The proximal femur is a common site for primary bone tumors and the most common site for metastatic lesions.11,20 Before the development of endoprostheses in the 1970s, the primary treatment of these lesions was hip disarticulation or hindquarter amputation. With advances in radiation therapy and chemotherapy, limb salvage became an option in the early 1980s. One study suggests that though there appears to be a higher incidence of local recurrence, the overall patient survival is similar for amputation and limb salvage.21 The development of new operative techniques, better patient selection, and improved prosthetic design has been instrumental in improving the treatment options available.
The optimum method for reconstruction of the hip after resection of the proximal femur remains controversial. Multiple reconstructive options exist including autografts,2,12,22 allografts,8,9 custom megaprostheses,14,16,23 and modular endoprostheses. The first reconstructive procedures were performed using allograft or allograftprosthetic composites.4,10 Several studies suggest non-union, fracture, and infection are frequent complications associated with the use of allografts, and thus they may not be the ideal reconstructive option.8,9 These were followed by the development of custom endoprostheses, which while associated with fewer such complications are costly.14,16,23 More recently, modular endoprostheses for limb sparing procedures have permitted reconstruction of a wide variety of skeletal defects using standard components without the expense or time required to manufacture custom implants. Intraoperative assembly provides flexibility for the uncertainty of the margin of tumor resection, and an element of expandability when used for reconstruction in skeletally immature patients.
Improvements in treatment for primary bone neoplasms have led to increased long term survival of patients. Many of these patients are young and expected to lead active lives, placing greater demands on their implants. In addition, patients with metastatic bone disease are now surviving longer and many undergo reconstructions that will last for years rather than months. Because of these factors, implant durability is important in reducing the number of revisions patients will undergo over their lifetime.
We asked whether the survivorship of modular proximal femur endoprostheses would compare to that of custom implants and whether gender, age, diagnosis, or type of implant would predict failure. We additionally assessed the intermediate results with regards to complications and functional outcome.
MATERIALS AND METHODS
We retrospectively reviewed the database maintained by the Orthopaedic Oncology Service and identified 96 consecutive patients who underwent limb salvage with proximal femoral endoprosthetic reconstruction from July 1992 to November 2003. The inclusion criteria for this study included all patients who underwent reconstruction of the proximal femur with a modular system for either a metastatic lesion or primary bone tumor. Patients who underwent reconstruction with a custom implant, had revision of a preexisting implant, or had reconstruction for nonneoplastic disease were excluded from this analysis. All patients had been referred to the senior author for further management at our tertiary care institution. Forty-five patients were male and 51 were female. The mean age was 59 years (range, 14-86 years). Twenty-one patients (22%) had malignant primary bone tumors, three (0.03%) had benign bone tumors, and 72 (75%) underwent resection for metastatic disease (Table 1). The study design and protocol for human subjects were approved by the institutional review board.
At the time of the review, 73 patients (76%) had died of their disease, of which 60 succumbed to metastatic disease and 13 to primary malignant bone neoplasms. Fifty-three patients had a minimum 1 year followup, 22 patients had a 2-year followup, five patients were followed for 5 years, and one patient was followed for 10 years. Seven patients were followed for less than 6 months due to the patients succumbing to their disease. The average length of followup for patients with metastatic disease was 14 months (range, 1.5-86.2 months) because they tended to die sooner after their index procedure. In contrast, patients with primary bone tumors were followed for an average of 29 months (range, 1-129 months). The overall mean length of followup was 18.1 months (range, 1-129 months).
Patients with primary bone neoplasms were staged according to the staging system adopted by the Musculoskeletal Tumor Society (MSTS).7 Fifteen patients had Stage IIB disease at the time of diagnosis, five patients had Stage III disease, and one patient had Stage IB disease. The three patients with benign bone tumors were all Grade 3. When indicated, patients received the standard preoperative chemotherapy regimen in use at the time of their treatment.
All operative procedures were performed by the senior author (LRM). Surgical resection of the tumor was performed with the goal of obtaining adequate margins. Pathologic fracture was not considered a contraindication for performing limb salvage and reconstruction with an endoprosthesis. Wide resection with negative margins was obtained for all patients and all resections of the hip were intraarticular.
