Salvage of Failed Treatment of Hip Fractures George J. Haidukewych, MD, and Daniel J. Berry, MD Abstract With contemporary techniques of open reduction and internal fixation, most femoral neck and intertr ochanteric hip fractures heal uneventfully. Neverthe- less, because the number of hip frac- tures is large and continually increas- ing, a small percentage of patients experience nonunion or early fixation failure. 1,2 Unfavorable fracture pat- terns, poor implant placement, and poor bone quality all increase the like- lihood of failure of fracture fixation. 1,3,4 Effective salvage is important because patients typically are severely disabled. The main management options are re- vision internal fixation (with or with- out bone grafting) and prosthetic re- placement. The choice of salvage method depends on whether the frac- ture occurred at the femoral neck or at the intertrochanteric level. Treatment is then individualized, according to physiologic age, activity level, remain- ing bone quality, viability of the fem- oral head, and status of the hip joint articular surface. Preoperative Evaluation Although most nonunions with failed fixation devices and persistent frac- ture instability are easy to diagnose, occasionally nonunion can be subtle and difficult to recognize. Several months after internal fixation, patients may present with persistent pain and difficulty with ambulation. Radio- graphs may demonstrate settling of the fracture or backing out of hard- ware (Fig. 1, A). Alho et al 5 reviewed the radiographic signs that predict failure in patients with internally fixed femoral neck fractures; they consid- ered 3 months to be the critical time for prognosis. Change in fracture po- sition by 10 mm, change in screw po- sition by 5%, backing out of the screws by 20 mm, and perforation of the femoral head each correlated with a high rate of revision. When plain ra- diography is equivocal, computed to- mography (CT) can help determine whether bony union has occurred (Fig. 1, B). Usually, revision is con- sidered for acute failure of fracture fixation, unacceptable fracture align- ment, or established fracture non- union. Although 3 months is a rea- sonable time to expect union in most patients, fixation failure may be ev- ident much earlier; in some patients, however, especially those with radio- graphic evidence of progressive but incomplete healing, a longer period of observation may be necessary. In evaluating any patient with failed internal fixation of a hip frac- ture, occult infection should be con- sidered as a potential cause of the fail- ure. Prudent preoperative evaluation includes complete blood count with manual dif fer ential count, erythrocyte sedimentation rate, and C-reactive Dr. Haidukewych is Orthopaedic Traumatologist and Adult Reconstructive Surgeon, Florida Or- thopedic Institute, Tampa, FL. Dr. Berry is Pro- fessor of Orthopaedics, Mayo Clinic College of Medicine, and Consultant, Orthopaedic Surgery, Mayo Clinic, Rochester, MN. Neither Dr.Haidukewych nor the department with which he is affiliated has received anything of val- ue from or owns stock in a commercial company or institution related directly or indirectly to the subject of this article. Dr. Berry or the department with which he is affiliated has received research or institutional support from DePuy , Zimmer, and Stryker. Dr. Berry or the department with which he is affiliated has received royalties from DePuy. Reprint requests: Dr. Haidukewych, Florida Or- thopedic Institute, 13020 Telecom Parkway, Tem- ple Terrace, FL 33637. Copyright 2005 by the American Academy of Orthopaedic Surgeons. Typically, patients with failed internal fixation of a hip fracture have marked pain and disability. These patients may present treatment challenges. Salvage is tailored to the anatomic site of the nonunion, the quality of the remaining bone and articular surface, and patient factors such as age and activity level. In younger patients with either a femoral neck or intertrochanteric fracture nonunion with a satisfactory hip joint, treatment typically involves revision internal fixation with or without osteotomy or bone grafting. In older patients with poor remaining proximal bone stock or a badly damaged hip joint, conversion to hip arthroplasty can restore function effec- tively and reduce pain. For femoral head salvage procedures, choosing a fixation de- vice and accurate preoperative planning are the major challenges in decision mak- ing. For conversion to arthroplasty, the major challenges are assessing the need for acetabular resurfacing, selecting the femoral implant, and managing the greater tro- chanter. Technical challenges include broken hardware, deformity, and femoral bone defects. Attention to technical details can minimize potential complications. J Am Acad Orthop Surg 2005;13:101-109 Vol 13, No 2, March/April 2005 101 protein level. Aspiration of the non- union site does not need to be per- formed routinely because it is tech- nically difficult to obtain an adequate specimen, and the reliability of the re- sults of such aspirates