➢ Percutaneous pinning, while technically demanding, provides a minimally invasive approach to the treatment of select displaced two, three, and four-part fractures, with reliable results.
➢ The advent of locked plate fixation has expanded the indications for open reduction and internal fixation of displaced proximal humeral fractures but still requires anatomic reduction and appropriate implant placement in order to avoid complications. The incorporation of tension band rotator cuff sutures into the plate construct may augment tuberosity and fracture fixation.
➢ Anatomic tuberosity position and healing, along with the use of a dedicated fracture implant, are key factors for the successful treatment of proximal humeral fractures with use of hemiarthroplasty.
➢ Reverse shoulder replacement is an alternative to hemiarthroplasty for the treatment of severe fractures for which arthroplasty is indicated, especially in elderly patients (those over the age of seventy-five years), but it requires tuberosity healing for optimum function.
Proximal humeral fractures are common in the elderly population, and the societal burden is expected to increase. These fractures represent 4% to 5% of all fractures and one-third of fractures among patients over the age of sixty years1,2. Epidemiological studies have indicated that the number of such fractures in women who are eighty years of age or older may nearly double within the next twenty years3.
In his original description of proximal humeral fractures, Neer affirmed that most fractures are nondisplaced or minimally displaced and therefore can be treated nonoperatively4. Displaced fractures represent only 15% to 20% of all proximal humeral fractures4. Decision-making is influenced by many factors, including fracture pattern, comminution, bone quality, surgeon preferences, and the age and activity level of the patient5. Numerous options are currently available to the surgeon for the treatment of displaced fractures, including percutaneous fixation, open reduction and internal fixation, and arthroplasty6,7. This review will briefly describe and discuss the outcomes of current options for the treatment of proximal humeral fractures.
Pathoanatomy and Classification
The proximal humeral anatomy as it relates to fractures and their treatment can be divided into distinct anatomic “parts” as described by Neer4. These parts include the articular segment, the greater and lesser tuberosities and their cuff insertions, and the shaft. An understanding of the normal anatomy of the proximal part of the humerus, including inclination, version, and tuberosity-head and tuberosity-tuberosity relationships, is critical to identifying displaced fractures and understanding pathoanatomy. Most fractures are two-part fractures involving the surgical neck, with progressive displacement of the tuberosities as the severity of injury increases. Three-part fractures are characterized by displacement of either the greater or lesser tuberosity from the articular segment and shaft, and four-part fractures demonstrate displacement of both tuberosities from the articular segment and shaft. The fracture plane, which separates the greater tuberosity from the lesser tuberosity, is most commonly found posterior to the bicipital groove and is an important landmark as the anterior vascular supply to the humeral head courses along the bicipital groove and usually is spared coursing just anterior to this fracture plane8,9. Although these vessels may be spared in many fracture patterns, osteonecrosis remains a common consequence of proximal humeral fractures, with rates ranging from 3% to 37% in association with two and three-part fractures and with higher rates observed in association with more complex and comminuted injuries10. It is critically important to appreciate the likely risk of osteonecrosis on the basis of the fracture pattern seen on radiographic imaging. Hertel et al. devised a method of assessing the risk of osteonecrosis and concluded that the critical components include the integrity of the medial hinge and the length of the posteromedial metaphyseal head extension11. Fractures resulting in disruption of the medial hinge and those with <8 mm of calcar bone attached to the articular segment were more likely to be associated with the development of osteonecrosis. Displaced fracture patterns that may warrant special consideration include the valgus-impacted four-part fracture, in which the medial calcar connection of the shaft to the articular segment is potentially preserved, lowering the rate of osteonecrosis and allowing the surgeon greater flexibility in terms of treatment options without resorting to arthroplasty (Fig. 1)12,13. Varus-angulated fractures, especially those associated with calcar comminution, are inherently unstable, have been shown to be difficult to stabilize, and are prone to redisplacement14.
An understanding of the neural anatomy, especially the course of the axillary nerve, as it relates to the pathoanatomy is also critical to the successful treatment of proximal humeral fractures. While rare, injuries to the brachial plexus, especially the axillary nerve, are observed following these injuries, and it is critical to assess the neurological function of the injured arm prior to determining the final course of treatment.
