Increased Resection or Component Shift
Increasing the bone resection from the proximal part of the tibia or the distal end of the femur theoretically is a simple strategy to remove the bone deficiency. However, one must realize that bone was already removed at the time of initial knee replacement, additional bone loss occurred because of the failure of the primary total knee replacement, and further bone is typically lost during component removal. Aggressive bone resection further decreases the strength of supporting cancellous bone and may require a decrease in the size of the tibial component selected. Harada et al. found an abrupt decrease in tibial bone strength over the first 5 mm of resection15. Smaller tibial components lead to a decreased surface area for fixation and a corresponding increase in the unit load across the tibial tray and fixation interface. For these reasons, removal of substantial bone in revision total knee arthroplasty is to be avoided. We recommend that no more than 1 to 2 mm from the most prominent femoral or tibial condyle be removed and then any remaining deficits of the contralateral condyle are managed with use of other methods discussed below.
Shifting the tibial component in the coronal plane away from an osseous defect to an area of greater osseous support is another option of dealing with small (<5 mm) defects. Component shift can be detrimental if it leads to a smaller tibial component. The risk of subsidence increases if the tibial tray is supported by more cancellous bone. Shifting the tray also impacts ligament kinetics. For example, shifting the tibial tray medially to avoid a lateral tibial defect lateralizes the tibial tubercle and increases the risk of patellofemoral instability. Shifting the tibial tray should be limited to no more than 3 mm16. We reserve shifting the tibial tray to the lateral direction and limit this shift to no more than 2 mm. We do not recommend decreasing the size of the tray to accommodate this shift. Because of the risk of patellar instability and collateral ligament irritation, femoral component shift is rarely indicated. Downsizing the femoral component creates a larger flexion gap and can elevate the joint line.
Cement and Screw Reconstruction
Small defects in the knees of elderly patients can be filled with polymethylmethacrylate (PMMA) and screws13,14. The defect construct is strengthened by the addition of screws to the PMMA. Filling a bone deficiency with PMMA is indicated for a peripheral deficiency of ≤10% of the condylar area, or for small central defects5. Filling with PMMA is simple, economical, and easily contoured to the osseous defect. Large masses of PMMA, however, can lead to the potential for osseous thermal necrosis secondary to the heat of polymerization. It can also be difficult to pressurize PMMA in the setting of sclerotic uncontained defects. As PMMA cures, it can lose approximately 2% of its volume, leading to decreased support17. Although PMMA has inferior load transfer compared with custom implants or metal augments, use of PMMA has led to favorable clinical results, at least in primary total knee arthroplasty17. Thirteen (28%) of forty-seven total knee replacements treated with PMMA and screws and observed for an average of 6.1 years had radiolucent lines, but the lines did not progress and there was no prosthetic loosening13. These same authors evaluated 125 total knee replacements in which PMMA and screws were used to fill large medial tibial defects14. In this group, there were two failures with varus collapse at a mean of 7.9 years but no loosening in the remainder of the patients. The authors consider the use of PMMA with screws in Type-I or II defects involving <50% of condylar width and <10 mm in depth.
Localized Autograft or Particulate Allograft
Particulate allograft or local autograft can easily be molded to fit cystic and small contained areas of bone deficiency9-12. Bone-grafting is an attractive option for younger patients who may need further revision because of its potential to restore bone stock12. Bone grafts are cost-effective compared with metal augments, are useful in large defects, and have increased physiologic load transfer compared with PMMA18. Donor site options for autograft include the resected condyles, intercondylar notch, and iliac crest. There are numerous sources for allograft such as the femoral head, distal end of the femur, or proximal part of the tibia19. Allograft material can be fresh-frozen, frozen with radiation, or freeze-dried. There is a risk of nonunion, malunion, or late collapse with the use of bone graft. Allograft also has a minimal risk of disease transmission20. Whiteside reported on the use of morselized femoral head allograft in fifty-six cementless revision total knee arthroplasties over two years21. He observed new osseous trabeculation in fifteen knees with femoral grafting and in twenty-one knees with tibial grafting. Forty-seven (84%) of the fifty-six knees had mild to no pain, and nine (16%) had moderate to severe pain. Lotke et al., in a prospective study of forty-eight patients treated with impaction allograft for substantial bone loss, reported that all radiographs showed incorporation and remodeling of bone graft with no mechanical failures11. Six complications, including two infections and two periprosthetic fractures, were noted. Morselized cancellous bone grafts are best used for contained Type-I or II defects, in which the graft can be captured and, when compacted, provides some degree of structural support. Extensive Type-II contained defects can be managed with impaction bone graft as long as stem extensions are used7 (Figs. 1-A, 1-B, and 1-C). The addition of wire mesh can also be used to create a contained defect when the peripheral rim is not intact11.