has not been well documented. Intraoperative tis- sue from the nonunion site is obtained for frozen-section histology. When there is evidence of infection, all hardware should be removed, deep cultures obtained, and necrotic tis- sue débrided; antibiotic-impr egnated polymethylmethacrylate beads or spa- cers may be placed. If arthroplasty is contemplated as the final method of reconstruction, then a Girdlestone resection with placement of an antibiotic-impregnated spacer may be considered when the femoral head is thought to be infected. The definitive reconstruction is then performed af- ter a period of organism-specific in- travenous antibiotic administration. A staged approach is usually prefer- able when infection is present, wheth- er arthroplasty or an attempt to sal- vage the femoral head is planned. Symptomatic malunion is uncom- mon following hip fracture. Howev- er, shortening of the femoral neck, shortening through the intertrochan- teric area, and malunion of the great- er trochanter all can occur after hip fracture. Any of these can lead to limb-length discrepancy or adverse hip biomechanics, resulting in limp or pain. In most cases, moderately suboptimal hip biomechanics are ac- cepted as the trade-off to gain good bone apposition in a stable position and fractur e union. Little information is available about the options for sal- vage of a severe malunion; most data have been gathered from the treat- ment of neglected intertrochanteric hip fractures. In one series of 48 treat- ed hips, 6 corrective osteotomy was recommended for symptomatic inter - trochanteric malunions in younger patients, wher eas older patients were treated with hip arthroplasty. More studies are needed to determine the ideal methods to prevent and salvage malunions after hip fracture. Generally, the viability of the fem- oral head can be assessed with plain radiographs, using the radiographic criteria described for osteonecrosis. 7 If necessary, bone scintigraphy or magnetic resonance imaging (when titanium implants are present) can be useful. 7 However, such additional im- aging modalities are rarely required because in the younger patient with- out collapse of the femoral head, ev- ery attempt is made to preserve the femoral head, even if small areas of avascular bone are present. When evaluating the patient with a failed hip fracture, certain patient- specific issues also should be ad- dressed. When osteosynthesis is at- tempted, tobacco use in any form should be discontinued. Achieving optimal medical and nutritional sta- tus, especially in elderly, debilitated patients, also is critical. Salvage of Failed Femoral Neck Fractur es Young Patients Usually, femoral neck fractur e non- unions in physiologically young pa- tients are treated with methods de- signed to salvage the femoral head and preserve the hip joint. Preserving the femoral head is preferable to prosthet- ic replacement. The most common techniques used for femoral neck non- unions in young patients fall into two categories: those designed to improve the mechanical environment at the fracture site (ie, valgus-producing os- teotomies) and those designed to im- prove the biologic environment of the nonunion site by bone grafting (non- vascularized, free vascularized, or muscle pedicle–type grafts). 7 The Mey- ers quadratus femoris pedicle graft, the most widely studied graft, pro- vides a vascularized local bone graft to improve the biology at the nonunion site. 8-10 Its use may be indicated when there is loss of bone stock posterior- ly or when patients have well-aligned fractures with low shear angles. Sev- eral series have evaluated individu- al methods of bone grafting for fem- oral neck nonunions 8,9,11-17 (Table 1). The indications for these techniques have yet to be fully elucidated; how- ever, they may be useful for neglect- ed fractures, failed fixation attempts, or well-aligned nonunions with os- teonecrosis. The clear superiority of any of the bone grafting choices is un- substantiated by the curr ent literature. Figure 1 A, Anter oposterior radiograph demonstrating femoral neck nonunion in a 35-year- old woman. She continued to have groin pain with ambulation for more than 1 year post- operatively. Note the varus alignment and the backing out of the cannulated screws. B, Coro- nal CT scan demonstrating nonunion. Salvage of Failed Treatment of Hip Fractures 102 Journal of the American Academy of Orthopaedic Surgeons Valgus intertrochanteric osteoto- mies can convert shear forces at the nonunion site to compressive forces, which then promote fracture healing (Fig. 2). Marti et al 18 reported on a se- ries of 50 patients (mean age, 53 years) who were treated with valgus inter- trochanteric osteotomy for femoral neck nonunion. Eighty-six percent of nonunions united in a mean of 4 months. Of the 22 patients who had radiographic evidence of osteonecro- sis (without collapse) at the time of osteotomy, only 3 (14%) showed pro- gressive collapse of the femoral head requiring hip replacement. Anglen 19 reported on a series of 13 patients fol- lowed up for