Imaging assessment begins with a standard series of radiographs, including anteroposterior, true anteroposterior, axillary lateral, and scapular-Y radiographs of the proximal humeral fracture. Computed tomography (CT) can provide additional information for both classification and preoperative planning15. Additionally, CT also may be useful for the evaluation of fractures associated with articular surface involvement and for the enumeration of fracture fragments.
Nonoperative treatment is indicated for fractures with minimal displacement or angulation of the “parts,” such as impacted two-part surgical neck fractures. The fracture is immobilized in a sling for two to four weeks, followed by a graduated physical therapy regimen beginning with passive range of motion16,17. Active range of motion typically is instituted at six weeks, followed by strengthening modalities at three months. However, the fracture must be carefully followed as displacement or progressive angulation can occur, especially in patients with varus fractures, and, if missed, can yield suboptimal outcomes18,19 (Fig. 2 and Fig. 3). We recommend serial evaluation and radiographic assessment at one, two, and four weeks after the injury to carefully monitor fracture orientation and healing prior to the initiation of therapy. Full recovery following a fracture can take up to nine months or a year.
Operative treatment of displaced two, three, and four-part fractures is indicated to maximize functional outcome and pain relief. The goal of operative treatment is to restore the normal proximal humeral anatomy, including the version of the articular segment, the tuberosity relationships, and the neck-shaft angle. Fixation options include percutaneous techniques with pins or screws and open reduction and internal fixation with contoured locking plates or intramedullary fixation. Fractures that are not amenable to operative fixation, including comminuted fractures, head-split injuries, and fractures associated with high rates of osteonecrosis, are often indications for reconstruction with either hemiarthroplasty or reverse total shoulder arthroplasty.
Closed reduction and percutaneous pinning is a minimally invasive technique that has been associated with reliable healing and low rates of osteonecrosis when used for the treatment of select two, three, and four-part proximal humeral fractures20-22 (Table I). While this technique is technically demanding, the minimal soft-tissue dissection may be associated with less stiffness or scarring, allowing for excellent functional outcomes after fracture-healing. However, several biomechanical studies have demonstrated that pin fixation is not as robust as open reduction and internal fixation with use of a plate, highlighting the importance of proper operative indications23-25.
Fractures that are amenable to closed reduction and percutaneous pinning include select two-part surgical neck fractures (those without varus collapse or calcar comminution) and greater tuberosity fractures, three-part fractures (although such fractures may be the most challenging pattern to treat percutaneously because the humeral head not only is displaced but also is malrotated), and valgus-impacted four-part fractures. For closed reduction and percutaneous pinning to be successful, several conditions must be satisfied: (1) satisfactory reduction, (2) sufficient bone quality (without osteoporosis), and (3) minimal comminution, with an intact medial calcar.
Preferred Operative Technique
A detailed description of the operative technique for closed reduction and percutaneous pinning has been published previously21.
Proper positioning is the first step of successful closed reduction and percutaneous pinning. The patient is placed in a beach-chair or supine position so that adequate images, including a true anteroposterior (Grashey) radiograph, a scapular lateral radiograph, and an axillary radiograph, can be made. Fracture reduction is obtained by means of traction and closed manipulation, but for three and four-part fractures a small fracture-reduction portal can be helpful for manipulating fragments with elevators or pins used as joysticks26.
After fracture reduction, the fracture is secured with two to three 2.8-mm terminally threaded pins that are placed in a retrograde fashion from the anterolateral aspect of the humeral shaft. Retrograde pins are inserted distal to the axillary nerve path in the deltoid and are placed through a protective sleeve to avoid nerve injury. The pins are placed under image-intensifier guidance and should be advanced by means of hand power to reduce the risk of inadvertent subchondral penetration. Fixation of the tuberosities is performed with cannulated screws that are placed from the top of the tuberosity into either the calcar or the humeral head22. The pins are cut beneath the skin, and the arm is immobilized postoperatively for the first three to four weeks (Fig. 4 and Fig. 5). The retrograde pins are removed four weeks after the operation in the operating room with the patient under sedation and local anesthesia. Progressive active-assisted range of motion is initiated after pin removal. Strength training may be instituted once radiographic union has been achieved, typically at twelve weeks.
