Proximal femoral replacement prostheses have been used extensively in the reconstruction of proximal femoral deficiencies after oncologic resection and in revision hip arthroplasty. When these prostheses fail, options for subsequent salvage of the hip are limited because of the loss of proximal femoral bone stock. Resection arthroplasty is unsatisfactory in patients with extensive bone loss of the proximal femur because of limb shortening, limited ambulatory ability, and persistent pain.4 Salvage options in these complex cases include the use of another cemented proximal femoral replacement prostheses or reconstruction with an allograft prosthesis composite. Long term results of proximal femoral replacement prostheses in revision hip arthroplasty have shown high failure and complication rates.7,10 In distinction, the early and midterm results of allograft prosthesis composites in revision arthroplasty have been more favorable.5,8
The current study reviews the early to midterm results (mean of 5.6 years) in patients who underwent reconstruction of a failed femoral megaprosthesis with an allograft prosthesis composite. Hip salvage in these patients is among the most complex problems encountered in revision hip surgery because of the extensive nature of the proximal femoral bone loss and severe disruption of the soft tissue attachments about the proximal femur.
MATERIALS AND METHODS
The Total Joint and Bone Bank Registries at the authors' institution were reviewed retrospectively for all years before 1995. All patients who had undergone revision of a failed femoral megaprosthesis with implantation of an allograft prosthesis composite were identified. Eleven patients fit these criteria, and all had undergone surgery between May 1986 and October 1993. The mean age at surgery of the three men and eight women in this cohort was 59 years (range, 43-73 years). Before reconstruction with the allograft prosthesis composite these patients had undergone a mean of 4.4 previous hip operations (range, 2-9 operations). The underlying diagnosis, type of failed femoral megaprosthesis, and indication for surgery for each patient are detailed in Table 1. The most frequent indication for surgery was aseptic loosening (nine patients).
The mean clinical followup was 67 months (range, 33-96 months) for the seven patients in whom the allograft prosthesis composite had not been removed. Four patients had undergone removal of the allograft prosthesis composite at a mean of 16 months (range, 5-41 months). Clinical followup for the entire group of 11 patients was 49 months (range, 5-96 months). Mean radiographic followup of 65 months (range, 35-92 months) was available for five of seven patients at latest followup.
Proximal femoral bone loss was categorized preoperatively according to the American Academy of Orthopaedic Surgeons classification.3 Radiographs were available for all 11 patients. Of these, 10 had Type III combined segmental and cavitary defects and one patient had a Type I segmental defect. The mean length of the previous proximal resection as measured from the center of the femoral head was 11 cm (range, 7.5-23 cm). The greater trochanter had been resected previously in eight patients. In two patients an ununited greater trochanteric fragment was present and in one patient the trochanter was intact.
The surgical technique used varied during the study period because of the extent of the proximal bone loss, the quality of the remaining host bone, surgeon preference, and evolving techniques. A direct lateral approach was used in all patients with skeletonization of the proximal femur. The specific details of each reconstruction including allograft length, prosthesis type, prosthesis fixation in the allograft and host bones, and supplemental fixation at the allograft-host junction are summarized in Table 2.
Clinical data were collected prospectively before surgery, and at the latest followup. Radiographs obtained preoperatively, immediately postoperatively, and at latest followup were reviewed. Radiographic criteria for femoral loosening were defined as previously described.11 Heterotopic ossification was graded according to the classification of Brooker et al.1
The allograft prosthesis composite was removed in four of 11 patients at a mean of 16 months. Kaplan-Meier survival estimates representing the cumulative probability of remaining free of revision were 73% at 3 years (95% confidence interval, 0.48 to 1.00) and 61% at 5 years (95% confidence interval, 0.36 to 1.00).
Pain scores were documented for all 11 patients preoperatively and for all seven patients at latest followup who had not undergone subsequent revisions of the allograft prosthesis composite. Preoperatively, all patients had moderate or severe pain; whereas, at latest followup, six of seven patients had no pain or only slight pain (Table 3).
Gait aid requirements were documented for 10 of 11 patients preoperatively, and for all seven who had not undergone subsequent revisions. Preoperatively, eight patients had required the use of two crutches or a walker and two required a cane or single crutch. At the latest followup, three patients used either no gait aid or required only a cane for long walks. Three other patients used either a cane or a single crutch full time and only one patient required the use of two crutches (Table 3).