a mean of 25 months af- ter valgus osteotomy for failed inter- nal fixation of a femoral neck fracture. All fractures healed, and 11 of the 13 patients (85%) had good to excellent results. Later, two patients (15%) un- derwent arthroplasty because of os- teonecrosis. Ballmer et al 20 reported on a series of 17 patients with nonunions of the femoral neck treated with valgus- producing osteotomies. Twelve of 17 (71%) healed with one procedure. Three patients required revision fix- ation but eventually healed, increas- ing the overall union rate to 88%. Three patients (18%) had progressive os- teonecrosis and required hip arthro- plasty. Thus, even with areas of os- teonecrosis, the results of salvage of the femoral head can be good. When segmental collapse of the femoral head is present, valgus osteotomy is rare- ly a satisfactory alternative because the results are then less predictable. Additionally, the osteotomy deforms the proximal femur, which may make later revision to total hip arthr oplasty, if needed, more difficult. Wu et al 21 compared the use of a sliding compression screw with and without subtrochanteric valgus os- teotomy for femoral neck nonunions in 32 patients (mean age, 38 years). All of the nonunions healed at a mean of 4.6 months. Even though there were fewer complications in the nonosteotomy group, the authors rec- ommended valgus osteotomy for pa- tients with shortening of more than Table 1 Bone Grafting Techniques for Nonunion of the Femoral Neck: Summary of Results Study No. of Patients Mean Follow-Up (mo) Mean Age (yrs) No. (%) Preoperative Osteonecrosis Type of Graft No. (%) Fracture Union No. (%) Progression of Osteo- necrosis No. (%) Converted to Total Hip Ar- throplasty LeCroy et al 11 22 85 29 16 (73) stage I and II, 6 (27) stage III and higher Free vascularized fibula 20 (91) 13 (59) 2 (9) Nagi et al 12 40 59 35 8 (20) Free vascularized fibula 38 (93) 7 (18) 3 (8) Hou et al 13 5 24 24 0 Iliac crest pedicle (deep circumflex iliac artery) 5 (100) 0 0 Leung and Shen 14 15 60-84* 38 0 Iliac crest pedicle (deep circumflex iliac artery) 15 (100) 1 (7) 1 (7) Nagi et al 15 26 29 39 4 (15) Autograft fibula nonvascularized 25 (96) 0 0 Baksi 8 56 35 42 34 (61) stage I and II Quadratus femoris muscle-pedicle 42 (75) 2 (4) Not stated Meyers et al 9 32 14 followed >1 yr 16-79* Not stated Quadratus femoris muscle pedicle 23 (72) Not stated Not stated Bonfiglio and Voke 16 77 60 31-79* 77 (100) Autogenous tibial strut, nonvascu- larized 72 (94) Not stated Not stated Henderson 17 77 69 followed to union 46 Not stated Autograft fibula or tibia, nonvascu- larized 46 (67) Not stated Not stated * Mean not stated; therefore, the range is given. George J. Haidukewych, MD, and Daniel J. Berry, MD Vol 13, No 2, March/April 2005 103 1.5 cm because the valgus osteotomy helps gain leg length. Although studies of valgus os- teotomy have focused on union rates and progression of osteonecrosis, lit- tle has been written about clinical func- tion after such salvage procedur es. Re- cently, Mathews et al 22 evaluated the functional outcome in 15 patients with valgus-producing osteotomies for fem- oral neck nonunions at a mean of 4 years after sur gery. Although fracture union without progression of osteone- crosis was achieved in most patients, a persistent limp was common, prob- ably caused by loss of femoral offset and abductor moment arm (Fig. 2). Most femoral neck nonunions in younger patients result primarily from mechanical, not biologic, factors. The original fractures and subsequent non- unions typically have high shear an- gles (Pauwels type III 18 ), have become shortened, and are aligned in varus. The preferred salvage operation there- fore should be the valgus-producing intertrochanteric osteotomy. The technique of valgus-pr oducing intertrochanteric osteotomy has been well described. 23 It involves convert- ing a vertically oriented fracture to a more horizontal orientation, thus min- imizing the shear forces at the frac- ture site and promoting union. The recommended horizontality of the nonunion after osteotomy should be approximately 20° to 30°. 18 Thus, the size of the intertrochanteric wedge re- moved would be calculated as the dif- ference between the curr ent nonunion verticality and the desired horizon- tality. For example, a patient with a 70° nonunion verticality would have a 40° to 50° wedge resected from the intertrochanteric region to properly reposition the proximal fragment. Fracture shear angles may be quite dif- ficult to measure accurately because of leg rotation and should be measured from a line perpendicular to the fem- oral shaft. 