➢ There is growing evidence that a strong relationship exists between the use of bisphosphonates and atypical femoral fractures. Most atypical femoral fractures can be treated successfully with intramedullary nailing and discontinuation of bisphosphonate therapy, with an expected delay in healing.
➢ Prodromal thigh pain and radiographic evidence of a radiolucent line (incomplete fracture) in patients with a history of atypical femoral fracture or chronic bisphosphonate use are strong indicators of an impending fracture, and prophylactic fixation should be considered.
➢ Prodromal thigh pain in a patient receiving chronic bisphosphonate therapy without radiographic evidence of incomplete fracture warrants magnetic resonance imaging for further investigation.
➢ The management of patients who have an incomplete fracture with minimal or no pain remains controversial. The authors of this review recommend a trial of conservative therapy, including protected weight-bearing, discontinuation of bisphosphonate therapy, calcium and vitamin-D supplementation (800 to 1000 IU per day), and possible teriparatide therapy. If there is no symptomatic or radiographic improvement after two to three months of conservative therapy, then prophylactic fixation is recommended.
➢ No recommendation currently exists regarding the duration of bisphosphonate therapy. However, the authors of the present report recommend that, on the basis of patient risk factors for future fracture, it may be optimal to discontinue bisphosphonate therapy after five years of continuous use to decrease the risk of associated complications.
Bisphosphonates are widely used as a primary treatment modality for osteoporosis. These agents prevent bone resorption via both the inhibition of osteoclast function and the promotion of osteoclast apoptosis1,2. The Fracture Intervention Trial Research Group investigated the effectiveness of bisphosphonates for the reduction of fragility fractures of the hip, distal part of the radius, and vertebrae3. Their research demonstrated that, compared with calcium supplementation alone, bisphosphonates with calcium supplementation decreased the prevalence of hip fractures, distal radial fractures, and vertebral fractures by 51%, 44%, and 46%, respectively. Additional indications for chronic bisphosphonate therapy include other metabolic bone diseases such as Paget disease, glucocorticoid-induced osteoporosis, and metastatic bone disease4. Unfortunately, despite the great utility of this class of medications, there has been growing evidence of an association between prolonged bisphosphonate use and atypical femoral fractures5-8. This review will focus on the definition, epidemiology, pathophysiology, diagnosis, and treatment of atypical femoral fractures related to prolonged bisphosphonate exposure.
Atypical femoral fractures are a distinct fracture type and should be distinguished from other more commonly encountered subtrochanteric and femoral shaft fractures. In 2010, The American Society for Bone and Mineral Research (ASBMR) formed a task force that defined atypical femoral fracture as an atraumatic or low-trauma fracture located in the subtrochanteric region or femoral shaft. Their definition excluded all fractures associated with high-energy trauma, femoral neck fractures, and intertrochanteric fractures with subtrochanteric extension. Major characteristics of atypical femoral fracture include location in the subtrochanteric region and/or diaphysis of the femur, association with minimal or no trauma, a transverse or oblique fracture orientation without comminution, and the presence of a complete fracture extending through both cortices or an incomplete fracture involving only the lateral cortex. Minor characteristics include localized periosteal reaction or “beaking” of the lateral cortex following fracture (Fig. 1), generalized cortical thickening of the femoral shaft, history of prodromal pain, bilateral fracture, and delayed healing9.
As of 2014, recent revisions to the ASBMR definition include minimal comminution as a major criterion, upgrading the presence of periosteal thickening from a minor to a major criterion, and defining the location of the fracture from distal to the lesser trochanter to just proximal of the supracondylar flare. In addition, there has been further characterization of the fracture line, which typically originates from the lateral cortex and is mainly transverse, although it may become oblique as it travels medially10 (Table I). The ASBMR task force also recommends that four of five major features must be present in order to make the diagnosis of an atypical femoral fracture, whereas no minor features are needed for this diagnosis10.
In the United States, the annual incidence of subtrochanteric and femoral shaft fractures ranges from twenty to thirty per 100,000 patient-years11,12. The vast majority of these fractures involve high-energy trauma in younger populations. In the osteoporotic population, most femoral fractures involve the femoral neck and/or intertrochanteric region, with subtrochanteric and femoral shaft fractures accounting for merely 3% and 5%, respectively9. Atypical femoral fractures have been associated with Asian descent, bilateral involvement, and prodromal pain in patients receiving bisphosphonate therapy8. Lo and colleagues reported that atypical femoral fractures were more common in Asian females (32% to 50% of cases) and were bilateral in 40% of cases13. Giusti et al., in a systematic review, found that prodromal pain was noted in 63% to 70% of cases and that the concurrent use of glucocorticoids and steroids was found in 34% and 39% of cases, respectively5.
