From September 2002 to October 2003, a porous tantalum metaphyseal cone was used to reconstruct a severe tibial defect during revision total knee replacement in fifteen patients (fifteen knees). All surgical procedures were performed by the two senior authors (D.G.L. and A.D.H.). This study was approved by the institutional review board at our hospital. All study patients provided informed consent for participation in the study, and the metaphyseal cones were approved by the Food and Drug Administration prior to use.
There were eight women and seven men with an average age of 68.1 years (range, forty-one to eighty-one years) at the time of the revision surgery with the porous metaphyseal cones. There were eight left and seven right knees. The patients had had an average of 3.5 total knee replacements (range, one to eight replacements) prior to the revision procedure with use of the porous tantalum metaphyseal cone. The indications for the revision procedure included a second-stage reimplantation for deep infection (five knees), aseptic loosening of the tibial component (four knees), severe tibial osteolysis in the presence of a well-fixed tibial component (three knees), fracture of the tibial component (two knees), and severe global knee instability with associated bone loss (one knee).
Tibial bone loss was categorized according to the Anderson Orthopaedic Research Institute bone defect classification19. On the basis of an intraoperative assessment, eight knees were categorized as having a Type-3 defect and seven knees as having a Type-2B defect. Type-2B defects include moderate-to-severe cancellous bone defects of both tibial plateaus and/or segmental cortical defects of one tibial plateau, and Type-3 defects have combined cavitary and segmental bone loss of both tibial plateaus (Figs. 1-A through 1-F).
All patients were followed prospectively at three and twelve months postoperatively and annually thereafter, and knee motion was estimated by the evaluating surgeon. Knee function was assessed preoperatively and postoperatively according to the Knee Society clinical rating system21. Immediate postoperative anteroposterior and lateral knee radiographs were made and analyzed along with radiographs made at three months, twelve months, and annually thereafter. All radiographs were assessed by an orthopaedic surgeon who was not one of the operating surgeon-developers. The radiographic assessment was performed according to a previously reported modification22 of the Knee Society total knee arthroplasty radiographic evaluation system23 for long-stemmed revision prostheses. The interface between the porous tantalum cone and the host bone was also assessed for initial and progressive radiolucencies as well as areas of initial radiolucencies that resolved with continued follow-up. Clinical and radiographic follow-up averaged thirty-four months (range, twenty-four to forty-seven months), and no patient was lost to follow-up.
Porous Tantalum Metaphyseal Cones
Porous tantalum (Trabecular Metal; Zimmer, Warsaw, Indiana) is a biomaterial with mechanical and biological properties that include low stiffness, high porosity, and a high coefficient of friction24. The excellent potential for biologic fixation with porous tantalum has been demonstrated in a wide variety of adult reconstructive applications5. The porous tantalum metaphyseal cones used in this study were designed and developed by the two senior authors in conjunction with Zimmer.
The cones were developed in various shapes and sizes to accommodate the variety of bone defects encountered as well as to match the size of the proximal tibial metaphyseal region. The basic shapes include full and stepped designs of various widths and heights. The cones were designed to be impacted into the proximal part of the tibia to allow for osseous ingrowth from the adjacent host bone and simultaneously provide for support of a revision stemmed tibial component. The prostheses used in conjunction with these porous tantalum cones in this series included seven hinged (Rotating Hinge Knee [RHK]; Zimmer), six posterior stabilized (Legacy PS [Zimmer] in five and a Sigma PS [DePuy, Warsaw, Indiana] in one), and two constrained condylar knee designs (LCCK; Zimmer).
Surgical Technique
The surgical technique included intraoperative assessment of the bone defect, use of an intramedullary guide to facilitate proper alignment and insertion of the metaphyseal cone, use of a variety of trial sizes and shapes to guide careful contouring of prominent areas of tibial bone with a high-speed burr, and final impaction of the porous tantalum cone (Fig. 1-D) with size-specific impactors. The rotation of these implants was guided by the size, shape, and location of the bone defects and was not related to the final rotation of the tibial prosthetic component. Once the porous tantalum cone had been impacted into a stable position, the internal surface of the cone provided a receptive surface for cementation of the tibial implant. The stability of the porous tantalum cone was assessed during the impaction to ensure there was no inducible axial subsidence by means of manual force in its final position. Rotational stability was assessed and confirmed with manual manipulation by the surgeon and was obtained by the intact peripheral tibial bone stock.
