➢ The precise etiology and pathogenesis of osteonecrosis continue to be a source of controversy.
➢ Nontraumatic osteonecrosis runs through a pathway of endothelial dysfunction leading to critical ischemia and eventual bone death.
➢ Recent advancements in osteonecrosis research have led to the development of nonoperative treatment methods, including bisphosphonate therapy, the use of anticoagulation, and biophysical treatments.
➢ There is currently no high-level evidence for the use of any treatment, but statins and hyperbaric oxygen have been associated with the best results among nonoperative treatment options.
➢ An approach that involves a combination of nonoperative modalities with bone and joint-preserving procedures incorporating bone biology therapies may be considered.
Osteonecrosis of the femoral head, previously referred to as avascular or aseptic necrosis, is a progressive disease of young adults that produces considerable morbidity. Approximately 15,000 to 20,000 new cases of osteonecrosis of the femoral head are diagnosed every year in the United States1. Bilateral disease is present in 42% to 72% of cases2,3 and, without treatment, 59% of asymptomatic cases will progress to femoral head collapse4. Magnetic resonance imaging (MRI) has improved the rate of early detection of osteonecrosis of the femoral head, providing orthopaedic surgeons with time to intervene prior to end-stage joint destruction. Furthermore, an improved understanding of molecular mechanisms involved in osteonecrosis of the femoral head has led to the evaluation of preexisting medical therapies that halt and potentially reverse disease progression. The purposes of the present review are (1) to provide an overview of the current understanding of the development of osteonecrosis of the femoral head, (2) to review the presentation, diagnosis, and classification of precollapse osteonecrosis of the femoral head (Association Research Circulation Osseous [ARCO] stage-I and II osteonecrosis), and (3) to provide evidence-based recommendations based on the existing literature for medical and surgical interventions.
The pathogenesis of osteonecrosis of the femoral head may be divided into traumatic and nontraumatic causes. The pathogenesis of traumatic osteonecrosis is well understood, with prolonged femoral head dislocation causing an interruption of blood flow and with proximal femoral fractures interrupting blood flow through the formation of a compressive intracapsular hematoma. The pathogenesis of nontraumatic osteonecrosis is more complex and is divided into two competing theories: intravascular coagulation and, to a lesser extent, extravascular compression5. Intravascular coagulation and occlusion can occur as a result of local thrombus formation, the accumulation of abnormally shaped red blood cells, fat embolism, or nitrogen bubbles6. Extravascular compression arises secondary to damaged femoral head blood vessels7 that permit fat and blood to accumulate in the extravascular space.
Because of its association with intravascular coagulation, osteonecrosis of the femoral head has been linked to a variety of hypercoagulable conditions, including sickle cell disease, hereditary thrombophilia, antiphospholipid antibodies, malignancy, and inflammatory bowel disease6,8. Genetic mutations involving the coagulation cascade (protein C, protein S, plasminogen activator inhibitor-1)9-12, angiogenesis stimulation (nitric-oxide synthase, vascular endothelial growth factor)13,14, and bone formation (type-2 collagen A1, vitamin-D receptor)15 have been described in patients with osteonecrosis. Collectively, genetic associations probably share responsibility in cases of idiopathic osteonecrosis of the femoral head. However, their contribution appears to be overshadowed by the effects of exogenous agents including corticosteroids and alcohol.
Corticosteroid administration represents the most common cause of nontraumatic osteonecrosis of the femoral head16, increasing the risk of developing the disease twentyfold17. Evidence suggests that the vascular hypothesis18, whereby endothelial dysfunction and fatty emboli impair femoral head blood flow, is largely responsible. Corticosteroid administration has been shown to preferentially induce vasoconstriction19 and to increase procoagulant factor production, collectively inducing endothelial damage and intravascular coagulation18. The formation of fatty emboli is believed to arise through increased adipogenesis20, which in turn decreases osteogenesis21 and downregulates osseous repair and remodeling. Interestingly, similar changes have been found in association with alcohol consumption, albeit through a different underlying mechanism22. Unlike corticosteroids, alcohol has a dose-dependent effect on marrow and osteoblast function23. How these two exogenous agents lead to the same disease condition remains to be determined.