At latest followup, one patient was able to ambulate between 4 and 6 blocks, and six patients were able to ambulate between 1 and 3 blocks.
Radiographic union occurred at the junction site between the allograft and host bone in all eight patients in whom the prosthesis was still in place at 1 year followup (range, 8-16 months).
Radiographs were available at the latest followup for five patients (mean, 5.4 years; range, 2.9-7.7 years). Femoral component subsidence occurred in two patients (29%). Subsidence of 20 mm occurred in one patient who had the prosthesis implanted without cement into the allograft and native bone (Case 2). A second prosthesis which also was uncemented into the allograft and did not cross the allograft-host bone junction had subsided 30 mm (Case 7). In this case subsidence had occurred in the presence of massive allograft lysis measuring 80 mm in Gruen Zone 1 and 30 mm in Gruen Zone 7.6 Both of these prostheses were considered to be definitely loose based on radiographic criteria.11 No other prosthesis met radiographic criteria for probable or possible loosening and no other cases of allograft lysis had occurred (Fig 1). Heterotopic ossification occurred in three patients (43%). Of these, two patients (29%) had Class I ossification develop (Cases 4 and 7) and one patient (14%) had Class III ossification develop (Case 2).1
Additional surgery was performed after the implantation of the allograft prosthesis composite in five patients (45%) and included removal of the allograft prosthesis composite in four cases (Table 4). Deep infection developed in two patients (18%) at 2 months and 13 months, respectively (Cases 3 and 9). Both patients were treated with resection arthroplasty and delayed reimplantation. One hip was reconstructed with implantation of a cemented modular proximal femoral replacement prosthesis (Case 3) and the other was salvaged with implantation of a saddle prosthesis (Case 9). Two patients (18%), both of whom had a short prosthesis that did not bypass the allograft-host bone junction, suffered periprosthetic fractures of the allograft at 2 months and 41 months, respectively (Cases 1 and 11 [Fig 2]). In both patients the hip was reconstructed with implantation of a second allograft prosthesis composite.
Hip instability developed in three patients (27% [Cases 5, 9, and 10]). Two patients (18%) had recurrent episodes of subluxation and dislocation beginning at 3 months and 12 months, respectively (Cases 5 and 10). The first patient was managed nonoperatively and continued to have persistent episodes of subluxation (Case 5). The second patient was treated successfully with revision of the acetabular component and implantation of a constrained liner (Case 10). The other patient (Case 9) experienced only one dislocation at 12 months and was treated nonoperatively. However, this patient had one of the deep infections develop, as mentioned before, that resulted in resection arthroplasty and delayed reimplantation. The other major complication seen in this cohort occurred in the patient who had undergone transfer of the autologous vascularized fibula (Case 7). Intraoperative thrombosis of the ipsilateral femoral artery occurred but was treated successfully with thrombectomy.
Revision hip arthroplasty after failure of a femoral megaprosthesis is a complex problem. Resection arthroplasty in patients with extensive proximal bone loss is a poor solution because of persistent pain, limb shortening, and poor ambulatory ability.4 In addition to the extensive bone defects, previous violation of the soft tissue attachments about the hip, in particular the abductor mechanism, complicates successful reconstruction. Options for hip salvage include reimplantation of another proximal femoral replacement prosthesis or reconstruction with an allograft prosthesis composite. Long term results of the use of proximal femoral replacement prostheses in revision hip arthroplasty have been poor with a predicted survival rate free from revision of 64% at 12 years.10 The same authors reported high complication rates including dislocation in 22% of patients (11 of 50) and deep infection in 6% of patients (three of 50).10 Early and midterm results of the use of allograft prosthesis composites in general revision hip arthroplasty have been cautiously optimistic.2,5,8,9 A revision rate of 10.1% was reported in 168 allograft prosthesis composites at an average followup of 4.8 years.5 However, significant complications also have been associated with these techniques and in the same series a dislocation rate of 5.3% and an infection rate of 3% were reported.5 Infection rates of 17% (three of 18) and 7.7% (one of 13) have been reported by other authors.9,12
The extensive proximal bone defects and significant disruption of the soft tissues about the hip encountered in the patients in the current series of failed femoral mega-prostheses were among the most severe examples of these problems encountered in revision hip arthroplasty. Therefore, the high complication and failure rates reported in the current series, which exceed previously documented rates in general revision hip arthroplasty, might have been anticipated. In particular, instability and infection may be caused by the previous violation of the soft tissue attachments about the hip resulting from implantation of the previous proximal femoral replacement type prosthesis, the significant soft tissue stripping involved in skeletonization of the proximal femur, the large size of the allografts used in these reconstructions, and the multiple prior operations that these patients had undergone. In addition, none of these patients had even a thin layer of native proximal bone to enhance graft healing and soft tissue repair.
