Femoral intertrochanteric fractures are one of the most common orthopedic procedures, and most intertrochanteric fractures can be successfully treated with intramedullary nailing or plating. However, a minority of patients have fractures that do not heal because of the initial type of fracture, comminuted fracture, poor fracture fixation, or poor bone quality. Failed hip fracture treatment causes dysfunction and pain.1 The two main treatment options for patients with failed treatment of intertrochanteric femoral fractures are revision with internal fixation and salvage treatment with hip arthroplasty. Thirty-two patients who underwent hip arthroplasty after failed internal fixation of femoral intertrochanteric fractures admitted to our hospital from July 2004 to June 2006 were followed up and their outcomes, surgical techniques and complications were retrospectively analyzed. Data and methods I. General data From July 2004 to June 2006, 32 patients with failed internal fixation treatment for intertrochanteric fracture of the femur underwent hip arthroplasty in our hospital. There were 24 male cases and 8 female cases; the age at the time of fracture ranged from 54 to 80 years, with an average of 68 years. The interval from fracture to arthroplasty averaged 40 months (5-70 months). Fifteen cases were treated with slide hip screws, 10 cases with intramedullary nailing, 5 cases with plate fixation, and 2 cases with multiple screw fixation. Failure mode: tension screws penetrated the femoral head in 8 cases, non-healing in 9 cases, ischemic necrosis of the femoral head in 7 cases, and traumatic arthritis in 8 cases. Total hip arthroplasty was performed in 28 cases (all were biologic acetabulum) and bipolar artificial femoral head replacement in 4 cases. Cemented femoral stems in 12 cases and non-cemented femoral stems in 20 cases. Standard femoral prosthesis was used in 25 cases and long-stem femoral prosthesis in 7 cases. II. Surgical methods All surgeries were performed by the same group of surgeons, and all patients used the posterior-lateral surgical approach to the hip joint. The plate screws were left in place prior to hip dislocation to reduce the risk of intraoperative fracture. After hip exposure, the lesser trochanter was identified. Based on preoperative plain radiographs, the femoral neck is osteotomized around the tension screw by measuring the femoral neck length directly above the lesser trochanter. Because there may be a medial bone defect in the proximal femur, sometimes the bone is osteotomized right at the lesser trochanter. The femoral head and tension screws are removed together in a retrograde fashion. The lateral femoral muscle is split longitudinally to expose and remove the plates and screws from the lateral femoral cortex. Bone remodeling around the nail path of the tension screw often results in intramedullary sclerosis and closure of the femoral medullary cavity. It is sometimes necessary to apply a power abrasive drill to remove the sclerotic bone and recanalize the femoral medullary cavity, followed by a minimal reaming drill or file to prepare the medullary cavity, a step that requires great care due to the increased risk of intraoperative femoral fracture during hip arthroplasty due to disuse osteoporosis, cortical bone defects and alterations, and deformities. An appropriately sized femoral prosthesis should be selected preoperatively. Sometimes to compensate for the proximal femoral bone defect and restore limb length, maintaining hip stability may require a femoral spur replacement prosthesis or an extended eccentric spur stem, requiring consideration of a long stem femoral prosthesis through the distal screw hole. Sometimes bone grafting of the resected femoral head is required. Fixation of the femoral prosthesis can be with a cemented or non-cemented prosthesis. The use of cemented and uncemented femoral prostheses depends on the quality of the femur and the geometry of the femoral medullary cavity. With cemented femoral prostheses, care must be taken to avoid extrusion of the cement from the screw holes during compression and implantation of the prosthesis. With non-cemented femoral prostheses, intraoperative fractures may occur when larger femoral prostheses are implanted, especially in patients with poor femoral quality and a large number of bilateral cortical screw holes. Prophylactic trapping of steel cables/wires distal to the screw holes will reduce the risk of fracture. Therefore, regardless of the choice of femoral fixation prosthesis, intraoperative radiographs are recommended after prosthesis placement. For acetabular prostheses, we usually use uncemented fixation and acetabular preparation is mostly done as a primary total hip arthroplasty. Two intraoperative issues to be aware of are disuse osteoporosis and acetabular defects due to screw penetration. Careful, noninvasive reaming with frequent observation of the acetabular bed is required to avoid medically induced penetration. Screw-penetrated bone defects are usually contained and are easily managed with femoral head autogenous bone grafts. When converting to total hip arthroplasty we prefer a larger femoral head (32 or 36 mm) so that the risk of postoperative dislocation is reduced due to the patient’s usually poor preoperative abductor tone and limb shortening. The decision to