19 These osteotomies should be per- formed on a fracture table that allows excellent fluor oscopic visualization of the pr oximal femur. Car eful preoper- ative templating is performed to de- termine the appropriate blade plate angle. Blade plates with multiple an- gles are available, and the selected an- gle of the plate should allow excel- lent fixation of the proximal fragment and the appropriate neck shaft angle after correction. After the original hardware is removed, the proximal femur is prepared with the seating chisel to accept the blade plate before the osteotomy is performed (Fig. 2, C). It is important to mark the correct leg rotation, usually with Kirschner wires in the proximal and distal frag- ments or before making the osteoto- my. The chisel that creates a path for the blade is seated to the appropri- ate depth and is then removed. The osteotomy is then performed parallel to the chisel tract, taking care to leave at least 2 c m of bone between the inferior aspect of the blade tract and the superior aspect of the os- teotomy. This minimizes the chance of fracture o f this inferior bony bridge Figure 2 A, Early postoperative anteroposterior radiograph following valgus-producing intertrochanteric osteotomy. Note the medializa- tion of the femoral shaft, which should be minimized if possible. B, Femoral neck nonunion with the typical foreshortening and verticality of the nonunion site. C, Appropriate seating of the chisel in the proximal fragment based on templating. To avoid fixation failure, it is im- portant to leave sufficient bone between the planned blade plate and the osteotomy. In this situation, the intertrochanteric wedge size re- moved is planned to allow horizontal orientation of the nonunion site. D, Nonunion verticality has been decreased from 70° (panel B) to approximately 30°. Note the lateralization of the femoral shaft and fixation with the angled blade plate. Salvage of Failed Treatment of Hip Fractures 104 Journal of the American Academy of Orthopaedic Surgeons (Fig. 2, C). Commercially available pro- tractors are available for exact calcu- lation of the intertrochanteric wedge. These are typically placed along the anterior femur, and a fluoroscopic im- age is taken (Fig. 3). Kirschner wires are then used to mark the appropri- ate wedge trajectory, and the os- teotomy is performed with a saw. It is important to cool the saw with pe- riodic irrigation because the bone in this anatomic region can be dense, and thermal necrosis could occur. After the appropriate wedge has been removed, a blade plate of appro- priate length and angle is impacted into the femoral head. A secondary proximal screw is placed below the blade; then, distal screws are placed in the usual fashion (Fig. 2, D). Good compression across the osteotomy site usually results as the distal screws are placed because of the osteotomy obliq- uity. Care should be taken to keep the bone both proximal and distal to the osteotomy well aligned on the later- al view to avoid creating a deformi- ty that would be difficult later to con- vert to a hip arthroplasty. It is wise to place bone graft at the osteotomy site by morcellizing the cancellous bone from the resected wedge and placing this along the osteotomy line. The wound is closed in the usual layered fashion. Patients should be cautioned that, although union rates are high, a persistent limp is common. The amount of femoral shaft medi- alization should be minimized when performing such osteotomies. This can be accomplished by choosing a slightly longer blade. When seated to the appropriate depth, the plate re- mains lateral, which helps keep the shaft lateral. Shaft medialization de- creases offset, thereby decreasing ab- ductor muscle efficiency and increas- ing the joint r eactive force. In addition, excessive shaft medialization may cause valgus alignment at the knee. Occasionally, despite all efforts to preserve the femoral head in the young patient, there may be no rea- sonable alternative to hip arthroplas- ty or hip arthrodesis. For example, a patient with total collapse of the femoral head and a nonunion would not be a good candidate for a joint- preserving procedure. Hip arthro- plasty in young patients should be re- served for those in whom several well-done attempts to preserve the joint have failed and for those with collapse of the femoral head. Older Patients Typically, in physiologically older patients, femoral neck fracture non- unions are salvaged with hip arthro- plasty, either hemiarthroplasty or to- tal hip arthr oplasty. Hemiarthroplasty has the advantage of being a less ex- tensive surgery and likely has a lower risk of instability. In cases of badly damaged articular cartilage of the hip (from degenerative arthritis or er osion because of hardware penetration), to- tal hip arthroplasty is usually pre- ferred. When the articular cartilage of the acetabulum is well preserved, the decision between hemiarthroplasty and total hip arthroplasty is at the sur- geon’s discretion. Scrutiny of preop- erative radiographs and intraopera- tive inspection of the acetabular cartilage may guide decision making. Either bipolar or unipolar compo- nents may be used, based on surgeon preference. A bipolar implant is more commonly used if total hip arthro- plasty is not performed because of the excellent hip stability and low rates of acetabular er osion it offers. If hemi- arthroplasty is planned, it is wise to have total hip arthroplasty compo- nents available as well because pre- operative radiographs may