The Howmedica Modular Segmental Replacement System (MSRS) endoprosthesis (Rutherford, New Jersey) was used in all patients to reconstruct large segmental defects because of tumor resection. The articulating head segments were a bipolar endoprosthesis (62 patients) or a total hip arthroplasty (THA) (34 patients). Selection of head segment was based on preoperative expectation and intraoperative assessment of the acetabulum and surrounding soft tissue attachments. Our intention was to use bipolar head segments in all patients; however, when there was involvement of the acetabulum due to tumor or the presence of degenerative joint disease of the acetabulum, we opted to convert to a total joint arthroplasty. All implants were trialed prior to implantation, confirming intraoperative stability to 80° of internal rotation at both 0 and 90° of hip flexion and shuck of less than 0.5 cm. All femoral stems were fixed with a 1 to 2 mm polymethylmethacrylate cement mantle after reaming the host femur 2 mm larger than the stem diameter. The prosthesis was implanted with the femoral neck at neutral position with respect to an imaginary perpendicular line from the prosthesis and a line drawn from the linea aspera through the body of the prosthesis. No anteversion was built into the implant design. Acetabular components when used were cemented. Reconstruction of the abductor muscles was performed by suturing the tendons to the remaining host soft tissue without direct attachment to the endoprosthesis.
Postoperatively, patients were allowed full weightbearing as tolerated immediately after surgery. Due to the achievement of intraoperative stability throughout hip range of motion, we did not place patients in a brace postoperatively. Routine followup exams were performed every 3 months for the first 2 years postoperatively, then every 6 months between 2 and 5 years after treatment, and yearly thereafter.
Failure of the endoprosthetic reconstruction was defined as revision of any or all components of the implant, removal of the prosthesis, or amputation of the limb. Charts were reviewed for the onset of complications and causes of failure leading to revision including infection, tumor recurrence, aseptic loosening, fatigue failure, periprosthetic fracture, and dislocation. Results were compared to historical controls from the existing literature.
Functional outcome was assessed for all surviving patients using the revised MSTS functional rating system.6 This system uses a 30-point scale to equally weight each of six parameters, including pain, functional limitation, walking distance, use of support, emotional acceptance, and gait. Patients were interviewed by telephone or at their latest followup visit and administered the questionnaire by an investigator who was not involved in patient care to avoid any patient bias toward the treating surgeon.
Statistical analysis was performed using the Kaplan-Meier method to generate survivorship curves with corresponding 95% confidence intervals for implant and patient survival using the GraphPad Prism® software (San Diego, California). Patients who died with their original implant in place were censored. Time zero was defined as the date of implantation of the initial prosthesis. The endpoint was defined as the need for revision, removal of the implant, or amputation. The endpoint for limb survival was amputation. The Kaplan-Meier estimates included patients who had followup until their endpoint and all those patients in whom the initial implant was retained as of the date of termination of this study. The log-rank test and Cox proportional-hazards regression model were used to analyze the association between various clinical variables and implant failure, such as age, gender, diagnosis (metastatic versus primary bone tumors), and bipolar versus total hip arthroplasty implants. A p value of < 0.05 was considered significant.
Modular implant survivorship based on Kaplan-Meier time estimates was 94% at 2 years, 82% at 5 years, and 82% at 10 years. Overall, nine prostheses (9.3%) were considered failures at the time of latest followup (Table 2). The rate of revision was 7.3% (7/96), amputation 1% (1/96), and prosthesis removal was 1% (1/96). When compared with overall patient survival, the endoprostheses greatly outlived the patients in whom they had been implanted (p = < 0.001) (Fig 1). For those patients who underwent reconstruction for metastatic lesions implant survivorship was 96% at 2 and 5 years respectively. Patients with primary bone tumors had implant survivorship of 80% at 2 years and 70% at 5 years. There was a trend (p = 0.087) toward an increased rate of failure in those patients treated for primary bone neoplasms versus those treated for metastatic disease but with the limited power of the study was not significant. The average age of patients who ultimately had failed initial endoprosthetic reconstruction was 53 years (range, 23-80 years). The average interval between initial surgery and failure was 9.7 months (range, 1.5-36.5 months). Overall survivorship of the limb without amputation in this series of 99% at 5 years (Fig 2).
Gender, age, and diagnosis did not predict implant failure. The type of articulating head segment was, however, an indicator for failure, with bipolar implants having a greater (p = 0.018) survivorship than total hip endoprostheses (Figs 3). We found no difference in the rates of revision in those patients who died versus those who survived. At the time of death, 92% of patients had the original endoprosthesis in place. Of the patients who were alive at latest followup, 91% retained their original implants.