perform an artificial femoral head replacement or a total hip arthroplasty depends on the treating surgeon. Total hip arthroplasty was routinely performed if significant damage to the acetabular cartilage was found intraoperatively, and some patients with good acetabular articular cartilage also underwent total hip arthroplasty as a precautionary measure. All patients were treated with postoperative anticoagulation and received prophylactic antibiotics during the perioperative period. Postoperative plan For patients with no special considerations, gait exercises and walking were the same as for the initial total hip arthroplasty. For patients who underwent sliding osteotomy and fixation of the greater trochanter due to non-union of the greater trochanter, partial weight-bearing with an abduction brace for 6 weeks was recommended, followed by gradual increase in weight-bearing as tolerated. The hip joint function was scored according to the Harris scale, and the fixation of the prosthesis was evaluated by X-ray. The fixation of cemented femoral prosthesis was evaluated according to the Harris5 criteria, and the fixation of uncemented femoral prosthesis was evaluated according to the Engh6 criteria. Loosening of uncemented acetabular prosthesis was defined as prosthesis movement, complete translucent line at the prosthesis-bone interface or breakage of fixation screws. For patients with bipolar artificial femoral head replacements, acetabular cartilage wear was evaluated according to LaBelle7 criteria. The Brooker8 classification system was used to grade the heterotopic ossification. Results The mean operative time was 170 minutes (105-250 minutes), and the mean blood loss was estimated at 1100 mL (500-2800 mL). There were 4 complications and 2 intraoperative fractures of the greater trochanter. The fractures occurred during expansion and opening of the femoral medullary cavity, and the fractures were usually fixed with a cable clasp device. 1 femoral fracture occurred during preparation of the femoral medullary cavity and was treated with a looped wire. 1 postoperative dislocation occurred and was given a closed reduction. Twenty-eight patients (87.5%) were followed up for at least 2 years after hip arthroplasty, with a mean follow-up of 5 years (4-6 years). All patients had moderate or severe pain in the hip joint before hip replacement, with a mean preoperative Harris score of 37 (32-45) in 28 patients and a mean score of 88 (84-95) at final follow-up. 3 had mild pain and 3 had occasional minimal pain. 2 had mild limping during routine walking. Follow-up radiographs showed normal position of the joint prosthesis, mean abduction angle of the artificial acetabulum of 44o (42-48o), and no loosening of the acetabular prosthesis. There was no loosening of the cemented femoral prosthesis and no instability of the non-cemented femoral prosthesis. 9 cases of cemented femoral prosthesis with cement fixation were grade C, probably due to the presence of multiple cortical screw holes and difficulty in cement pressurization. In 3 hips, heterotopic ossification was found at the 6-month postoperative review, and Brooker’s grade II was 2 hips and grade III was 1 hip. Discussion In most patients with failed internal fixation for femoral intertrochanteric fractures, hip arthroplasty provides significant pain relief and functional improvement. Failure of internal fixation treatment for femoral intertrochanteric fractures presents a significant challenge to orthopaedic surgeons, who should consider that occult infection may be the cause of failure when faced with a femoral intertrochanteric fracture with failed internal fixation. The current clinical pathway at the author’s joint center includes preoperative routine blood tests, sedimentation and C-reactive protein measurements, and intraoperative frozen section analysis. If there is evidence of infection, all metal fixation material should be removed, irrigated and thoroughly debrided, and the arthroplasty staged after intravenous administration of specific antibiotics. Many idiosyncratic problems can occur during hip arthroplasty for failed internal fixation of femoral intertrochanteric fractures. This is because the internal fixation device must first be removed and the operation takes longer. In addition, the need to expose the internal fixation device through scar tissue dissection increases blood loss and there is the potential for misuse of peripheral neurovascular structures and muscles. The anatomy of the proximal femur is often deformed, especially if the intertrochanteric fracture is poorly repositioned or if the medial bone is crushed. Bone quality is usually poor due to the patient’s pre-existing osteoporosis, which is further exacerbated by disuse after failed internal fixation. Poor healing of the greater trochanter or re-crushing during hip replacement will affect the function of the abductor muscles, leading to an increased rate of dislocation and adversely affecting the walking function after hip replacement. Internal fixation devices for femoral intertrochanteric fractures may pose some problems for hip arthroplasty. Removal of the previous internal fixation device caused a femoral stem bone defect, which increased the stress leading to intraoperative femoral fracture. In addition, these hips are very stiff and may require