Regardless of the underlying cause, all forms of osteonecrosis of the femoral head culminate in a terminal pathway involving critical interruption of blood flow24, cell necrosis, and structural collapse. After the onset of ischemia, histological signs of marrow necrosis and osteocyte death become apparent within twenty-four to seventy-two hours25 and are detectable on MRI26. Saponification of free fatty acids with extracellular calcium subsequently occurs, leading to visible deposits on radiographs27,28. An inflammatory response is initiated by the surrounding tissues, leading to osseous remodeling29. Acellular trabecular bone is replaced with mechanically inferior woven bone27 that does not tolerate normal loading. Eventually, structural integrity fails, leading to osseous collapse, cartilage damage, and the development of osteoarthritis.
Patients with early-stage osteonecrosis of the femoral head are often in their third to fifth decade of life and commonly complain of pain localized in the groin, sometimes radiating to the anteromedial aspect of the thigh and/or knee. Pain may arise acutely or insidiously, can occur at night, and typically is described as deep, throbbing, and exacerbated by weight-bearing activities. Such patients can present with a variety of symptoms, ranging from slight discomfort to incapacitating pain. Medical and family history and medication use should be determined to assess for known risk factors (Table I)30.
Examination should focus on gait and hip motion. Active and passive range of motion may be limited by pain, with passive internal rotation being the most affected. Because of the high prevalence of bilateralism, examination of the contralateral hip is warranted31,32. Stigmata of chronic alcohol use (jaundice, scleral icterus, ascites, palmar erythema) and steroid use (truncal obesity, muscle weakness, easy bruising, thick skin, depression) should be noted.
Positive findings on history and physical examination should prompt additional diagnostic tests. Radiographic evaluation is critical and includes orthogonal radiographs of the affected hip. Subtle subchondral collapse in the weight-bearing zone is best appreciated on a frog-leg lateral view. Images should be scrutinized for irregularities in the femoral head and for the presence of the pathognomonic crescent sign, indicating impending structural collapse. Although laboratory tests are not routine, the erythrocyte sedimentation rate (ESR) and the C-reactive protein (CRP) level are usually elevated. The levels of acute-phase reactants such as von Willebrand Factor and Factor VIII are often abnormal.
MRI is the most reliable and widely used imaging modality. T1 and T2-weighted images should be made for both hips in order to screen for bilateral involvement. Computed tomographic (CT) scanning, bone scintigraphy, and bone biopsy are not performed routinely. The differential diagnosis includes, but is not limited to, subchondral fracture, chondrolysis, osteoarthritis with subchondral cysts, inflammatory arthritis, transient osteoporosis of the hip, and traumatic femoral head bone bruises.
The classification of osteonecrosis of the femoral head is established with use of radiographs and/or MRI. Establishing the disease classification determines the prognosis and assists in treatment planning. To date, there is no universally accepted classification system, with sixteen different systems having been described in the literature33. The most commonly used system is that of Ficat and Arlet34 (cited in 63% of studies), followed by the Steinberg35 (University of Pennsylvania) system (20%), the ARCO36 (Association Research Circulation Osseous) system (12%), and the Japanese Orthopaedic Association37 system (5%).
Despite its popularity, the Ficat system lacks intraobserver and interobserver reliability38 and does not consider the size and location of the necrotic area. The ARCO classification system, although also exhibiting poor to fair intraobserver and interobserver reliability, includes the parameters missing from the Ficat system to further assist in prognostication and treatment planning36 (http://arco-intl.org/Newsletters/Gardeniers-1993-5-2/classofosteon.jpg). ARCO stages 0, I, and II represent precollapse osteonecrosis of the femoral head and are targets for hip-preservation procedures and medical intervention. ARCO stages III and IV represent articular destruction and generally require salvage procedures, including hip resurfacing or arthroplasty.
Observation alone of osteonecrosis of the femoral head often results in progression and poor prognosis. Therefore, interest has arisen in the development of medical adjuncts designed to reverse disease mechanisms and to prevent femoral head collapse (Fig. 1). This section summarizes the findings related to several proposed medical therapies.
Activity Modification and Physical Therapy
Restricting weight-bearing reduces joint-reactive forces on the femoral head and theoretically promotes remodeling in patients with osteonecrosis of the femoral head. A meta-analysis of twenty-one studies involving 819 hips demonstrated no differences between full, partial, and non-weight-bearing groups39. Furthermore, only 22% of hips treated with non-weight-bearing had a satisfactory clinical outcome and 76% required arthroplasty or other salvage procedures40. These results indicate that a modified weight-bearing regimen does not delay progression.
