fractures had taken bisphosphonates (97%), and longer duration of use further increased the risk. Although presented only in abstract form and not yet peer reviewed, these preliminary data appear to support a causal relationship between bisphosphonate use and atypical femoral fractures. How will we ever make progress here? A very important step is to establish a rigorous definition of atypical femur fractures and to summarize the existing preclinical and clinical data supporting or refuting the relationship with bisphosphonate use. The deliberations from a Task Force organized by the American Society of Bone and Mineral Research (ASBMR) have just been published (14). The group of 28 experts reviewed the existing literature regarding atypical fractures and identified 310 cases reported to date. Although fraught with concerns about reporting bias and limited by inconsistencies in ascertainment, a qualitative summary of those cases reveals that most (291 or 94%) received oral or iv bisphosphonate with a median duration of 7 yr. Among cases where the information was sought and reported, 70% had prodromal symptoms (typically thigh pain), 28% had bilateral involvement, and many reported other medication use (corticosteroids in 34% and proton-pump inhibitors in 28%). Two thirds of the reported cases had osteopenia or normal BMD, and the handful of reports that measured bone turnover markers or vitamin D levels found that markers were usually within the premenopausal range and only 6% had serum 25-hydroxyvitamin D levels less than 20 ng/ml. Based upon these 310 cases, the ASBMR Task Force offers a preliminary case definition for atypical fractures of the femur (Table 1) and reviews potential mechanisms, such as oversuppression of remodeling. After reviewing these data, the Task Force concludes that: 1) bisphosphonates might be causally related, particularly with more than 5 yr of use; 2) the potential benefit for patients with indications for therapy far outweigh the potential risks; 3) oversuppression of bone turnover, as assessed by biochemical markers or even histomorphometry, is not consistently present; 4) definitive risk factors are unknown; and 5) an international registry is needed. Although it is a superb effort and a huge step in the right direction, the Task Force recommendations do not address or quantify how often one or more of the proposed major criteria are not present (such as minimal or no associated trauma, which is very difficult to reliably capture), nor do they sufficiently acknowledge that even comprehensive and publicly supported international registries will be terribly prone to reporting bias (atypical fractures among those taking bisphosphonates will be reported more often than atypical fractures among those not exposed to bisphosphonates). The Task Force recommendation to evaluate complaints of thigh pain among bisphosphonate-treated individuals is likely to lead to a substantial increase in unrevealing radiographic procedures, including expensive bone scan and magnetic resonance imaging tests, because nonspecific musculoskeletal complaints are very common and the condition being sought is very rare. Proposed criteria for a case definition of atypical femoral fractures (14 ) Proposed criteria for a case definition of atypical femoral fractures (14 ) At the end of the day, it seems likely that a causal relationship between prolonged bisphosphonate use and atypical femoral fractures will be established, and it is also likely that most confirmed cases will have one or more additional risk factors, such as coexisting diseases or concomitant medication use. Even so, because these atypical fractures are so infrequent, there will almost always be net benefit from bisphosphonate use among individuals at high risk for fracture (i.e. those with vertebral fracture or BMD T-scores below −2.5). Clinicians must now try to convey that idea while openly discussing the risk of atypical femur fractures in all patients treated with bisphosphonates. For now, clinicians need not radically change their prescribing habits and should continue to follow evidence-based treatment guidelines (15, 16), but the guidelines need to address the optimal duration of bisphosphonate therapy and the utility and mechanics of drug holidays, even if based upon limited data. Most importantly, clinicians, investigators, and regulatory authorities need to remind the public and each other that the potential benefits of bisphosphonate use greatly exceed the risks among those who qualify for antifracture therapy. Clearly, additional preclinical and mechanistic studies are needed, as are additional clinical studies that define both the prevalence and clinical risk factors for atypical femoral fractures among treated and untreated individuals. Such studies will be necessary to help us fine-tune the appropriate use of bisphosphonates and allow us to further maximize the benefits and minimize the risks. For article see page 5258 Disclosure Summary: D.C.B. has received research support from Amgen, Novartis, and Merck. Bone mineral density.