In 2005, the first reports of bisphosphonate-related atypical fractures were published. Odvina et al. reported on four patients with atypical femoral fractures who had been receiving bisphosphonate therapy for a mean of 6.5 years14. Bone marrow biopsies were performed, and histomorphometric analysis revealed minimal callus formation and decreased bone turnover (estimated to be 100 times less than that of a normal postmenopausal woman). Furthermore, several retrospective studies have shown a relationship between bisphosphonates and atypical femoral fractures. Lenart et al. performed a retrospective review of forty-one fractures in postmenopausal women and found that 37% of subtrochanteric and femoral shaft fractures were associated with bisphosphonate use, compared with only 11% of femoral neck and intertrochanteric fractures15. Not surprisingly, the duration of therapy was longer for patients with subtrochanteric and femoral shaft fractures as compared with intertrochanteric fractures and fractures of the femoral neck. Furthermore, in a retrospective chart review, Girgis and Seibel reported an incidence of 2.3 to sixteen atypical femoral fractures per 100,000 patient-years, with an increased incidence in patients older than sixty-five years of age16.
In 2010, the ASBMR reported an annual incidence of atypical femoral fracture of two per 100,000 patient-years after two years of bisphosphonate exposure9. There was a dramatic increase in the frequency of these fractures after eight years of use, to seventy-eight per 100,000 patient-years annually, with most fractures occurring after five years of use. Schilcher and Aspenberg analyzed the Swedish national registry and reported the highest absolute risk of fifty cases per 100,000 patient-years, whereas Meier et al. retrospectively reviewed a single university hospital database and reported the lowest absolute risk of 3.2 cases per 100,000 patient-years17,18. However, in a similar study, when bisphosphonates were taken for a mean of eight to nine years, the absolute risk increased to 113.1 per 100,000 patient-years19. Therefore, although the incidence of atypical femoral fractures is very low in the osteoporotic population, this number drastically increases with prolonged bisphosphonate therapy. In its most recent report, the ASBMR task force stated that there is limited evidence to prove a causal relationship between bisphosphonates and atypical femoral fractures10. For example, some patients with atypical femoral fractures have no history of bisphosphonate use; however, an association between the two is convincing and is unlikely to be a result of unmeasured confounders.
The exact mechanism of atypical femoral fractures is unknown. However, several theories have been proposed, including microdamage accumulation, increased mineralization, reduced heterogeneity of mineralization, reduced rate of bone turnover, and alterations to the normal pattern of collagen cross-linking10. Atypical femoral fractures are thought to occur as a result of brittle failure and do not show signs of plastic deformation forces prior to failure as do other long-bone fractures in osteoporotic patients20. The radiographic features of atypical femoral fractures, including transverse orientation, general lack of comminution, and the nearly universal presence of periosteal reaction and cortical hypertrophy, support the notion that these fractures are stress fractures that fail because of the lack of remodeling potential and the brittleness of bone.
In atypical femoral fractures, the lateral femoral cortex is disrupted first as a result of tension. Tensile forces cause microdamage to the bone, which requires remodeling and repair to maintain the bone’s original strength. However, in patients undergoing bisphosphonate therapy, bone resorption is markedly reduced, affecting both bone turnover and remodeling potential. The result of this cascade is an accumulation of microdamage, which can be visualized in biopsy specimens21. In addition, bisphosphonates are preferentially delivered and accumulate at these areas of microdamage because of the local hyperemia that occurs in preparation for fracture-site remodeling and callus formation. This accumulation of bisphosphonates in turn prevents intracortical repair and may result in further propagation of the stress fracture. Interestingly, the impairment of fracture-healing decreases by up to 70% per year when bisphosphonate therapy is terminated17,22.