The outcomes of closed reduction and percutaneous pinning are directly correlated with the adequacy of fracture reduction, and the union rates associated with this method are comparable with those associated with other fixation techniques5,20-22. Keener et al. reported on twenty-seven patients at a minimum of one year after closed reduction and percutaneous pinning21. The fractures included seven two-part, eight three-part, and twelve valgus-impacted four-part proximal humeral fractures27. At the time of the latest follow-up, the average elevation was 142° and the average external rotation was 51°. Four patients had healing with malunion, and four had development of posttraumatic arthritis. Fracture type, age, malunion, and osteoarthritis had no significant influence on measured outcomes. The osteonecrosis rate was 3.7%.
Harrison et al. recently reported the intermediate-term results for these same patients at an average of fifty months (range, eleven to 101 months) postoperatively5. An increased rate of osteonecrosis was observed at the time of this later follow-up (26%, seven of twenty-seven), especially among patients with four-part fractures (50%, five of ten). Patients with osteonecrosis demonstrated lower ranges of motion (average forward elevation, 144° compared with 129°; average external rotation, 44° compared with 33°) and a lower American Shoulder and Elbow Surgeons (ASES) score (average, 77 compared with 84) (p = 0.26). Although a higher rate of osteonecrosis was observed at the time of this later follow-up, only three of seven patients underwent hemiarthroplasty.
In a similar study, Resch et al. reviewed the records for twenty-seven patients with three-part fractures (nine patients) and four-part fractures (eighteen patients) after an average duration of follow-up of twenty-four months22. All three-part fractures went on to union, and the average Constant score at the time of the latest follow-up was 91%. Osteonecrosis was observed in association with two (11%) of the eighteen four-part fractures, and both patients required revision to an arthroplasty.
Locked Plate Fixation
Open reduction and internal fixation of proximal humeral fractures historically has been fraught with complications and the need for revision surgery. Many of these fractures occur in elderly patients with osteoporotic bone, making stabilization difficult and leading to screw loosening or pullout as well as unsatisfactory clinical outcomes27. Recent biomechanical data on locked plate fixation of the proximal part of the humerus have demonstrated improved construct stability, with a lower rate of implant failure28-32 (Table I). Lescheid et al.33 found that removal of medial cortical support resulted in a decrease in axial, torsional, and shear stiffness at all varus angulations of fractures, whereas Gardner et al.34 concluded that achieving mechanical support in the inferomedial region of the proximal part of the humerus resulted in improved maintenance of fracture reduction (Fig. 6).
Locked plate fixation is commonly indicated for displaced or angulated fractures that are amenable to fixation, including osteoporotic and comminuted fractures, representing an evolution in technique compared with historical options. Fracture patterns that are associated with high rates of osteonecrosis, such as classic Neer four-part fracture-dislocations or head-split fractures, also may be treated with locked plate fixation35. Varus fractures with medial calcar comminution are particularly suitable for treatment with locked plate fixation because of the potential for these fractures to redisplace into varus.
Preferred Operative Technique
Multiple operative approaches, including deltopectoral, anterolateral, deltoid-splitting, and two-incision techniques, have been described43-48. Although some surgeons have advocated the use of a deltoid-splitting approach, those approaches may increase the risk of injury to the axillary nerve44-46. In one series, the use of a deltoid-splitting approach was found to result in lower Constant scores49, decreased abduction, and decreased forward elevation in comparison with the use of a deltopectoral approach47. We prefer the use of the deltopectoral approach for the open treatment of fractures.
After fracture exposure, the rotator cuff may be tagged with heavy, nonabsorbable sutures in the supraspinatus, infraspinatus, and subscapularis insertions to aid in the reduction of fracture fragments37,39,40,50. These sutures are later incorporated into the plate to augment tuberosity fixation and to help neutralize the deforming forces of the rotator cuff. The head is anatomically reduced via traction or is manipulated with elevators or joysticks. The fracture plane between the lesser and greater tuberosities provides good access to the head for reduction. Once the head is anatomically reduced, it is provisionally held with Kirschner wires and the tuberosities are then reduced with aid of the cuff sutures51,52.