underes- timate the amount of articular surface damage. Several important technical issues must be considered when a total hip arthroplasty is done for failed femo- ral neck fracture. The original hard- ware usually needs to be removed, thereby leaving a defect in the shaft of the femur. Also, acetabular bone quality in patients with femoral neck nonunion often is very poor because of disuse osteopenia. Most of these pa- tients do not have degenerative hip Figure 3 Anteroposterior fluoroscopic image demonstrating calculation of intertrochan- teric wedge and placement of Kirschner wires. George J. Haidukewych, MD, and Daniel J. Berry, MD Vol 13, No 2, March/April 2005 105 arthritis and so do not have the scle- rotic subchondral bone typically present in patients undergoing elec- tive hip r eplacement for degenerative arthritis. Therefore, when a cement- less cup is used, poor press-fit fix- ation or even acetabular fracture during implant insertion can occur. Judicious acetabular r eaming, with an effort made to preserve the subchon- dral bone, is recommended. Care should be taken to avoid forceful ac- etabular component impaction, and augmentation of fixation with screws should be considered. Standard fem- oral components typically can be used; however, proximal defects from pri- or hardware can pose intraoperative fracture risk during canal pr eparation. Little has been written about the results and complications of hip ar- throplasty for failed tr eatment of fem- oral neck fractures. 24-28 McKinley and Robinson 29 reported a matched-pair series of 214 patients: 107 patients with failed open reduction and inter- nal fixation of a femoral neck fracture were treated with early-salvage ce- mented total hip arthr oplasty; another group of 107 patients with fracture were treated with arthroplasty. The salvage arthroplasty group had sig- nificantly higher dislocation rates (21% versus 8%) and more superficial infections (P < 0.05) than did the pri- mary arthroplasty group. Functional scores and implant survivorship were inferior for the salvage group, as well. Mabry et al 30 reported on the long- term follow-up of 99 patients with femoral neck nonunions treated with Charnley hip arthroplasties between 1970 and 1977. The mean age at time of arthroplasty was 68 years (range, 36 to 92 years). At a mean 12-year follow-up of 84 patients, 12 had un- dergone revision arthroplasty. Im- plant survivorship free of revision for any reason was 93% at 10 years and 76% at 20 years. Implant survivorship was better for older patients (age >65 years). Instability occurred in 9% of patients, half of whom had recurrent dislocation. Thus, reported results clearly document the value of total hip arthroplasty for salvage of fem- oral neck nonunion in older patients. The use of larger-diameter femoral heads and surgical approaches that reduce dislocation risk may be use- ful to reduce the risk of dislocation in this patient population, although no published data substantiate this speculation. Salvage of Failed Intertrochanteric Hip Fractures Young Patients Nonunion of the intertrochanteric hip fracture in young patients is un- common. For those with proximal bone quality adequate for internal fix- ation, the most common treatment is revision internal fixation with select- ed bone grafting. 31 A fixed-angle de- vice, such as the angled blade plate or dynamic condylar screw, is pre- ferred, usually accompanied by au- togenous bone grafting. These devic- es can target the bone in the inferior region of the femoral head, which usually has not been violated by pri- or implants (Fig. 4). Few studies of intertrochanteric nonunions have been published. 32,33 Mariani and Rand 34 reported on 1 1 pa- tients (mean age, 53 years) whose in- tertrochanteric nonunions were treat- ed with repeat open reduction and internal fixation. Nine of 11 (82%) achieved union at a mean of 6 months. A variety of implants was used suc- cessfully, based on the location of re- maining bone stock in the femoral head. Wu et al 35 reported on 14 inter- trochanteric fractures with cutout of a lag screw of a dynamic hip screw fixation. All were treated with rein- sertion of a lag screw inferiorly in the femoral head, cement augmentation, and valgus-producing subtrochanteric osteotomy. All nonunions healed at a mean of 5 months. Sarathy et al 36 reported on seven patients with in- tertrochanteric nonunions treated with valgus osteotomy, medial displace- ment, and 130° blade plate fixation. Figure 4 A, Anteroposterior radiograph demonstrating failure of internal fixation of an in- tertrochanteric fracture 3 weeks postoperatively in a 52-year-old woman. Note the excellent remaining proximal bone stock. B, Anteroposterior radiograph in another patient demon- strating salvage with a 95° angled blade plate. Note the fixation targeting the inferior fem- oral head bone. (Reproduced with permission from Haidukewych GJ, Berry DJ: Salvage of failed internal fixation of intertrochanteric hip fractures. Clin Orthop 2003;412:184-188.) Salvage of Failed Treatment of Hip Fractures 106 Journal