The proximal femur and femoral diaphysis are common sites for primary bone tumors, including Ewing sarcoma, osteosarcoma, and chondrosarcoma. In addition, metastatic bone disease frequently affects the proximal femur. Carcinomas are the most common metastatic tumor in these patients and are generally managed with irradiation either instead of or after surgical management. In some patients with metastatic tumors who have massive bone loss and failed healing of pathologic fractures, replacement of bone may be the best option to relieve pain and improve function. Failed fixation of pathologic fractures is a common indication for proximal femur replacement. Proximal femur tumors and the surgery to remove them can both cause significant loss of bone, including tendon attachments. Reconstruction at this site should attempt to address these soft-tissue deficits as well as provide skeletal stability. Modular implants have become the standard endoprosthetic approach to reconstruct large bone defects. Allograft-prosthetic composite (APC) reconstructions are used in selected patients. Less common choices include resection only, arthrodesis, and, in skeletally immature patients, rotationplasty.

Proximal femur replacement has shown relatively good implant success. Loosening is less frequent than with cemented distal femur endoprostheses, and the failure rate from other causes such as infection has been relatively low compared with other megaprosthetic replacements. Dislocation has historically been a problem for these procedures, especially with loss of muscle about the hip, but is lowest with meticulous soft-tissue reconstruction and hemiarthroplasty. Trendelenburg gait is typical but not universal after these procedures.

When the proximal femur is resected, a posterolateral approach is used with the patient in the lateral position.² The extent of bone involvement must be determined with MRI and plain radiography to plan the distal resection level and assess whether any part of the greater trochanter can be left attached to the abductor tendons. This piece of greater trochanter provides length for the abductor mechanism and solid tissue to help anchor the abductors to the endoprosthesis. Hip joint involvement must be evaluated to determine if capsule can be spared and to plan the best hip reconstruction. Preserving the hip capsule to help secure a hemiarthroplasty is highly favored because the muscles that stabilize the hip are compromised. The extent of soft-tissue resection is determined by the extent of soft-tissue involvement as seen on MRI and by the need for negative or wide soft-tissue margins. In the setting of metastatic disease, negative soft-tissue margins may not be clinically relevant, and thus muscle and capsule should be spared as much as possible.

During the resection, several structures deserve particular attention once bone and soft-tissue resection decisions have been made. The sciatic nerve should be noted and protected. Branches of the lateral circumflex artery that supply the quadriceps muscle can be large and are more likely to be encountered if any part of the medial quadriceps needs to be resected. If the gluteus medius is compromised, the tensor fascia and the gluteus maximus are the remaining soft-tissue structures that cross the joint in this area and become important hip stabilizers. Reconstruction may involve suturing these muscles together and/or to the prosthesis.

Proximal femur reconstruction may be performed with an endoprosthesis or allograft, or both. Allograft use and success have been limited by availability, infection, and delayed diaphyseal healing but have the advantage of providing soft-tissue attachments and the ability to size the allograft intraoperatively. Allografts are often used as part of an APC and can be used to salvage failed hip replacements and failed tumor megaprostheses. A study of APC use for total hip revision showed the rate of complications requiring revision was 26.5%, including 6.9% with deep infections. The 10-year survival of the APCs was 69%.³ Another study of APCs for revision total hip arthroplasty in patients with developmental dysplasia noted 93% implant survival at 10 years.⁴

For patients with tumors who are undergoing segmental resection, another alternative is the use of the patient’s own bone (autoclaved or irradiated autograft) in the APC after the bone has been cleaned and irradiated during surgery. At one center where this technique was used to reconstruct the proximal femur, nonunion occurred in only 1 of 13 patients. The average time to heal was 20 weeks, and the mean Enneking score was 72.1%.⁵

Modular prostheses can require a minimum of 70 mm of resection from the proximal femur for implantation of the smallest construct (Figure 1). Smaller amounts of bone loss may be managed with the use of revision hip stems. It is helpful to keep the specimen on the back table when putting together trial prostheses to most accurately restore length. With loss of muscle, soft-tissue tension becomes a less reliable determinant of limb length. Various commercial guides use markers in the pelvis and femur to help assess correct length. Modular prostheses allow for almost any length up to and including a total femur arthroplasty. Implant systems typically require a special linkage piece for the total femur that joins hip and knee joint attachments.

Hip dislocation after proximal femur replacement has been a problem in part because of the amount of soft-tissue release and resection. Hemiarthroplasty is commonly used in proximal femur replacement primarily to prevent dislocation. The low rate of conversion of hemiarthroplasty to total hip arthroplasty may be related in part to compromised patient survival from the underlying conditions, patient populations with little baseline arthritis,
and/or use in patients who have undergone larger surgeries with muscle loss who self-limit their activity.⁶ However, the use of a unipolar replacement, which has a low dislocation rate, has been associated with a high risk of acetabular revision in patients younger than 21 years, with no hip implant in that age group lasting longer than 11 years in one study.⁷ Bipolar hip replacements or the more recently available large head total hip arthroplasties may provide more durability with acceptable stability in patients with longer life expectancy. In a series of 90 patients treated with bipolar hemiarthroplasty, 12 were converted to THA at a mean of 47 months. Patients age 35 to 65 years were at the highest risk for conversion.⁸ Preservation of the acetabulum in growing children is still advocated because it allows for acetabular growth with the acknowledgment that these children will likely need revision surgery at some point.⁹ One study highlights these points in primary bone malignancies, reporting risk of conversion from a hemiarthroplasty to a total hip arthroplasty at 10 years at only 10% but also noting that the risk of death at 10 years for these patients was 70%.¹⁰

Once the method of hip reconstruction has been chosen and the size of the prosthesis has been determined, the prosthesis must be secured to the bone. Cemented stems have long been preferred. One series had a revision rate of only 4.7% for loosening or osteolysis.¹¹ Loosening of cemented stems has been more of a problem with the distal femur than with the proximal femur. Noncemented stems have been widely available and are often used when good bone stock permits the use of a press-fit stem, in patients with relatively longer anticipated longevity, and in patients in whom postoperative radiation therapy is not anticipated. In a series of 25 patients using conical fluted stems, including three proximal femur replacements, there was no loosening or subsidence at a mean of 2.5 years; however, marked stress shielding was observed in 20% of stems.¹² Although not specific to the proximal femur, a novel noncemented bone interface for these prostheses has been developed that statically loads the cut end of the bone, leading to hypertrophy instead of bone loss from stress shielding. This compressive osseointegration interface also encourages bone growth onto the implant interface, sealing it from wear debris, but the biggest advantage may be the long-term stability of a biologic attachment of bone to implant. These implants require good bone stock but need less length for attachment than a typical noncemented stem. Risk of mechanical and overall failure of compressive osseointegration implants in one study is lower than reports of other stem types.¹³ In a 2021 series of 12 proximal femur replacements attached with compressive osteointegration, there was no aseptic loosening at a mean of 6 years of follow-up despite minimal bone hypertrophy in the compressed segment.¹⁴ Although some innovations are becoming widespread, not all options for bone ingrowth, bone-implant interfaces, and prosthetic expansion are available in any one implant format.