Bisphosphonates inhibit the ability of osteoclasts to resorb bone by increasing osteoclast apoptosis and reducing osteoblast and osteocyte apoptosis, thereby reducing bone remodeling and turnover31. The benefits of bisphosphonates have been assessed in animal models41. Agarwala et al.42,43 evaluated 395 hips after a mean of four years of follow-up. Ninety-two percent of the patients had a satisfactory result without operative intervention, whereas failure requiring arthroplasty occurred in four (2%) of 215 Ficat stage-I hips, ten (8%) of 129 Ficat stage-II hips, and seventeen (33%) of fifty-one Ficat stage-III hips. Furthermore, a randomized controlled trial of fifty-four hips demonstrated significantly (p < 0.001) lower rates of disease progression and arthroplasty in hips that were treated with bisphosphonates for two months44. However, a multicenter, double-blinded, prospective study of sixty-five hips showed no significant difference between alendronate and placebo with regard to the rate of total hip arthroplasty, disease progression, or quality of life at two years of follow-up45.
Results are less favorable when bisphosphonates are combined with other modalities. The combination of alendronate, extracorporeal shock wave therapy, and hyperbaric oxygen demonstrated no difference in comparison with extracorporeal shock wave therapy alone46. The combination of extracorporeal shock wave therapy and alendronate also showed no increased benefit in comparison with extracorporeal shock wave therapy alone47.
It appears that bisphosphonates may play a role in improving the strength of osteoporotic or diseased bone48. However, this effect may be of minimal value in patients with osteonecrosis of the femoral head, as found in the largest Level-I study of which we are aware45. Prospective studies are warranted before the routine use of bisphosphonates for early-stage osteonecrosis of the femoral head can be recommended.
Statins reduce bone marrow adipocyte size and therefore could be beneficial in the treatment of osteonecrosis by reducing intraosseous pressure49. Animal models have demonstrated that statins protect against corticosteroid-induced osteonecrosis through pro-osteoblastic and anti-adipogenic effects on bone marrow40,50-54. However, to our knowledge, no clinical study has shown protective or therapeutic effects of statins in patients with osteonecrosis of the femoral head. Ajmal et al.51 retrospectively reviewed the records for 2881 patients who had undergone transplantation and were receiving high-dose steroids, and, although osteonecrosis of the femoral head developed in only 4.4% of patients who were receiving statins (compared with 7% of those who were not receiving statins), the difference was not significant. In summary, there is no high-level evidence to support a recommendation of routine use of statins in high-risk patients receiving corticosteroids.
The prostaglandin I2 analog iloprost causes systemic vasodilatation and inhibits platelet aggregation. Its use for the treatment of osteonecrosis of the femoral head followed from evidence of decreased bone marrow edema in the acetabulum and proximal part of the femur55,56. Disch et al. treated forty hips with iloprost over a twenty-five-month period. All patients had improvements in terms of pain and clinical function, and no patient exhibited osseous collapse or required surgery57. Jäger et al. noted improvement in the ARCO stage for twenty-four (42.9%) of fifty-six femoral heads that had not collapsed58,59. Additional studies are necessary to evaluate the use of iloprost for the treatment of early osteonecrosis and specifically to differentiate cases of osteonecrosis from cases of transient osteoporosis of the hip, a different condition that spontaneously resolves 80% of the time.
Glueck et al. evaluated the use of enoxaparin in hypercoagulable patients with Ficat stage-I or II osteonecrosis of the femoral head60. Twenty patients with presumed osteonecrosis of the femoral head of mixed etiology were compared with historic controls. At the time of the two-year follow-up, only one hip demonstrated progression, which compared favorably with historic controls. These findings were in keeping with those of a previously described rat model61. Additional study is required, particularly in this subset of patients with hypercoagulability.
Increasing oxygenation theoretically could be beneficial in the treatment of osteonecrosis of the femoral head as it has been shown to reverse cellular ischemia, to increase angiogenesis, and to reduce intraosseous pressure through vasoconstriction62,63. Reis et al.63 followed twelve patients with Steinberg stage-1 involvement who received hyperbaric oxygen. Ten hips (83%) reverted to normal MRI findings, and two hips demonstrated progression. Camporesi et al.62 also showed positive results in a prospective observational study of twenty Ficat stage-II hips, with pain decreasing following twenty oxygen treatments. At the time of the seven-year follow-up, all patients remained asymptomatic and did not require arthroplasty. Despite the positive results in those studies, the effect of hyperbaric oxygen in the treatment of osteonecrosis of the femoral head remains unclear and requires additional prospective evaluation of larger cohorts of patients.