significant force to dislocate the hip; dislocation before removal of the internal fixation device may reduce the risk of fracture. In order to deal with the possible presence of broken nails, it is often necessary to have specific tools such as ring drills, top drills and grasping tools to remove the broken nail. There are some technical considerations for hip replacement surgery for failed internal fixation of intertrochanteric fractures of the femur. Depending on the condition of the acetabular articular cartilage, the surgeon must decide whether to perform an artificial femoral head replacement or a total hip arthroplasty. Secondary hip injury caused by internal fixation penetrating the femoral head is not uncommon, and if the patient had previous comorbid osteoarthritis, a total hip replacement should be performed. If the acetabular cartilage is good, an artificial femoral head replacement may be considered. In the case of the acetabulum, the quality of the acetabulum is often poor due to disuse osteoporosis. In addition, since most patients do not have degenerative hip osteoarthritis, there is no sclerotic subchondral bone. Therefore, implantation of a non-cemented acetabular prosthesis may be difficult at the time of acetabular replacement due to poor compression fit. The acetabulum should be carefully filed to preserve as much subchondral bone as possible and to avoid violent impingement of the acetabular prosthesis, in addition to the need for reinforced fixation with screws. Non-union of the fracture may occur below the level of the standard osteotomy for initial total hip arthroplasty. Sometimes it may be necessary to replace the femoral prosthesis with a revision of the femoral spur used for the surgery to restore limb length and eccentric spacing and maintain hip stability. Theoretically, a long-stemmed femoral prosthesis exceeding the screw hole by at least twice the diameter of the femur (usually 6 cm) should be used to reduce the risk of intraoperative fracture . However, Zhang10 performed arthroplasty with a standard femoral prosthesis in 19 patients with failed internal fixation of femoral intertrochanteric fractures. 15 patients were followed up for a mean of 7.4 years, and the Harris score increased from a mean of 38.4 preoperatively to 79.8 at the final follow-up with no femoral fracture complications. Overall, cemented or non-cemented femoral prostheses were able to achieve successful femoral prosthetic fixation. Cemented prosthetic fixation may be chosen in older patients, especially if the femur is of poor quality and the medullary cavity is wide, and cemented fixation enables early mobility in these patients. However, if a cemented femoral prosthesis is chosen, the surgeon must be prepared to dispose of the cement extruded from the screw hole. A resected femoral head can be used to graft over a large number of lateral defects, such as those caused by sliding hip screws. If a biologic femoral prosthesis is used, an extensively porous coated stem provides stem fixation and is able to pass through any injury, deformity or defect in the proximal femur. However, intraoperative fractures may occur when inserting larger non-cemented prostheses, especially in patients with poor bone quality and a large number of bilateral cortical screw holes. Prophylactic trapping of the steel cable/wire distal to the screw hole will reduce the risk of fracture. Therefore, regardless of the choice of femoral fixation prosthesis, routine intraoperative radiographs are recommended after prosthesis placement. The condition of the greater trochanter is another important issue to consider. It may manifest as a separate, nonhealing bone mass or as an aberrant healing that impedes access to the medullary cavity during femoral preparation. In these patients, a slip osteotomy of the greater trochanter will help maintain the lateral femoral muscle, greater trochanter and adductor as a single tissue sleeve. However, patients should be informed preoperatively of the potential for permanent postoperative nonunion of the greater trochanter or painful greater trochanteric fixation device. A final technical aspect to consider is the proximal femoral deformity associated with fracture scab, fracture displacement, or deformity healing. The risk of femoral fracture may be increased during medullary preparation. The proximal femur is safer to shape with high-speed grinding and drilling than with rasp. Care must also be taken to avoid fractures of the proximal femur or penetration due to deflection of the reaming drill or open reamer by the sclerotic bone tract of a previously placed fixation device. In summary In most elderly patients, hip arthroplasty is a reliable salvage surgical option after failure of internal fixation therapy for intertrochanteric femoral fractures. Factors that influence decision making include the physiologic age of the patient, the remaining proximal femoral bone mass, the presence of deformities, the condition of the hip cartilage, and the viability of the femoral head. Hip arthroplasty after failed internal fixation treatment for femoral intertrochanteric fractures is technically more difficult than conventional initial total hip arthroplasty, and attention to specific technical details can improve success rates and reduce complications in treating these difficult problems.