of the American Academy of Orthopaedic Surgeons Six of seven healed. Haidukewych and Berry 31 reported on a series of 20 in- tertrochanteric nonunions revised with open reduction and internal fixation and selected bone grafting. Fixed-angle devices were used in 75% of cases. Nineteen of 20 nonunions healed. The available literature therefore suggests that a variety of differ ent implants may be used successfully to salvage the in- tertrochanteric nonunion as long as stable fixation of the proximal frag- ment is obtained. Older Patients Most intertr ochanteric hip fracture nonunions occur in older patients with poor proximal bone quality and fail by implant cutout from the fem- oral head. 1 The decision to perform revision internal fixation versus pr os- thetic r eplacement is based on patient characteristics, fracture pattern, re- maining bone quality, and status of the hip joint. In older patients, arthro- plasty has some advantages because it allows earlier patient mobilization. When hip arthroplasty is per- formed for salvage of failed intertro- chanteric fractures, specific technical considerations must be addressed. The initial decision is whether to per- form a total hip arthroplasty or a hemiarthroplasty. It is not uncommon to have had the cutout of the previ- ous internal fixation cause secondary damage to the hip joint. Usually, in this circumstance or in patients with markedly sever e preexisting arthritis, a total hip arthroplasty is performed. With well-preserved articular carti- lage, hemiarthroplasty may be con- sidered. The same advantages and disadvantages of hemiarthroplasty versus total hip arthroplasty dis- cussed for salvage of femoral neck nonunion also pertain to intertro- chanteric nonunion. Defects fr om previous internal fix- ation devices on the lateral femoral shaft create stress risers that can lead to intraoperative fracture of the fe- mur, particularly with torsion. Pre- liminary dislocation of the hip before hardware is removed may reduce fe- mur fractur e risk in these hips, which often are quite stiff and can require much force to dislocate. Frequently, broken screws are present. It is help- ful to keep instruments, including tre- phines and grasping tools, available to remove broken screws. Most patients with failed intertro- chanteric fracture fixation have bone loss below the standard resection level for a routine, primary total hip arthroplasty. Therefore, many need a calcar-replacing implant to restore leg length and hip stability. To pre- vent the chance of subsequent frac- ture when using longer stems, it is wise to bypass screw holes in the fe- mur by two cortical diameters 37 (Fig. 5). Successful femoral component fixation can be obtained with either cemented or cementless implants. For many older patients, cement- ed fixation is advantageous, particu- larly when bone quality is poor and the canal diameter is large. Cement- ed fixation also allows rapid mobili- zation in this patient population. If a cemented stem is chosen, the surgeon needs to be aware that cement can ex- trude from the empty screw holes. 38 Bone graft from the resected femoral head can be used to graft large lat- eral defects, such as those created by the barrel of a sliding hip screw. If a cementless implant is used, ex- tensively porous-coated stems have the advantage of providing fixation in the diaphysis of the femur, bypass- ing the damaged, deformed, or defi- cient proximal bone. Intraoperative fracture is possible with insertion of large cementless implants, especial- ly in patients with poor bone with multiple previous bicortical screw holes. Intraoperative radiographs af- ter implant placement are recom- mended, regardless of the type of femoral fixation chosen. Management of the greater tro- chanter has been problematic and warrants special discussion. The greater trochanter may be a separate, ununited piece of bone, or it may be Figure 5 A , Anter oposterior radiograph demonstrating intertrochanteric nonunion with cut- out and poor proximal bone stock in a 78-year-old woman. B, Anteroposterior radiograph in another patient showing salvage with a long-stem, calcar-replacing bipolar hemiarthro- plasty and fixation of the greater trochanter. George J. Haidukewych, MD, and Daniel J. Berry, MD Vol 13, No 2, March/April 2005 107 malunited, preventing entrance into the femoral canal for femoral prep- aration. In these circumstances, the trochanteric slide technique is pre- ferred because it r etains the vastus lat- eralis muscle, greater trochanter, and abductor muscles as a single sleeve of tissue. Patients should be coun- seled in advance that trochanteric problems relating to either persistent nonunion or painful trochanteric fix- ation devices are not infrequent after such reconstructions. 