After anatomic reduction has been confirmed with fluoroscopic imaging, the locking plate is applied. The plate may be positioned just lateral to the bicipital groove and inferior to the top of the greater tuberosity to prevent plate impingement in the subacromial space. When proximal locking screws are being placed into the humeral head, care should be taken to prevent subchondral screw penetration. We routinely shorten screws by approximately 4 mm to prevent later joint penetration37,41,47,50,53-56. A small arthrotomy of the rotator interval at the base of the coracoid process can allow the surgeon to palpate the humeral articular surface to ensure that no screws have penetrated the joint. Orthogonal fluoroscopic images are used to further evaluate screw position57. Most studies have demonstrated that fractures with varus malangulation are at higher risk for prosthetic failure and worse outcomes. For these fractures, the orientation of a so-called calcar screw along the inferior aspect of the head has been shown to be biomechanically important to ensure optimum fracture stabilization34.
Postoperatively, the arm should be placed in a sling. Passive exercises are initiated during the first week postoperatively. Active shoulder motion is begun six weeks postoperatively. Strengthening exercises should be initiated twelve weeks postoperatively to allow for adequate healing.
Most studies of locked plate fixation of proximal humeral fractures have demonstrated improved results, with lower rates of fixation failure and complications in comparison with historical controls41,51,55,58. However, locked fixation does not obviate the need for three critical factors: (1) anatomic fracture reduction with calcar support, (2) appropriate placement and length of implants, and (3) augmentation of fracture fixation with rotator cuff sutures incorporated into the plate.
Fankhauser et al. reported their early results following the use of locking plates in a series of twenty-nine proximal humeral fractures38. Most of these fractures were complex, with fifteen AO Type-B and nine Type-C fractures. The average Constant score at one year was 74.6. Simple fracture patterns (two-part fractures) were associated with the best outcomes (average Constant score, 82.6). Redisplacement of the fracture occurred in four patients, and screw cutout occurred in three patients, all of whom had varus fractures.
Solberg et al. retrospectively reviewed the records for seventy patients who had been managed with locked plate fixation of three and four-part proximal humeral fractures56. After an average of three years of follow-up, the average Constant score was 71. The complications of osteonecrosis, humeral head perforation, loss of fixation, tuberosity displacement, and varus subsidence occurred in association with nineteen (79%) of the twenty-four fractures in the varus group, compared with nine (20%) of the forty-six fractures in the valgus group, further highlighting the challenges associated with varus fractures.
More recently, Ricchetti et al. reported improved results and a decreased complication rate in a retrospective case series of fifty-two patients (fifty-four shoulders) after an average duration of follow-up of thirteen months40. The authors utilized rotator cuff sutures to aid in fracture reduction and to augment fixation. Postoperatively, the average forward elevation was 130°, the average external rotation was 28°, and the average ASES score was 70.8. Osteonecrosis developed in one patient, and an asymptomatic varus malunion occurred in five patients. There were no cases of implant failure or nonunion.
Primary arthroplasty for the treatment of three and four-part proximal humeral fractures was first advocated by Neer4. With the recent advent of locking plate technology, the indications for the use of arthroplasty for fracture treatment have continued to narrow. Although hemiarthroplasty consistently has been shown to provide pain relief, functional outcomes have been modest and inconsistent59, leading some surgeons to consider reverse shoulder arthroplasty as an alternative treatment.
Hemiarthroplasty is indicated for fractures for which internal fixation may be unsuccessful, such as fractures with severe articular incongruity (head-split fractures), comminuted fractures, and fractures associated with a substantial risk of osteonecrosis (four-part fracture-dislocations)59-63. Hemiarthroplasty requires anatomic restoration of the humeral head height, version, and tuberosity position and subsequent healing of the tuberosities for acceptable functional outcome (Fig. 7). Boileau et al. identified several key preoperative factors as predictors of poor outcome following hemiarthroplasty: female sex, an age of more than seventy-five years, and initial displacement of the greater tuberosity60.
Preferred Operative Technique
A detailed description of hemiarthroplasty for the treatment of proximal humeral fractures was published previously64.
Tuberosity healing and positioning are two of the most critical factors for successful hemiarthroplasty (Table I). The reported rate of tuberosity malunion or nonunion following hemiarthroplasty has ranged from 4% to 50%60,61,65. Boileau et al., in a retrospective review, found that tuberosity malpositioning occurred in thirty-three (50%) of sixty-six patients and was correlated with an unsatisfactory result, superior migration of the prosthesis, and persistent pain60. Tuberosity malpositioning was associated with poor initial positioning of the prosthesis (in terms of height and version), poor positioning of the greater tuberosity, and advanced age (especially in women over the age of seventy-five years).