Any areas or voids between the periphery of the porous tantalum cone and the adjacent bone of the proximal part of the tibia were filled with morselized cancellous bone graft (seven knees) (Fig. 1-E) or demineralized bone matrix (DBX; Synthes, Paoli, Pennsylvania) (five knees) to prevent any egress of bone cement between the cone and the host bone during cementation of the stemmed tibial component. The revision tibial component was then inserted through the cone with use of either cementless25,26 or cemented stem extensions27. Twelve stems were cemented, and three stems were press-fit without cement into the tibial diaphysis with insertion of cement only in the porous tantalum cone and the remaining proximal metaphyseal portion of the tibia. Antibiotics were used in the cement (Simplex P; Stryker, Mahwah, New Jersey) in all knees. Gentamicin (1.2 g per cement batch) was used in knees with aseptic loosening of the tibial component, and a combination of gentamicin (1.2 g) and vancomycin (1 g per cement batch) was used in knees that had second-stage reimplantation following infection. The tibial stem extensions ranged from 30 to 155 mm. Regardless of the stem fixation type, polymethylmethacrylate was placed between the porous cone and the tibial tray and the proximal keel of the tibial component, in a fashion analogous to metaphyseal cementation26, in order to join the stemmed tibial implant mechanically to the porous cone.
Clinical Function
Preoperative range of motion in all fifteen patients consisted of an average flexion contracture of 6.2° (range, 0° to 35°) and an average flexion of 86.5° (range, 30° to 120°). At the time of the final follow-up, the range of motion had improved to an average residual flexion contracture of 1° (range, 0° to 10°) and an average flexion of 99.7° (range, 80° to 120°).
When analyzed according to diagnosis, the ten knees with aseptic failure had an average extension of 3.3° (range, 0° to 11°) and an average maximum flexion angle of 97.7° (range, 80 to 120 points) preoperatively. At the time of the final follow-up, all of these knees had full extension and the maximum flexion angle was an average of 104° (range, 90° to 100°). In contrast, the five knees in patients who were being treated for infection had an average extension of 9.8° (range, 0° to 35°) and an average maximum flexion angle of 73.2° (range, 30° to 90°) preoperatively. At the time of the final follow-up, the extension in these five knees averaged 2.1° (range, 0° to 10°) and the maximum flexion angle averaged 93.4° (range, 80° to 100°).
Overall, the average Knee Society clinical score was 52 points (range, 25 to 81 points) preoperatively and improved to 85.2 points (range, 49 to 100 points) at the time of the final follow-up. Again, when considered according to diagnosis, the knee scores for the five patients treated for infection averaged 43 points (range, 25 to 47 points) preoperatively and improved to 79.6 points (range, 49 to 100 points) at the time of the final follow-up. In the ten patients with a diagnosis of aseptic loosening, the average knee score was 60 points (range, 25 to 81 points) preoperatively and improved to 90.5 points (range, 79 to 100 points) at the time of the final follow-up.
Radiographic Analysis
Preoperatively, the average anatomic tibiofemoral angle for the knees in varus alignment on the anteroposterior radiographs was 11.3° (range, 7° to 19°) and improved to 3.3° (range, 0° to 5°) of varus alignment postoperatively. For the knees in valgus alignment, the average tibiofemoral angle was 4.5° (range, 0° to 9°) preoperatively and improved to 5.4° (range, 2° to 8°) of valgus alignment postoperatively. On the immediate postoperative anteroposterior radiographs, there were three incomplete radiolucent lines at the bone-cement interface adjacent to the tibial stem extensions. Two of these radiolucencies measured <2 mm, and one measured 3 mm. At the time of the final follow-up, all three radiolucent lines were stable and nonprogressive and no other radiolucencies were noted at the time of the final follow-up.
On the immediate postoperative radiographs, all fifteen porous cones appeared to be closely apposed to the adjacent host bone of the proximal tibial metaphysis. At the time of the final follow-up, all fifteen cones had evidence of osseointegration with evidence of reactive osseous trabeculation at their points of osseous contact (Fig. 1-F). No radiolucencies were observed between the cones and the adjacent tibia at the time of the final follow-up, and this finding was considered evidence of osseointegration at those points identified on the radiographs. There was no evidence of loosening or migration of any of the fifteen tibial reconstructions at the time of the final follow-up.
Reoperation
Four patients required a reoperation. Two patients who had a recurrent deep infection develop at eighteen months postoperatively were treated with débridement, polyethylene liner exchange, and retention of the prosthetic components with subsequent chronic suppression with oral antibiotics. The decision to pursue component retention and chronic suppression was based on the anticipated technical difficulty and potentially severe bone loss involved with removing well-fixed cemented stems, identification of an organism amenable to antibiotic suppression, and the suboptimal general medical health of the patients. At the time of reoperation, both patients had well-fixed prosthetic components and the porous tantalum metaphyseal cones had excellent circumferential osseointegration with the surrounding tibial bone. Both patients had poor, yet functional, knee scores of 63 and 49 at the time of the last follow-up.