Extracorporeal Shock Wave Therapy
Extracorporeal shock wave therapy, although developed for the purpose of breaking down symptomatic renal stones, has been incidentally observed to increase bone density around the pelvis. Furthermore, extracorporeal shock wave therapy increases osteoblastic activity, bone morphogenetic protein-2 (BMP-2), vascular endothelial growth factor (VEGF), and angiogenesis64. Extracorporeal shock wave therapy is believed to decrease inflammation, neovascularization, and bone formation, potentially preserving bone density, which may be beneficial in the treatment of osteonecrosis of the femoral head.
Wang et al.47 and Massari et al.65 found that extracorporeal shock wave therapy preserved the precollapse femoral head when used alone or with fibular grafting. Furthermore, a systematic review of five studies comprising 133 patients demonstrated encouraging results66. A survey of 219 European orthopaedic departments indicated that extracorporeal shock wave therapy is still not widely used67. Only 33% of respondents admitted to using the method routinely, whereas 67% advocated early operative treatment. This heterogeneity in usage is likely a reflection of limited benefits and the desire to act quickly before disease progression.
Initially described by Ficat as a method of acquiring biopsy specimens in order to make the diagnosis of osteonecrosis68, core decompression is the most commonly used procedure for the treatment of precollapse osteonecrosis of the femoral head69. Core decompression is believed to work by reducing elevated intraosseous pressure and restoring vascular inflow. Surprisingly, despite more than four decades of research, there is no general consensus regarding the indications for core decompression. The literature suggests that better results are associated with early treatment. Ficat’s original review of 133 hips demonstrated good results in 90% of patients with stage-I and II osteonecrosis, with minimal disease progression after a mean of 9.5 years68. Stulberg et al., in a randomized trial of fifty-five hips that were treated with core decompression or nonoperative measures, found improved outcomes in the operative treatment group, regardless of disease stage70. Furthermore, Mont et al. pooled findings from the literature and reported satisfactory clinical results for 63.5% of 1206 hips that were treated with core decompression, compared with 22.7% of 819 hips that were treated nonoperatively39.
The influence of modern techniques of core decompression remains unclear. While the procedure was initially performed with use of an 8 to 10-mm trephine, new techniques involve the use of single or multiple small drill holes, producing similar intermediate-term outcomes71,72. Using smaller drill holes permits easier access to the anterior portion of the femoral head, lowers the risk of iatrogenic cartilage injury, and decreases the possibility of a subtrochanteric fracture73. Furthermore, the insertion of adjuvants into the core decompression tract has been studied. In the study by Gangji et al.74, twenty-four Ficat stage-I or II hips were randomized to treatment with core decompression alone or core decompression combined with the injection of harvested iliac crest bone marrow cells that had been previously sorted and concentrated. After five years of follow-up, the femoral heads in the bone marrow group took significantly longer to show radiographic signs of osseous collapse (p = 0.038). However, the eventual timing of total hip arthroplasty for the femoral heads that did collapse was not significantly different between the groups. Zhao et al.75 harvested mesenchymal stem cells from subtrochanteric bone at the time of core decompression, expanded them in vitro, and injected them into the femoral head. Despite requiring a secondary procedure, injected hips exhibited less pain postoperatively and had a significantly lower rate of progression at five years postoperatively (p < 0.001). Although these modern techniques show promise, prospective follow-up with documentation of initial staging, progression, and time to hip arthroplasty are required.
The use of core decompression as a first-line treatment has been found to be a more cost-effective choice than observation76. A recent review77 examining only studies with validated outcome scores and disease staging demonstrated that core decompression produced the best results in hips without evidence of subchondral fracture or the crescent sign and with a small femoral head lesion. On the basis of these findings, we recommend utilizing core decompression as the first-line treatment for ARCO stage-I and II osteonecrosis of the femoral head.