The leading cause of failure resulting in revision was instability. Overall, ten dislocations (10.4%) occurred with three resulting in revision of the implant. The remaining patients were successfully treated with closed (two patients) or open reduction (five patients). None of the patients with bipolar head segments underwent revision, while all three of the revisions for instability were in patients with THA head segments. Of the dislocations that occurred, four articulating head segments were bipolar and six were THAs. In six patients, dislocation occurred within two months of the index procedure as a result of a fall. Overall, the mean time from endoprosthetic replacement to dislocation was 4.7 months (range, 1-15.2 months).
Other complications which occurred in this study were infection (6.3%) and local recurrence (3.1%). Fatigue failure of the implant, periprosthetic fractures, and wound complications did not occur. Of the six patients who developed prosthetic infections, three were successfully treated with débridement alone, two patients underwent two-stage revision, and one patient underwent hip disarticulation. The majority of infections developed relatively soon after the initial endoprosthetic reconstruction with an average time from surgery to the onset of infection of 6 months (range, 2.5-17.1 months). Local recurrence occurred in three patients, one with recurrent hemangioendothelioma and two with recurrences of metastatic lesions (renal cell cancer, squamous cell cancer). Two of these patients underwent revision and the remaining patient had all hardware removed without reimplantation. The average time from initial resection to recurrence was 14.6 months (range, 2.3-36.5 months). At the time of this study, all patients with local recurrence had died of their disease. Overall, seventeen patients had a total of 19 complications for a complication rate in this series of 19.8%. There was no difference in the average age of those patients who developed complications (57 years) and those patients who did not (57.2 years). Patients with primary bone tumors had a 37.8% (9/24) rate of complications, while those with metastatic disease had only an 11% (8/72) complication rate.
The mean MSTS functional outcome scores for those patients surviving at the time of this study was 22 (range, 15-25), indicating they were able to achieve 73.3% (22/30) of normal function. Fifty nine patients (57%) had completed the questionnaire at their last followup visit, of which 21 (36%) had primary bone tumors and 38 (64%) had metastatic disease. The average followup of those patients completing the questionnaire was 21 months (range, 6-129 months). For those patients with metastatic disease the average MSTS score was 19 (range, 15-22). All patients were able to participate in activities of daily living and occupation-related activities. Ambulatory supports were used by 23 patients, all of whom had metastatic disease. Supports were used primarily for balance and generalized weakness due to progression of disease. The remaining 39 patients ambulated without the use of supports. Weakness in hip abductor strength accounted for the majority of functional deficit.
Modular segmental reconstruction has been used exclusively at the authors' institution for reconstruction after resection of proximal femur neoplastic disease since 1992. The benefits of modular implants include intraoperative assembly allowing for flexibility when greater resection than previously anticipated is necessary to achieve negative margins, the possibility of expanding the implants in skeletally immature patients to account for growth, and decreased cost in comparison with custom implants. This series is among the largest to date reporting the results of a single type of modular endoprosthesis. Previous studies have combined their results of a variety of both modular and custom-made implants, or have reported on their experience with the use of custom implants alone.
Limitations of this study include the fact that many patients did not have long-term followup due to the large numbers of patients with metastatic disease who tended to succumb to their disease within 2 years after the procedure. However, because of the large number of metastatic patients, this study does allow for analysis of modular endoprosthetic reconstruction in this specific group. Therefore, although followup is limited due to the nature of the disease, we were still able to show for these patients, who often live less than 24 months after surgery, modular implants have acceptable rates of survivorship. The reconstruction outlived the patients and was a durable option that lasted the remainder of these patients' lives.
The implant survivorship rate, with 5-year and 10-year projected survival of 82% overall, compares favorably to the results of custom implants. In a study of 81 patients treated with custom-made endoprosthetic replacement of the proximal femur five-year and 10-year survivorship of the implant was reported as 73% and 63%, respectively.3 Another study reported the results of 16 patients who underwent implantation of a custom or standard proximal femoral endoprosthesis.13 Of these, 14 reconstructions were bipolar and two had a cemented acetabulum. Implant survivorship was 88% at 5 years and 65% at 10 years. In a study of 232 patients who underwent custom proximal femoral replacement, survivorship was 68% at 10 years.23 These results indicate modular implants are an acceptable alternative to custom reconstructions with comparable rates of implant survivorship. In the present study, it must be taken into account many patients did not survive for 5 or 10 years because the majority of patients underwent reconstruction for metastatic disease. However, although patient survival was only 20% at 5 years, the majority of those patients retained their original reconstruction, indicating these implants will generally outlive the patients in whom they are implanted.