The third patient who required a reoperation initially had a revision with a hinged knee prosthesis with use of uncemented femoral and tibial stems, as well as a tantalum tibial cone. At one year postoperatively, because of progressive pain in the distal aspect of the thigh, the patient underwent an isolated femoral revision for aseptic loosening of the femoral component. The metaphyseal tantalum cone and the tibial component were both well fixed. At one year postoperatively, the patient had no pain and a knee score of 100.
The fourth patient sustained an acute periprosthetic tibial fracture during a fall at thirty-nine months postoperatively. The patient had undergone the index revision with a stepped metaphyseal cone for a Type-3 tibial defect and had a satisfactory clinical outcome at the latest follow-up evaluation prior to the periprosthetic fracture. The intraoperative findings at the time of the revision for the periprosthetic fracture revealed that the entire medial aspect of the metaphyseal cone was well fixed and was in continuity with the medial fracture fragment. The lateral aspect of the porous tantalum cone was detached from the lateral tibial cortex, yet a portion of the lateral cone had osseous ongrowth suggestive of osseointegration. The porous tantalum cone and medial fracture fragment were removed, a tibial component with a thicker polyethylene insert was implanted with cement, and the well-fixed femoral component was retained.
The number of failed total knee replacements with large bone defects is increasing in part because of the growing number of these procedures being performed in a more active patient population28-30. Currently, many of these bone defects are related to osteolysis secondary to particulate wear debris31. Osteolytic lesions associated with particulate wear debris frequently result in an unpredictable pattern of bone lesions that are often extremely large despite the fact that many are asymptomatic31-33. The absolute size of many osteolytic lesions is often underestimated preoperatively31,32,34. This uncertainty can complicate preoperative planning, as it is necessary to have a large variety of implants and structural bone grafts available for the revision surgical procedure6,35.
In the present study, the Type-2B and Type-3 tibial bone defects were severe and would have required the use of structural allograft bone3,6,7, the use of wire mesh combined with impaction bone-grafting15,16, or the use of custom-designed implants. Modular augments in current revision knee systems may not effectively address severe bone loss and the instability frequently encountered during revision surgery6. In a previous report, despite the use of metallic augmentation in 89% of sixty-five consecutive revision knee arthroplasties, large structural allografts were still required in 48% of the knees6. Currently, little is known about the long-term outcome of structural bone allografts used in conjunction with revision total knee arthroplasty. One report described fifty-two revision total knee replacements performed in conjunction with sixty-six structural bone grafts7. Twelve knees (23%) had a repeat revision at a mean of 70.7 months (range, twenty-six to 157 months). The allograft was retained in only two of these patients, and the survival rate for all allografts was 72% at ten years7. The authors concluded that allografts used in revision knee replacement in patients with the difficult problem of massive bone loss have an encouraging medium-term rate of survival.
The advantages of structural allografts include their biologic ingrowth potential, versatility, relative cost-effectiveness, potential for bone stock restoration, and potential for ligamentous reattachment36. Because structural allografts do not revascularize, the major advantage of allografts, compared with cement filling or augments, is their ability to unite to damaged host bone that has a poor cancellous structure19. The disadvantages of structural allografts include the risk of disease transmission and nonunion, malunion, and collapse or resorption of the graft36. Another disadvantage is the meticulous preparation required to maximize surface contact between the allograft and the host-bone interfaces36.
The potential advantages of the porous tantalum metaphyseal cones may include the simplified surgical technique required for implantation, compared with preparation of a structural bone allograft, resulting perhaps in shorter operative times with a potential benefit of a decreased infection risk37. In addition, by avoiding a structural allograft, the risk of disease transmission is minimized. Furthermore, it is unlikely that these implants will fail by resorption or implant collapse. The potential for successful osseous ingrowth into these implants, even in the face of severely damaged bone, was supported by the results of the present study. One distinct disadvantage may be the extreme difficulty in removing these implants should that be required for circumstances such as recalcitrant deep periprosthetic infection. We are unable to comment on this potential difficulty, as extraction of the porous tantalum cones has not been required to date.
We currently prefer the use of cemented stemmed extensions to maximize early implant fixation and allow for successful biologic ingrowth of the porous tantalum cones into the remaining proximal part of the tibia. Whether the outcome of these porous tantalum cones when used with uncemented diaphyseal engaging stems will provide similar results has yet to be determined. As our experience has increased, we have been cementing shorter metaphyseal stems to allow for initial, stable implant fixation until successful osseous ingrowth has occurred into the porous cones. Whether the ability to obtain proximal tibial biologic fixation with these porous devices will translate into more durable fixation of these implants is as yet unknown.
In summary, early results with the use of porous tantalum metaphyseal cones to reconstruct large tibial defects in revision knee arthroplasty appear encouraging. However, further study is needed in a larger cohort of patients with longer follow-up to determine whether this novel reconstruction method will provide durable long-term outcomes and clinical success. 