39 Finally, bone deformity of the prox- imal femur related to fracture callus, fracture translation, or malunion of- ten is pr esent, which increases the risk of femoral fractur e during canal pr ep- aration. Shaping of the proximal bone with a high-speed burr is safer than performing the same procedure with a rasp. The tracts of previously placed fixation devices often are sclerotic and can deflect reamers or broaches, lead- ing to proximal fracture or femoral perforation. There are few published series on the results of hip arthroplasty for re- vision after intertrochanteric non- unions. Mariani and Rand 34 reported on nine patients with intertrochan- teric nonunions treated with hip ar- throplasty. At an average follow-up of 6.6 years, all patients had functional improvement. Stoffelen et al 40 re- ported on seven hip arthroplasties for intertrochanteric nonunion. Seventy- two percent (5 patients) had good to excellent results. Mehlhoff et al 41 re- ported on 13 patients followed for a mean of 34 months; only 5 had good to excellent r esults. Three patients had dislocations and two of them required revision for instability. More recently, Haidukewych and Berry 39 reported on 60 patients (mean age, 78 years) treated between 1985 and 1997 with hip arthroplasty for failed treatment of intertrochanteric hip fractures. Thirty-two total hip arthroplasties and 27 bipolar hemiar- throplasties were performed. Forty- four patients were followed for a mean of 5 years. Two hip arthroplas- ties were revised for aseptic loosen- ing at 8 and 10 years. There was one dislocation. The 7-year survivorship of the arthroplasties free of revision for any reason was 100%; 10-year sur- vivorship was 88%. Importantly, a calcar-replacing stem or extra long neck-length stem was needed in 65% of cases, and long-stemmed implants were used in a high percentage of pa- tients, as well. A standard prosthesis was suitable only in 15% of cases. Se- rious complications were uncommon, and most patients’ ambulatory status and pain were markedly improved. The most common persistent com- plaint was discomfort over the great- er trochanter, which was present in 11% of hips. Summary In younger patients, salvage of the failed hip fracture typically involves efforts to preserve the hip joint with internal fixation, whereas in most old- er patients, prosthetic replacement is a r eliable salvage option. The location of the nonunion, physiologic age of the patient, quality of the remaining proximal bone, presence of deformi- ty, status of the hip joint, and viabil- ity of the femoral head all influence decision making. Regardless of the salvage method chosen, attention to specific technical details can improve the success rate and reduce the com- plications of treating these challeng- ing problems. The OKO video ″Approaches to Total Hip Arthr oplasty,″ by Bassam A. Masri, MD, Philip Mitchell, MD, and Clive Duncan, MD, is available at http://www5.aaos.org/oko/jaaos/ main.cfm. References 1. Kyle RF, Cabanela ME, Russell TA, et al: Fractures of the proximal part of the femur. Instr Course Lect 1995;44: 227-253. 2. Kyle RF, Gustilo RB, Premer RF: Anal- ysis of 622 intertrochanteric hip fractures. J Bone Joint Surg Am 1979;61:216-221. 3. Baumgaertner MR, Solberg BD: Aware- ness of tip-apex distance reduces fail- ure of fixation of trochanteric fractures of the hip. J Bone Joint Surg Br 1997;79: 969-971. 4. Haidukewych GJ, Israel TA, Berry DJ: Reverse obliquity of fractures of the in- tertrochanteric region of the femur. J Bone Joint Surg Am 2001;83:643-650. 5. AlhoA, Benterud JG, Solovieva S: Inter- nally fixed femoral neck fractures: Ear- ly prediction of failure in 203 elderly patients with displaced fractures. Acta Orthop Scand 1999;70:141-144. 6. Lifeso R, Younge D: The neglected hip fracture. J Orthop Trauma 1990;4:287-292. 7. Jackson M, Learmonth ID: The treat- ment of nonunion after intracapsular fractures of the proximal femur. Clin Orthop 2002;399:119-128. 8. Baksi DP: Internal fixation of ununited femoral neck fractures combined with muscle-pedicle bone grafting. J Bone Joint Surg Br 1986;68:239-245. 9. Meyers MH, Harvey JP Jr, Moore TM: The muscle pedicle bone graft in the treatment of displaced fractures of the femoral neck: Indications, operative technique and results. Orthop Clin North Am 1974;5:779-792. 10. Meyers MH, Harvey P Jr, Moore TM: Treatment of displaced subcapital and transcervical fractures of the femoral neck by muscle-pedicle-bone graft and internal fixation. J Bone Joint Surg Am 1973;55:257-274. 11. LeCroy CM, Rizzo M, Gunneson EE, Urbaniak JR: Free vascularized fibular bone grafting in the management of femoral neck nonunion in patients younger than fifty years. J Orthop Trau- ma 2002;16:464-472. 12. Nagi ON, Dhillon MS, Goni VG: Open reduction, internal fixation and fibular Salvage of Failed Treatment of Hip Fractures 108 Journal of the American Academy of Orthopaedic Surgeons autografting for neglected fracture of the femoral neck. J Bone Joint Surg Br 1998;80:798-804. 13. Hou SM, Hang YS, Liu TK: Ununited femoral neck fractures by open reduc- tion and vascularized iliac bone graft. Clin Orthop 1993;294:176-180. 14. Leung PC, Shen WY: Fracture of the fem- oral neck in younger adults: A new meth- od of treatment for delayed and non- unions. Clin Orthop 1993;295:156-160. 