Incorrect placement of the stem, either too high or too low, results in improper tensioning of the rotator cuff musculature. Substantial lengthening of >1 cm has been shown to result in tuberosity detachment and component migration60. In an effort to provide landmarks for implant height, Murachovsky et al. validated the use of the pectoralis major tendon insertion as an intraoperative landmark66. In a cadaveric dissection study of forty shoulders, the authors found that the pectoralis major tendon inserted a mean (and standard deviation) of 5.6 ± 0.5 cm distal to the top of the humeral head. The greater-tuberosity-to-humeral-head height also can be used to recreate that anatomical relationship. This relationship should be within 5 to 10 mm to improve functional outcomes60,62. In addition, preoperative templating with a full-length radiograph of the uninjured humerus has been suggested as a way to determine the height of the fracture stem67,68. Restoration of the native humeral version is also important for tuberosity healing and implant stability69,70. Excessive retroversion can place tension on the greater tuberosity during internal rotation, increasing the potential of tuberosity failure60.
Tuberosity fixation should incorporate cerclage sutures through the rotator cuff insertions of the tuberosities, which are passed through the implant stem and the humeral shaft. Frankle et al. showed that tuberosity reconstruction should incorporate the use of a circumferential cerclage suture that is passed medial to the humeral stem and around the tuberosities in order to reduce interfragmentary strain and displacement of the tuberosities71.
Postoperatively, the arm is placed in a sling in neutral rotation or slight external rotation. The initiation of passive motion therapy is dictated by tuberosity integrity and fixation and is often initiated within the first week, although some surgeons recently have advocated immobilization in a neutral rotation orthosis for long as four weeks in order to aid tuberosity healing72. Active-assisted motion is initiated at six weeks, and resistance and strengthening exercises are started at twelve weeks.
The outcomes of hemiarthroplasty for fracture treatment have been variable. Goldman et al. reported on the outcomes of twenty-two hemiarthroplasties that had been performed for the treatment of acute three and four-part fractures59. Active forward elevation averaged 107°, and active external rotation averaged 31°. Seventy-three percent of the patients reported slight or no pain. The authors concluded that hemiarthroplasty provides consistent results with regard to pain relief but that the results were less predictable in terms of return of function and range of motion. Similarly, Kontakis et al. performed a systematic review of sixteen studies involving 810 hemiarthroplasties73. The average age at the time of the operation was 67.7 years, and the average duration of follow-up was 3.7 years. Most of the fractures were four-part fractures or fracture-dislocations. The average active forward elevation was 105.7°. Poor tuberosity healing was reported in eighty-six (11.15%) of 771 cases. The authors concluded that the functional outcome after hemiarthroplasty is bimodal (either good or bad) and is dependent on tuberosity healing.
The importance of tuberosity healing was supported in studies by Mighell et al.62 and Krishnan et al.74. Mighell et al.62, in a retrospective review of eight shoulders, reported a tuberosity union rate of 96%, an average active forward elevation of 128° (range, 45° to 180°), and an average active external rotation of 43° (range, 0° to 80°). Krishnan et al.74, in a study involving a fracture-dedicated prosthesis and meticulous tuberosity repair, reported a tuberosity healing rate of 79% and an average active forward elevation of 129°.
Reverse Shoulder Arthroplasty
The use of reverse shoulder arthroplasty for the treatment of proximal humeral fractures was initiated as a result of the unpredictable functional outcomes reported in many studies of hemiarthroplasty. Although reverse shoulder arthroplasty may provide more predictable forward elevation, rotational function still requires tuberosity healing. Furthermore, implant survival, which was shown to be good following hemiarthroplasty in the study by Robinson et al.75, is less predictable following reverse shoulder arthroplasty, and some studies have demonstrated decreasing implant survival at 120 months, indicating that reverse shoulder arthroplasty should be reserved for the elderly patient76. Furthermore, reverse shoulder arthroplasty also is associated with unique complications, such as scapular notching, acromial fracture, and a higher rate of implant instability77-83. The rate of scapular notching has ranged from 44% to 96% in studies involving the Grammont-type reverse total shoulder prosthesis77,80-83 but may be lower for prostheses with a lateral offset of the center of rotation or a more varus neck-shaft angle compared with the classic Grammont design84,85.
