The nitrogen-containing bisphosphonates (N-BPs), stable analogs of the naturally occurring pyrophosphate, are the most widely utilized drug class for osteoporosis and the first line of treatment in metastatic bone disease. They selectively bind bone mineral and are ingested by osteoclasts to impair their ability to resorb bone. Therefore, in conditions where bone resorption exceeds bone formation, N-BPs slow down the ongoing bone loss and, in doing so, reduce the risk of fracture in osteoporosis and skeletal-related events in metastatic cancer. Equally effective is denosumab, a fully human monoclonal antibody inhibiting the receptor activator of nuclear factor B ligand (RANKL), a protein essential for osteoclast development, activity, and survival. However, an unusual atraumatic or minimal-trauma fracture, the atypical femoral fracture (AFF), has been reported with increasing frequency in bisphosphonate users since the first case reports were published in 2005 and were recently reported in denosumab users. This fracture occurs between the lesser trochanter and the femoral condyles and has specific radiological features (1, 2). It also occurs in bisphosphonateor denosumabnaive individuals and in patients on weak non-bisphosphonate antiresorptive agents (3, 4). Despite its rarity, there is considerable concern among physicians and patients regarding the long-term use of bisphosphonates for fracture prevention. The first suggestions of an association of AFFs with bisphosphonates came from case reports and case series followed by observational studies that, while being hypothesis-generating, remain fraught with limitations. Because no ICD code is yet available for low-trauma AFF, certain studies have failed even to separate low-energy from high-energy fractures. The indication bias is inherent within most, if not all, such studies. Patients using antiresorptives are, by indication, at a high risk of fracture, which does not allow the attribution of AFFs to antiresorptive use as opposed to the pre-existing osteoporosis. Also, there is inconsistency of defining AFFs, and there is no radiographic adjudication for a good number of them. Furthermore, a Dutch cohort study over 11 years reported that AFFs occurred with the same frequency in non-bisphosphonate users (5). Other epidemiological studies from Europe and North America have had no access to radiographs, preventing the identification of features that are required to classify a fracture as atypical. These studies were thus performed on all subtrochanteric fractures. The outcomes, to no surprise, are at best contradictory. The rarity of AFFs prohibits the conduct of randomized safety trials that could potentially determine whether the risk of such fractures is increased in bisphosphonate or denosumab users. The only “prospective” evidence comes from the reanalysis of data from over 14 000 patients enrolled in the alendronate Fracture Intervention Trial (FIT), FIT Long-term Extension (FLEX), and the Zoledronic Acid Once Yearly (HORIZON) trials. There was no increase in subtrochanteric and femoral shaft fracture risk, or indeed the risk of any fracture, in bisphosphonatetreated patients compared with the placebo group (6). This analysis has nonetheless been rightly criticized for its lack of access to patient radiographs and the relatively short duration of bisphosphonate exposure. It was also underpowered, considering the rarity of these fractures. There is no apparent dose response between bisphosphonates and AFFs. This is highlighted in cancer patients with bone metastasis who are regularly treated with bisphosphonates or denosumab. The yearly amount of these potent antiresorptive medications prescribed could be up
Read full abstract