Instability remains a major complication associated with large segmental reconstruction of the proximal femur. Several factors influence the rate of dislocation including acetabular resurfacing, joint capsule resection, and the technique of abductor repair. A slightly greater rate of dislocation has been reported after acetabular resurfacing. An 11% rate of dislocation was reported in a study of 54 patients who underwent endoprosthetic reconstruction of the femur with acetabular resurfacing, joint capsule preservation, and repair of the hip abductors to the fascia lata.14 Similarly, in a series of 45 patients who underwent proximal femoral endoprosthetic reconstruction with acetabular resurfacing and abductor attachment to the implant in almost all patients, a 15.5% rate of dislocation was reported.19 In our series, dislocation was almost three times more common after THA than after bipolar replacement. The use of a bipolar endoprosthesis may decrease the risk of this complication as the larger head size appears to confer a greater degree of stability. Bickels et al1 reported the results of 57 patients who underwent bipolar or unipolar reconstruction of the proximal femur with acetabular preservation, capsular repair, and anatomic reconstruction of the abductor mechanism to the endoprosthesis. Only one patient (1.7%) in that series developed a dislocation. An additional contributing factor predisposing to instability is abductor weakness. Tumor resection about the hip often requires the abductor muscles be sacrificed to the point that reconstruction to the endoprosthesis is not possible. In our series, the abductors were not directly reattached to the prosthesis but instead were sutured to the remaining soft tissues.
Infection occurred in 6.3% of patients in this study, which is similar to other series previously reporting rates ranging from 3% to 13%.5,15,16 This is still considerably less than reported after allograft reconstruction with rates as high as 30%.9 The higher incidence of infection with endoprostheses compared with arthroplasty performed for degenerative disease may be related to several factors, including the magnitude of soft tissue resection and use of perioperative adjuvant chemotherapy leading to immunosupression. Malawer and Chou15 proposed administering long-term suppressive antibiotics during chemotherapy and other periods of immunosuppression to potentially decrease the rate of infection. This proposal is controversial, but additional investigation may be worthwhile.
The local recurrence rate in this series of 3.1% compares favorably with that reported in other studies. Previous reports noted recurrence rates ranging from 3% to 25%.15,18 Risk factors associated with increased tumor recurrence after resection of malignant bone and soft tissue neoplasms include marginal resection, pathological fracture, poor response to chemotherapy, and intravascular tumor extension. The rate of limb salvage in our series was 99%, which was most likely the result of careful preoperative planning, aggressive tumor resection, and improvements in chemotherapy and radiation. We believe the most important factor in obtaining local tumor control and limb preservation is achieving negative margins by wide excision.
No patients in this study developed aseptic loosing, mechanical failure, or periprosthetic fractures. This improved durability may be related to advances in cementing techniques and improvements in prosthetic design and metallurgy. Aseptic loosening has been minimized by several factors, including modern cementing techniques of pressurization and the use of an intramedullary cement plug, which were routinely used in this series. Extramedullary porous ingrowth surfaces on more recently developed implants allow for soft tissue ingrowth and the option for extracortical bone graft allowing for load sharing with the stem. The low rate of aseptic loosening in this study, however, also must take into account many patients died within five years after their procedure. With longer followup the rate of aseptic loosening may have increased, but for the patients in this study the implant was durable enough to outlive most patients. Additionally, rates of mechanical failure have been previously been reported to range from 10% to 16%.5,24 The decreased incidence of stem fatigue failure can be related to several factors in improved implant design, including the use of forged rather than casted stems, which has led to an increase in stem strength and the use of stronger alloys.17 Periprosthetic fractures have been reduced by switching from straight to curved intramedullary stems. The current advances in implant design and modes of fixation may help to explain why no patients in this series underwent revision because of implant-related failure or aseptic loosening.
Our study supports the role of modular endoprosthetic replacement of the proximal femur as a durable reconstructive option after tumor resection. For most patients with neoplastic disease, these implants will generally outlive the patient. Based on the results of this series, bipolar reconstruction appears to have a greater survivorship over total hip articulating segments and should be the preferred treatment whenever possible.
1. Bickels J, Meller I, Henshaw RM, Malawer MM. Reconstruction of hip stability after proximal and total femur resections. Clin Orthop Relat Res
2. Chen WM, Chen TH, Huang CK, Chiang CC, Lo WH. Treatment of malignant bone tumors by extracorporeally irradiated autograft- prosthetic composite arthroplasty. J Bone Joint Surg Br
. 2002;84: 1156-1161.