Donnelly et al., using Fourier transform infrared spectroscopy of biopsy samples, examined the mineral and collagen properties of cortical and cancellous bone of patients receiving bisphosphonate therapy23. The results showed a decrease in the heterogeneity of mineralization of cortical bone and an increase in the glycosylation of end products compared with bisphosphonate-naïve subjects. The decreased heterogeneity of mineralization of cortical bone and the surrounding matrix may lead to decreased bone strength and likely contributes to atypical femoral fractures. Furthermore, Ettinger et al. hypothesized that chronic bisphosphonate use resulted in a premature aging of the bone, rendering it more brittle and less likely to impede crack-tip progression24. The authors postulated that bisphosphonates were attracted to metabolically active bone sites, which preferentially suppresses the repair process at the crack site by decreasing the heterogeneity of the bone matrix. They also argued that it was not the accumulation of microcracks at the fracture site causing the impending fracture; instead, they put forth the theory of a “mother crack” that slowly progresses as a result of the altered protective effects of the bone caused by bisphosphonate use.
The biomechanics of the femur also play a pivotal role in atypical femoral fractures. Occurrences of these fractures between the lesser trochanter and the middle part of the femoral diaphysis are likely a result of the forces on the convex side of the bone. The lateral cortex of the subtrochanteric region of the femur has been shown to be under the greatest degree of tensile loading25. The occurrence of bilateral atypical femoral fractures at similar locations suggests a relationship between femoral geometry and the risk of fracture. Interestingly, femoral bowing is greater in Asian populations, suggesting that the increased mechanical stress on the lateral cortex may account for the higher incidence of fractures within this population26,27. A case series involving an Asian population demonstrated that patients who had atypical femoral fractures had greater curvature of the femoral diaphysis27. Another biomechanical factor is the tibiofemoral angle. Saita et al. reported that patients with a greater tibiofemoral angle were more likely to sustain an atypical femoral fracture in the diaphysis, whereas those with a smaller tibiofemoral angle were more likely to sustain such a fracture closer to the lesser trochanter28. More recently, Taormina et al. found a greater incidence of atypical femoral fractures in long-term bisphosphonate users with relatively varus neck-shaft angles, higher body mass indices, and narrower center-edge angles29. These biomechanical studies suggest that patient-specific variations in femoral geometry and biomechanical axis result in different mechanical forces on the femur and need to be considered when patients receiving bisphosphonate therapy present with pain. The difference between males and females with regard to the biomechanics and geometry of the hip and femur has not been investigated, but further research is needed as there are limited studies evaluating the geometry of the lower extremities and the role that it plays in the location and occurrence of atypical femoral fractures.
Atypical femoral fractures are classified as either complete or incomplete. Patients with complete fractures present similarly to those with typical femoral fractures in the acute setting. However, radiographs demonstrate thickening of the medial and lateral cortices; a transverse fracture starting laterally and extending medially, commonly with a medial spike; and/or a transverse and/or short oblique fracture with a lack of substantial comminution. As mentioned previously, four of five major criteria must be identified in order for a fracture to be classified as an atypical femoral fracture (Table I). Up to 70% of patients report prodromal pain preceding the completion of the stress fracture5. Patients presenting with atypical femoral fractures should be questioned about prodromal pain, pain in the contralateral extremity, and chronic bisphosphonate therapy6,8.
Incomplete atypical femoral fractures present greater diagnostic and management challenges. These fractures involve only the lateral cortex of the femur. This characteristic is important when considering an atypical femoral fracture because stress fractures in athletes and military recruits typically arise from the medial cortex30. Patients with incomplete fractures commonly present with thigh pain or radiographic evidence of stress fracture. Magnetic resonance imaging (MRI) is an important diagnostic tool when a patient presents with thigh pain without the typical radiographic features of the femur. MRI can be employed to better visualize a fracture line if one is not present on radiographs or to look for evidence of stress reaction. T1-weighted imaging shows diffusely decreased signal that is reflective of marrow edema resulting from water replacing the fatty components of the medullary cavity. Short tau inversion recovery (STIR) sequencing reveals increased signal secondary to inflammation and local hyperemia31. Dual x-ray absorptiometry (DXA) has also been evaluated as a diagnostic tool for the identification of atypical femoral fractures32. Kim et al., in a recent retrospective study of fifty-two atypical femoral fractures in forty-six patients, used prefracture DXA scans to look for lateral cortical changes. The overall rate of detection of atypical femoral fracture was 73% when both prodromal symptoms and DXA were considered, compared with 42% when only prodromal symptoms were considered33. Similarly, bone scintigraphy is useful for the diagnosis of atypical femoral fractures. Bone scintigraphy is diagnostic if it shows mild uptake and multifocal endosteal thickening of the lateral femoral diaphysis34. In summary, MRI should be performed for patients presenting with thigh and/or groin pain who are receiving bisphosphonate therapy and who have negative radiographs. If MRI cannot be performed because of contraindications, DXA or bone scintigraphy could be considered.