Failures caused by loosening and fracture might have been predicted based on the early techniques for allograft prosthesis composite reconstruction used in some of the cases in this series. The two periprosthetic allograft fractures were likely attributable to failure to bypass the allograft-host bone junction site with a long stem prosthesis and the use of plate and screw fixation (Fig 2). Both of these technical factors create stress risers in the allograft bone that can lead to catastrophic failure of the construct. The use of uncemented prostheses in the proximal allograft bone where ingrowth cannot be expected to occur probably contributed to the two cases of subsidence and radiographic failure seen at latest followup. The use of contemporary techniques for allograft prosthesis composite reconstruction including cementation of a long stem prosthesis into the allograft, bypassing of the allograft-host bone junction with the prosthesis stem, and rigid distal fixation in host bone (Fig 3) may eliminate some of the technical errors seen in this series.2,5
Despite significant complications and failures seen in several of the patients in the current study, encouraging clinical results were observed at early to midterm followup in those patients in whom the construct had not been removed. Improvements in pain and walking ability have been documented. At latest followup, all seven patients in whom the allograft prosthesis composite was still in place remained ambulatory outside of the home with the use of appropriate gait aids.
The management of femoral bone loss and disrupted soft tissue attachments about the proximal femur associated with failure of a femoral megaprosthesis represent complex problems. Despite the high complication and revision rates seen in the current series, continued use of allograft prosthesis composites in this patient population is appropriate. Furthermore, with the experience gained from this preliminary series and with advances in the understanding of allograft prosthesis composite construction gained in other published series improved outcomes should be anticipated.
The authors thank Dirk R. Larson, MS, Section of Biostatistics, Mayo Clinic, for the statistical analysis performed in this study.
1. Brooker AF, Bowerman JW, Robinson RA, Riley Jr LH: Ectopic ossification following total hip replacement. Incidence and a method of classification. J Bone Joint Surg 55A: 385-389, 1992.
2. Chandler HP, Ayres DK, Tan RC, Anderson LC, Varma AK: Revision total hip replacement using the S-ROM femoral component. Clin Orthop 319:130-140, 1995.
3. D'Antonio J, McCarthy JC, Bargar WL, et al: Classification of femoral abnormalities in total hip arthroplasty. Clin Orthop 296:133-139, 1993.
4. Grauer JD, Amstutz HC, O'Carroll PF, Dorey FJ: Resection arthroplasty of the hip. J Bone Joint Surg 71A:669-678, 1989.
5. Gross AE, Hutchison CR, Alexeeff M, et al: Proximal femoral allografts for reconstruction of bone stock in revision arthroplasty of the hip. Clin Orthop 319:151-158, 1995.
6. Gruen TA, McNiece GM, Amstutz HC: "Modes of failure" of cemented stem-type femoral components. A radiographic analysis of loosening. Clin Orthop 141: 17-27, 1979.
7. Haentjens P, DeBoeck H, Opdecam P: Proximal femoral replacement prosthesis for salvage of failed hip arthroplasty. Acta Orthop Scand 67:37-42, 1996.
8. Head WC, Berklacich FM, Malinin TI, Emerson RH: Proximal femoral allografts in revision total hip arthroplasty. Clin Orthop 225:22-36, 1987.
9. Jofe MH, Gebhardt MC, Tomford WW, Mankin HJ. Reconstruction for defects of the proximal part of the femur using allograft arthroplasty. J Bone Joint Surg 70A:507-516, 1988.
10. Malkani AL, Settecerri JJ, Sim FH, et al: Long-term results of proximal femoral replacement for non-neoplastic disorders. J Bone Joint Surg 77B:351-356, 1995.
11. McLaughlin JR, Harris WH: Revision of the femoral component of a total hip arthroplasty with a calcar replacement femoral component: Results after a mean of 10.8 years postoperatively. J Bone Joint Surg 78A:331-339, 1996.
© 1998 Lippincott Williams & Wilkins, Inc.
12. Zehr RJ, Enneking WF, Scarborough MT: Allograft-prosthesis composite versus megaprosthesis in proximal femoral reconstruction. Clin Orthop 322:207-223, 1996.