Bone-grafting for the treatment of early-stage osteonecrosis of the femoral head is considered to be theoretically superior to core decompression because it provides structural support for the remaining subchondral bone and facilitates remodeling. Although there is no general consensus for its use, bone-grafting is recommended when there is <2 mm of subchondral bone depression, when <30% of the femoral head is involved, and when core decompression fails78. Several grafting techniques have been described. With no published prospective comparisons of these techniques, recommendations for their use are limited.
Phemister popularized the idea of filling the tract for a core decompression procedure with cortical autograft taken from the ilium, tibia, or fibula79. This technique recently was evaluated80 in a series of eighty hips that were treated with core decompression and insertion of either tibial autograft or fibular allograft. The autografts exhibited a significantly better survival time at six years compared with the allografts (75% compared with 48%; p = 0.002).
The light-bulb technique, introduced by Merle D’Aubigné et al.81, involves harvesting cancellous autograft from the iliac crest and inserting it through a cortical window at the femoral head-neck junction. A recent retrospective study of 110 stage-II hips with three to four years of follow-up demonstrated survival rates of 100% for hips that had involvement of <15% of the femoral head, 93% for those that had involvement of 15% to 30% of the femoral head, and 54% for those that had involvement of >30% of the femoral head82. The trapdoor technique, described by Mont et al.83, differs from the light-bulb technique in that autograft is inserted through a window in the femoral head cartilage. Studies evaluating this procedure have included only ARCO stage-III and IV hips83,84. Another technique involves the insertion of porous tantalum rods following core decompression. A retrospective review showed that treatment with tantalum rods yielded a considerably shorter operative time, less blood loss, and shorter hospitalization compared with vascularized fibular grafting, with equivalent survival time after the two procedures at two years of follow-up85. A four-year follow-up study demonstrated a survival rate of 68% for tantalum rods86, and a pooling of six studies revealed significantly better Harris hip scores in association with tantalum rods as compared with nonvascularized bone-grafting (p = 0.002)87. Evaluation of this procedure in long-term studies is necessary before recommendations can be made.
Vascularized bone-grafting for the treatment of osteonecrosis of the femoral head involves the use of local muscle-pedicle graft or vascularized fibular graft. The local graft involves the insertion of iliac crest bone into the core decompression tract and overlaying a muscle-pedicle graft either from the fascia lata anteriorly or from the quadratus femoris posteriorly88. A vascularized fibular graft involves transferring the central portion of the fibula and its nutrient artery to the hip. The fibula is inserted into the core decompression tract and is held with a Kirschner wire, and the vascular pedicle is anastomosed to recipient branches of the lateral femoral circumflex vessels. Commonly cited indications for vascularized grafting include no evidence of osseous collapse (ARCO stage-I and II involvement), articular collapse of <3 mm, and involvement of <50% of the entire head88.
When performed for the appropriate indications, vascularized grafting has been shown to be effective for improving hip function and delaying disease progression. Aldridge and Urbaniak reviewed the records for more than 2600 patients and found that only 11% of hips with ARCO stage-I and II osteonecrosis of the femoral head required subsequent total hip arthroplasty at five years88. Soucacos et al. found that 0% of ARCO stage-I and 5% of ARCO stage-II hips progressed to arthroplasty five years after fibular grafting89. This success has been found to be less predictable with larger lesions, as reflected by five-year arthroplasty conversion rates ranging from 17% to 57%90,91. Nevertheless, a recent study of sixty-five ARCO stage-I and II hips that were followed for more than ten years after vascularized grafting showed that the majority (75%) of the hips had survived, indicating that the procedure can produce long-term hip preservation92. Furthermore, vascularized grafts provide greater symptomatic relief and increased delay to reoperation93-95 in comparison with nonvascularized grafts. No significant differences have been reported when iliac crest grafts have been compared with vascularized fibular grafts96. The main limitations to vascularized grafting techniques include the requirement of expertise of microvascular surgery and donor-site morbidity. Although the rate of ankle problems has been reported to be as high as 20% following free fibular transfer97, several studies have demonstrated that sensory and motor symptoms tend to resolve over time and that symptoms of ankle instability are quite rare98-101.
On the basis of these findings, we recommend that vascularized grafting should be considered as a supplemental technique following core decompression failure and is suited for patients with ARCO stage-I and II involvement as well as select patients with ARCO stage-III involvement who have small lesions and minimal collapse.