The treatment of atypical femoral fractures requires a defined protocol involving both medical and surgical measures. Fractures can be classified as either complete or incomplete, which will inevitably determine the treatment protocol. Once an atypical femoral fracture or impending atypical femoral fracture is identified, the first step is to immediately terminate bisphosphonate therapy. Dell et al. followed 126 patients with atypical femoral fractures and found that the prevalence of contralateral fracture was 53.9% when bisphosphonate therapy was continued for three years or more after the index fracture, compared with 19.3% when bisphosphonates were discontinued19. In addition to the termination of bisphosphonates, treatment with calcium and vitamin-D supplementation (800 to 1000 IU per day) should be started. Although not approved by the U.S. Food and Drug Administration for the treatment of atypical femoral fractures, the recombinant parathyroid hormone teriparatide (Forteo; Eli Lilly, Indianapolis, Indiana), which acts by stimulating osteoblasts while also increasing the absorption of calcium in the body, has been shown to shorten the time to healing in postmenopausal women with osteoporotic distal radial fractures when administered at a dose of 20 µg per day35. However, at a higher dose of 40 µg per day, no improvement was noted, which calls these results into question35. Chiang et al., in a recent prospective trial of fourteen consecutive patients, evaluated the use of teriparatide (20 µg per day) for the treatment of incomplete atypical femoral fractures36. Five patients opted for teriparatide therapy, and nine chose either operative or conservative management. Patients in the teriparatide group were shown to have a twofold to threefold increase in bone remodeling markers, with complete healing being noted in two (40%) of the five patients in the teriparatide group as compared with only one (11%) of the nine patients in the no-teriparatide group at one year of follow-up. Several other case reports have shown evidence of fracture-healing after discontinuation of bisphosphonate therapy and initiation of teriparatide37,38. In the absence of randomized controlled trials, the ASBMR was unable to make a definite conclusion with regard to the use of teriparatide for the treatment of atypical femoral fractures.
Evidence from the Fracture Intervention Trial Long-term Extension (FLEX) study suggested there was no significant difference in the rates of fragility fractures (other than vertebral fractures) when postmenopausal women who stopped bisphosphonate therapy at five years were compared with women who continued the medication for ten years3. Therefore, on the basis of a patient’s risk factors for future fractures, it may be optimal to discontinue bisphosphonate therapy after five years of continuous use to decrease the risk of associated complications.
Most orthopaedic surgeons recommend intramedullary nailing as the preferred method of treatment, although we are not aware of any randomized controlled studies that have compared intramedullary nailing with other types of constructs. It is common practice to treat complete atypical femoral fractures with full-length reamed cephalomedullary nailing in order to protect the entire length of the femur39.
The outcomes for patients who are managed surgically for complete atypical femoral fractures secondary to bisphosphonate use are a major concern. Weil et al. reviewed the records on seventeen bisphosphonate-related atypical femoral fractures in fifteen patients who were managed with intramedullary nailing and found that fracture-healing was poor, with only 54% of the fractures healing and with 46% of the patients requiring a second operation40. Similarly, Prasarn and colleagues, in a study in which the outcomes of intramedullary nailing of atypical femoral fractures were compared between patients receiving bisphosphonates and those without a history of bisphosphonate exposure, found that patients receiving bisphosphonates had a significantly higher rate of intraoperative fracture, malunion, nonunion, implant failure, and periprosthetic fracture41. In a more recent multicenter series, the revision rate was 9% among patients with complete displaced atypical femoral fractures, with 20% of the patients having fracture-healing after six months42. Conversely, Ha et al. reported complete union of ten atypical femoral fractures after initial internal fixation43. Lin et al., in a study of atypical femoral fractures that were treated with surgical nailing, reported that the time to healing was shorter for patients who were managed with cessation of bisphosphonates and addition of teriparatide than for those who were managed with continuation of bisphosphonate therapy (mean, 4.5 compared with 23.5 months)44. These data indicate that atypical femoral fractures associated with bisphosphonate therapy have a longer time to union than normal subtrochanteric and femoral shaft fractures do. It also highlights the importance of terminating bisphosphonate therapy at the time of the index fracture in order to promote fracture-healing.