3. Dobbs HS, Scales JT,Wilson JN. Kemp HB, Burrows HJ, Sneath RS. Endoprosthetic replacement of the proximal femur and acetabulum. J Bone Joint Surg Br
4. Donati D, Giacomini S, Gozzi E, Mercuri M. Proximal femur reconstruction by an allograft prosthesis composite. Clin Orthop Relat Res
5. Eckardt JJ, Eilber FR,Rosen G. Mirra JM, Dorey FJ, Ward WG, Kabo JM. Endoprosthetic replacement for stage IIB osteosarcoma. Clin Orthop Relat Res
6. Enneking WF, Dunham W, Gebhardt MM, Malawer MM, Pritchard DJ. A system for functional evaluation for reconstructive procedures after surgical treatment of tumors of the musculoskeletal system. Clin Orthop Relat Res
7. Enneking WF, Spanier SS, Goodman MA. A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop Relat Res
8. Fox EJ, Hau MA, Gebhardt MC, Hornicek JF, Tomford WW, Mankin HJ. Long-term followup of proximal femoral allografts. Clin Orthop Relat Res
9. Gebhardt MC, Flugstad DI, Springfield DS, Mankin HJ. The use of bone allografts for limb salvage in high-grade extremity osteosarcoma. Clin Orthop Relat Res
10. Gitelis S, Piasecki P. Allograft prosthetic composite arthroplasty for osteosarcoma and other aggressive bone tumors. Clin Orthop Relat Res
11. Harrington KD. Orthopaedic Management of Metastatic Bone Disease
. St. Louis, MO: Mosby; 1988.
12. Harrington KD, Johnston JO, Kaufer HN, Luck JV Jr, Moore TM. Limb salvage and prosthetic joint reconstruction for low-grade and selected high-grade sarcomas of bone after wide resection and replacement by autoclaved autogeneic grafts. Clin Orthop Relat Res
13. Horowitz SM, Glasser DB, Lane JM, Healey JH. Prosthetic and extremity survivorship after limb salvage for sarcoma. How long do the reconstructions last? Clin Orthop Relat Res
14. Kabukcuoglu Y, Grimer RJ, Tillman RM, Carter SR. Endoprosthetic replacement for primary malignant tumors of the proximal femur. Clin Orthop Relat Res
15. Malawer MM, Chou LB. Prosthetic survival and clinical results with use of large-segment replacements in the treatment of high- grade bone sarcomas. J Bone Joint Surg Am
16. Mittermayer F, Krepler P,Dominkus M. Schwameis E, Sluga M, Heinzl H, Kotz R. Long-term followup of uncemented tumor endoprostheses for the lower extremity. Clin Orthop Relat Res
17. Plotz W, Rechl H,Burgkart R. Messmer C, Schelter R, Hipp E, Gradinger R. Limb salvage with tumor endoprostheses for malignant tumors of the knee. Clin Orthop Relat Res
18. Quill G, Gitelis S, Morton T, Piasecki P. Complications associated with limb salvage for extremity sarcomas and their management. Clin Orthop Relat Res
19. Rechl H, Reinisch M, Plotz W, Burgkart R, Gradinger R. Soft tissue reconstruction about the proximal femur. Oper Techn Orthop
20. Schneiderbauer MM, von Knoch M,Schleck CD. Harmsen WS, Sim FH, Scully SP. Patient survival after hip arthroplasty for meta- static disease of the hip. J Bone Joint Surg Am
21. Simon MA, Aschliman MA, Thomas N, Mankin HJ. Limb-salvage treatment versus amputation for osteosarcoma of the distal end of the femur. J Bone Joint Surg Am
22. Smith WS, Struhl S. Replantation of an autoclaved autogenous segment of bone for treatment of chondrosarcoma. Long-term follow- up. J Bone Joint Surg Am
23. Unwin PS, Cannon SR, Grimer RJ, Kemp HB, Sneath RS, Walker PS. Aseptic loosening in cemented custom made prosthetic replacements for bone tumors of the lower limb. J Bone Joint Surg Br
24. Wirganowicz PZ, Eckardt JJ, Dorey FJ, Eilber FR, Kabo JM. Etiology and results of tumor endoprosthesis revision surgery in 64 patients. Clin Orthop Relat Res