15. Nagi ON, Gautam VK, Marya SK: Treat- ment of femoral neck fractur es witha can- cellous scr ew and fibular graft. J Bone Joint Surg Br 1986;68:387-391. 16. Bonfiglio M, Voke EM: Aseptic necrosis of the femoral head and nonunion of the femoral neck. J Bone Joint Surg Am 1968;50:48-66. 17. Henderson MS: Ununited fracture of the neck of the femur treated by the aid of the bone graft. J Bone Joint Surg 1940; 22:97-106. 18. Marti RK, Schuller HM, Raaymakers E: Intertrochanteric osteotomy for non- union of the femoral neck. J Bone Joint Surg Br 1989;71:782-787. 19. Anglen JO: Intertrochanteric osteotomy for failed internal fixation of femoral neck fracture. Clin Orthop 1997;341: 175-182. 20. Ballmer FT, Ballmer PM, Baumgaertel F, Ganz R, Mast JW: Pauwels osteoto- my for nonunions of the femoral neck. Orthop Clin North Am 1990;21:759-767. 21. Wu CC, Shih CH, Chen WJ, Tai CL: Treatment of cutout of a lag screw of a dynamic hip screw in an intertrochan- teric fracture. Arch Orthop Trauma Surg 1998;117:193-196. 22. Mathews V, Berry DJ, Trousdale RT, Cabanela ME: Poster: Clinical and func- tional results of valgus intertrochanter- ic osteotomy for femoral neck fracture nonunion. 70th Annual Meeting Proceed- ings. Rosemont, IL: American Academy of Orthopaedic Surgeons, 2003, p 380. 23. Müller ME: Intertrochanteric osteoto- my: Indication, preoperative planning, technique, in Schatzker J (ed): The Inter- trochanteric Osteotomy. New York, NY: Springer-Verlag, 1984, pp 25-66. 24. Franzen H, Nilsson LT, Stromqvist B, Johnsson R, Herrlin K: Secondary total hip replacement after fractures of the femoral neck. J Bone Joint Surg Br 1990; 72:784-787. 25. Hagglund G, Nordstrom B, Lidgren L: Total hip replacement after nailing fail- ure in femoral neck fractures. Arch Or- thop Trauma Surg 1984;103:125-127. 26. Nilsson LT, Jalovaara P, Franzen H, Niinimaki T, Stromqvist B: Function af- ter primary hemiarthroplasty and sec- ondary total hip arthroplasty in femo- ral neck fracture. J Arthroplasty 1994;9: 369-373. 27. Tabsh I, Waddell JP, Morton J: Total hip arthroplasty for complications of prox- imal femoral fractures. J Orthop Trauma 1997;11:166-169. 28. Turner A, Wroblewski BM: Charnley low-friction arthroplasty for the treat- ment of hips with late complications of femoral neck fractures. Clin Orthop 1984;185:126-130. 29. McKinley JC, Robinson CM: Treatment of displaced intracapsular hip fractures with total hip arthroplasty: Compari- son of primary arthroplasty with early salvage arthroplasty after failed inter- nal fixation. J Bone Joint Surg Am 2002; 84:2010-2015. 30. Mabry TM, Prpa B, Haidukewych GJ, Harmsen WS, Berry DJ: Long-term re- sults of total hip arthroplasty for fem- oral neck fracture nonunion. J Bone Joint Surg Am 2004;86:2263-2267. 31. Haidukewych GJ, Berry DJ: Salvage of failed internal fixation of intertrochan- teric hip fractures. Clin Orthop 2003;412: 184-188. 32. Haentjens P, Casteleyn PP, Opdecan P: Hip arthroplasty for failed internal fix- ation of intertrochanteric and subtro- chanteric fractures in the elderly pa- tient. Arch Orthop Trauma Surg 1994;113: 222-227. 33. Kim Y-H, Oh J-H, Koh Y-G: Salvage of neglected unstable intertrochanteric fractures with cementless porous- coated hemiarthroplasty. Clin Orthop 1992;277:182-187. 34. Mariani EM, Rand JA: Nonunion of in- tertrochanteric fractures of the femur following open reduction and internal fixation: Results of second attempts to gain union. Clin Orthop 1987;218:81-89. 35. Wu CC, Shih CH, Chen WJ, Tai CL: Treatment of femoral neck nonunions with asliding compression screw: Com- parison with and without subtrochan- teric valgus osteotomy. J Trauma 1999; 46:312-317. 36. Sarathy MP, Madhavan P, Ravichan- dran KM: Nonunion of intertrochanter- ic fractures of the femur. J Bone Joint Surg Br 1995;77:90-92. 37. Patterson BM, Salvati EA, Huo MH: Total hip arthroplasty for complications of intertrochanteric fracture: A techni- cal note. J Bone Joint Surg Am 1990;72: 776-777. 38. Eschenroeder HC Jr, Krackow KA: Late onset femoral stress fracture asso- ciated with extruded cement following hip arthroplasty. Clin Orthop 1988;236: 210-213. 39. Haidukewych GJ, Berry DJ: Hip arthro- plasty for salvage of failed treatment of intertrochanteric hip fractures. J Bone Joint Surg Am 2003;85:899-905. 40. Stoffelen D, Haentjens P, Reynders P, et al: Hip arthroplasty for failed internal fixation of intertrochanteric and subtro- chanteric fractures in the elderly pa- tient. Acta Orthop Belg 1994;60:135-139. 41. Mehlhoff T, Landon GC, Tullos HS: To- tal hip arthroplasty following failed in- ternal fixation of hip fractures. Clin Or- thop 1991;269:32-37. George J. Haidukewych, MD, and Daniel J. Berry, MD Vol 13, No 2, March/April 2005 109 . with failed internal fixation of a hip frac- ture, occult infection should be con- sidered as a potential cause of the fail- ure. Prudent preoperative evaluation includes complete blood count. broken hardware, deformity, and femoral bone defects. Attention to technical details can minimize potential complications. J Am Acad Orthop Surg 2005;13:101-109 Vol 13, No 2, March/April 2005 101 protein