Proximal Femoral Osteotomy
The goal of proximal femoral osteotomy is to offload the osteonecrotic segment of the femoral head, with attendant decrease in both intramedullary pressure and venous hypertension. Additional goals include preserving femoral head blood flow, preserving hip joint mechanics, and achieving rigid fixation at the osteotomy site. Although there is no consensus with regard to the indications for proximal femoral osteotomy, consistent results have been achieved in patients who are less than forty years old and have ARCO stage-II or III involvement, have a necrotic segment with a combined Kerboul angle of <200°, have no acetabular pathology, and have maintenance of normal hip range of motion31,102. When performed for patients with these indications, osteotomy has been associated with quality-of-life scores similar those associated with total hip arthroplasty103. Patients who do not meet these indications and who have corticosteroid-induced osteonecrosis of the femoral head tend to have poorer outcomes104.
Proximal femoral osteotomies that are dependent on the location of the necrotic segment have been described. Rotational transtrochanteric osteotomy is indicated for anterosuperior necrotic lesions, which are offloaded through anterior rotation of the proximal part of the femur. Positive results following this technically challenging procedure have come primarily from Japan, with a survival rate of 70% to 90% at three to eighteen years in selective patients with ARCO stage-II and III involvement105-108. Results outside of Asia have not been as positive, with five-year survival rates of <40%73. Although a lax posterior capsule in Asians has been suggested as an explanation for such geographic variability in results109, the lack of standardization of indications, surgical technique, and postoperative protocol is a more likely explanation. Intertrochanteric angular osteotomies have produced positive results in Europe and North America, leading to their preferential use in these regions. Valgus flexion osteotomy, which transfers the weight-bearing area medially and posteriorly, is indicated for patients with small anterolateral necrotic lesions. In a series of forty-eight patients, this procedure was associated with an improvement in the Harris hip score and an 87% success rate in terms of avoiding arthroplasty at five years postoperatively110. Meanwhile, varus osteotomy is indicated for patients with medially located necrotic lesions, a preserved lateral column, and a clinical examination demonstrating ≥30° of abduction. With use of these criteria, Mont et al. reported a 76% rate of good or excellent results in a study of thirty-seven ARCO stage-II and III hips that were treated with varus osteotomy and were followed for a mean of 11.5 years39. A lingering concern following proximal femoral osteotomy is the poorer outcomes that have been encountered with subsequent total hip arthroplasty, with increased rates of perioperative blood loss, operative time111, femoral shaft fracture, and component loosening112.
The lack of studies comparing osteotomies and various grafts makes it difficult to definitively place proximal femoral osteotomies within a treatment algorithm for osteonecrosis of the femoral head. However, on the basis of the current literature, we recommend proximal angular osteotomy for patients under the age of forty years who have ARCO stage-II or III changes, a combined Kerboul angle of <200°, normal range of motion, and no acetabular involvement. Osteotomies also should be considered for patients who have had a failure of previous core decompression and those who have necrotic lesions that cannot be adequately debrided and addressed with a vascularized graft.
Treatment of ARCO stage-I and II osteonecrosis of the femoral head remains controversial, with few high-level studies in the literature providing guidance for a treatment algorithm. Core decompression remains the most familiar and least morbid surgical procedure and currently is the first step in our treatment algorithm for precollapse osteonecrosis of the femoral head. Unfortunately, the remainder of the algorithm is less clear. Although vascularized grafts have been found to outperform nonvascularized grafts, donor-site morbidity remains a pertinent issue. The role of proximal femoral osteotomies is even less clear. Although such osteotomies are successful when performed for the right indications, the lack of comparative studies makes it difficult to determine whether osteotomies should be attempted prior to, instead of, or after vascularized grafts. Intertwined within this mélange of preservation procedures are pharmacological and nonpharmacological medical adjuvants, whose role in tandem with hip-preservation procedures remains untested and therefore cannot be currently recommended for clinical use (Table II). It is possible that medical adjuvants will play an essential role once early biological markers of osteonecrosis of the femoral head are found. Until that milestone occurs, prospective studies investigating specific treatment algorithms aimed at preserving the hip joint for as long as possible will have to suffice.
Source of Funding: No external funds were received in support of this work.
Investigation performed at McGill University Health Center, Montreal General Hospital, Montreal, Quebec, Canada
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. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, 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 © 2014 by The Journal of Bone and Joint Surgery, Incorporated