Reverse shoulder arthroplasty is indicated for the treatment of three or four-part proximal humeral fractures in elderly patients who have risk factors for poor outcome following hemiarthroplasty (Table I). Predictors of poor outcomes following hemiarthroplasty include substantial tuberosity comminution, osteopenia, known rotator cuff deficiency, and an age of more than seventy-five years60.
Preferred Operative Technique
Reverse shoulder arthroplasty can be performed through either a deltopectoral approach or a superior approach. To preserve the deltoid, we prefer the deltopectoral approach. As with hemiarthroplasty, it is important to achieve the correct height and version of the implanted prosthesis and to meticulously repair the tuberosities to enable rotational function of the shoulder. Unlike hemiarthroplasty, the goal is not to obtain the native retroversion. Many prostheses are designed to be placed in 0° to 20° of retroversion to maximize shoulder rotation86. To avoid scapular notching, the glenosphere baseplate is implanted as inferior on the native glenoid as possible as well as in neutral or inferior tilt. The technique of performing reverse shoulder arthroplasty has been extensively reviewed in the literature87-89 (Fig. 8).
Most studies on the use of reverse shoulder arthroplasty for the treatment of fracture have demonstrated a fairly uniform functional outcome, without a large discrepancy between patients, as compared with the bimodal distribution of functional outcome observed following hemiarthroplasty59,73,74,90-92. Stated in a different manner, functional outcome may be more predictable following reverse shoulder arthroplasty, albeit perhaps not quite as good as that following hemiarthroplasties in which the tuberosities are anatomically healed.
Gallinet et al., in a retrospective study of forty patients with three and four-part fractures, compared the results of reverse shoulder arthroplasty (nineteen patients) with those of hemiarthroplasty (twenty-one patients)90. Seventeen patients in the hemiarthroplasty group were followed for a mean of 16.5 months, and sixteen patients in the reverse shoulder arthroplasty group were followed for an average of 12.4 months. The reverse arthroplasty group demonstrated better results than the hemiarthroplasty group in terms of abduction (average, 91° compared with 60°) and the Constant score (average, 53 compared with 39). However, the hemiarthroplasty group demonstrated better results than the reverse shoulder arthroplasty group in terms of active external rotation (average, 13.5° compared with 9°) and active internal rotation (average, 54.6° compared with 31°).
Lenarz et al.91 reported the results of reverse shoulder arthroplasty in a retrospective study of thirty patients with displaced three and four-part proximal humeral fractures. The average age was seventy-seven years, and the minimum duration of follow-up was one year. The mean postoperative ASES score was 78. The average active forward elevation was 139° (range, 90° to 180°), and the average active external rotation was 27° (range, 0° to 45°). Complications were noted in three patients (10%) and included one case of Grade-1 scapular notching. Similarly, in the study by Garrigues et al.92, reverse shoulder arthroplasty demonstrated better results than hemiarthroplasty in terms of forward flexion, the ASES score, the University of Pennsylvania shoulder score93, and the Single Assessment Numeric Evaluation score94.
The treatment of proximal humeral fractures remains challenging for the surgeon. Advances in implants and techniques have improved the outcomes of operative treatment, but controversy remains with regard to the best treatment choice for these heterogeneous injuries. Surgeons have many choices available to them when fixing proximal humeral fractures. Surgeons with training in percutaneous pinning, fracture plating, or the various arthroplasty options have the flexibility necessary to face these challenges.
We are not aware of any Level-I randomized controlled trials or high-quality prospective studies that have evaluated any of the treatment options for proximal humeral fractures. The majority of the data presented in the current review were culled from prospective comparative studies, case-control studies, and retrospective comparative studies. These data represent fair evidence for or against recommending an intervention. Furthermore, all of the methods and techniques described in the present report have been presented in the peer-reviewed literature.
Source of Funding: No external funds were used for this study.
Investigation performed at the Leni and Peter W. May Department of Orthopaedic Surgery, Mount Sinai School of Medicine, New York, NY
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 © 2014 by The Journal of Bone and Joint Surgery, Incorporated