We recommend that all patients who have sustained an atypical femoral fracture should be evaluated with radiographs of the contralateral femur to assess for signs of involvement or impending fracture. The current ASBMR recommendation for patients with an incomplete fracture with minimal or no pain is a trial of conservative therapy, including protected weight-bearing, discontinuation of bisphosphonate therapy, calcium and vitamin-D supplementation, and, possibly, teriparatide therapy9,10. Saleh et al. found that fractures without evidence of a radiolucent line on radiographs responded well to conservative treatment, with a union rate of 100% (five of five)45. However, when a radiolucent line was present, the union rate after conservative therapy was only 22% (two of nine), and the majority of fractures required operative fixation. Similarly, Banffy et al. reported that five (83%) of six incomplete fractures that were associated with a radiolucent line failed at a mean of ten months46. In addition, patients for whom conservative therapy failed were found to have hospital stays that were three days longer than those who were managed with prophylactic fixation.
Outcomes following the fixation of incomplete atypical femoral fractures are impressive25. In a recent cohort study of thirty-seven patients with forty-one incomplete atypical femoral fractures who had been receiving bisphosphonate therapy for an average of nine years, Egol et al. found that 81% of the patients were pain-free and 100% displayed fracture union at an average of seven months following intramedullary nailing47. On the other hand, patients who were managed conservatively had an 18% rate of union and only 64% reported being pain-free at the time of the latest follow-up (average, eleven months). Functional outcomes were also significantly better following surgical treatment. This evidence suggests that prophylactic nailing of incomplete atypical femoral fractures provides more reliable healing, enhanced functional outcomes, and higher rates of pain cessation at earlier intervals in comparison with conservative treatment.
Bisphosphonates are used worldwide as a primary treatment modality for osteoporosis and fracture prevention. Unfortunately, there is a growing body of evidence that demonstrates a strong relationship between the use of bisphosphonates and atypical femoral fractures. Although the incidence of atypical femoral fractures is very low in the osteoporotic population, this number increases drastically with prolonged bisphosphonate therapy. The radiographic features of atypical femoral fracture, including transverse (or transverse-oblique) orientation, general lack of comminution, and nearly universal presence of periosteal reaction and cortical hypertrophy, support the notion that these fractures are stress fractures that fail due to a lack of remodeling and the brittleness of bone.
Patients presenting with atypical femoral fractures should be questioned about prodromal pain, pain in the contralateral extremity, and chronic bisphosphonate therapy. Treatment should include both surgical and medical measures, including cephalomedullary nail fixation to protect the entire femur, cessation of bisphosphonate therapy, calcium and vitamin-D supplementation, radiographs of the contralateral femur, and consideration of teriparatide. Multiple studies have demonstrated higher complication rates following atypical femoral fractures associated with bisphosphonate therapy and significantly longer times to union in comparison with normal subtrochanteric and femoral shaft fractures. Patients presenting with an incomplete fracture and groin and/or thigh pain should be managed with surgical prophylaxis as studies have shown superior outcomes in association with surgical management. Patients who have an incomplete fracture associated with minimal or no pain should be managed with a trial of conservative therapy, including protected weight-bearing, discontinuation of bisphosphonate therapy, calcium and vitamin-D supplementation, and possible teriparatide therapy. If there is no symptomatic or radiographic improvement after two to three months of conservative therapy, then surgical prophylaxis is recommended. We recommend the use of specific algorithms for the treatment of atypical femoral fractures and the evaluation of the contralateral femur in patients with atypical fractures (Fig. 2 and Fig. 3).
Bisphosphonates should continue to be used as first-line treatment for osteoporosis and fracture prevention. However, increased physician awareness of the relationship between prolonged bisphosphonate use and impending atypical femoral fractures is paramount. Increased communication and comanagement between prescribing primary care physicians, pharmacists, and orthopaedic surgeons is essential to prevent future impending and complete atypical femoral fractures.
Source of Funding: No external funds were received in support of this manuscript.
Investigation performed at the Department of Orthopaedics, Warren Alpert Medical School of Brown University, Providence, Rhode Island
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2015 by The Journal of Bone and Joint Surgery, Incorporated