Prosthetic Augments
Prosthetic augments are most commonly selected for noncontained unicondylar (Type-IIA) or bicondylar (Type-IIB) defects of moderate size. Tibial augments are available in various shapes including hemiplateau rectangular blocks, hemiplateau angular wedges, or a full plateau angular wedge22-26. Chen and Krackow demonstrated that alterations in defect configuration can affect the rigidity of the construct27. They observed that block augments were more stable than wedge augments because of a reduction in shear forces. Fehring et al. also found that block augments compared with angular wedge augments resulted in superior strain distribution to the supporting bone28. Bilateral block tibial augmentations can restore the anatomic joint line in Type-IIB defects and avoid the use of excessively thick modular tibial polyethylene inserts. Peripheral defects between 5 and 15 mm in depth that extend over the majority of the width of the medial or lateral tibial condyle are ideally managed with tibial augments. Femoral prosthetic augments range from approximately 5 to 15 mm in thickness, are typically block shaped, and can be used for both distal and posterior femoral bone defects. Unlike bone allografts, prosthetic augments have no risk of disease transmission, malunion, nonunion, or augment collapse. They have demonstrated good load transmission to underlying bone and provide immediate support and stability17. However, prosthetic augments are expensive and are limited in size and shape. There is a potential for debris creation from their modular attachment to the main component, and they can loosen if the bone supporting the augment is poor. Use of stem extensions can protect weakened condylar bone and enhance fixation. Augments are often combined with other treatment modalities, such as bone-grafting and stem extensions, so the results of the use of augmentation in revision total knee arthroplasty are somewhat difficult to precisely interpret. Patel et al. reported on seventy-nine revision total knee replacements in knees with AORI Type-II defects that were followed for a mean of seven years22. One hundred and seventy-six augments were implanted into the femur and tibia. Nonprogressive radiolucent lines were observed in 14% of the knees and were unassociated with clinical results. The survival rate of the implants at eleven years was 92%.
Structural Allografts
For noncontained Type-II defects that are too large to be managed with prosthetic augments and for Type-III deficits, structural allografts may be considered (Figs. 2-A, 2-B, and 2-C). Dorr et al. suggested that tibial defects involving >50% of the osseous support of either tibial plateau would benefit from allograft reconstruction12. Preoperatively, it is critical to match the specimen with the host defect size. Femoral head, distal femoral, or proximal tibial allografts can be used. The technical keys of bone-grafting include developing a healthy, bleeding host recipient site; maximizing allograft-host and prosthesis-host contact; optimizing mechanical interlock between graft and host; restoring the anatomic joint line; and supplying rigid implant fixation with no instability or malalignment. The use of additional stem extensions to enhance implant fixation and offload the structural allograft during incorporation is recommended.
Advantages of structural allograft reconstruction include the biologic potential for bone stock restoration and versatility, as the graft can be shaped to fit the host defect. The structural allografts can be used to restore the joint line and have the potential for ligament reattachment. Disadvantages include a minimal risk of disease transmission (<1 per 1,000,000 risk of human immunodeficiency virus transmission),20 and risks of allograft nonunion, malunion, collapse, or resorption.
While complications are frequent in complex cases requiring allograft reconstruction, multiple studies have demonstrated high union rates if rigid fixation is achieved29-31. Clatworthy et al. reported on revision total knee arthroplasty in fifty patients (fifty-two knees) with large, noncontained osseous defects treated with structural allografts29. Thirty-seven of the fifty-two knee revisions were classified as a success, with forty-six of fifty grafts surviving at five years and thirty-six at ten years. They reported eleven failures because of infection (four), graft resorption (five), and allograft nonunion (two). Engh and Ammeen evaluated the cases of forty-six patients who had revision total knee arthroplasty with use of structural allografts30. At a mean duration of follow-up of ninety-five months, they noted only four failed allografts, two of which were secondary to infection. No allograft collapse was observed. Bauman et al. reviewed seventy revision total knee replacements with structural allograft repair that were followed for a minimum of five years31. Eight failures related to the allograft were found, and they observed a revision-free survival rate of 75.9% (95% confidence interval, 65.6 to 87.8) at ten years.
Metaphyseal Sleeves or Cones
Contained cavitary and combined cavitary-segmental metaphyseal defects in the femur and tibia can be filled with metaphyseal sleeves or porous cones. These allografts provide metaphyseal implant support and fixation and are ideal for managing large central, cone-shaped deficiencies in the femoral or tibial metaphysis. Additional fixation is obtained on condylar bone or via diaphyseal-engaging stems. Unlike bulk allograft reconstructions, the metaphyseal sleeves and cones avoid the risk of nonunion and resorption. Porous metal cones achieve peripheral osseous ingrowth, and any prosthetic device can be cemented into their inner, central surface. Long and Scuderi described the cases of sixteen patients who were treated with tibial tantalum cones and followed for thirty-one months32. There were no revisions for aseptic loosening, and all patients showed evidence of osseointegration on follow-up radiographs. Meneghini et al. reviewed a series of fifteen patients who had revision knee arthroplasties with implantation of porous metal metaphyseal tibial cones and were followed for a mean of two years25. On radiographs, all of the tibial cones demonstrated evidence of osseointegration, and there were no reported failures.
Metaphyseal sleeves are designed for cementless fixation via a porous ingrowth surface and are implant-specific (Fig. 3). Progressive metaphyseal loading is achieved through a step design. The metaphyseal bone is prepared similar to broaching for a cementless femoral component in total hip arthroplasty. Progressively larger broaches are implanted until rigid fixation of the broach is obtained. This creates a precisely engineered cavity for later sleeve insertion (Fig. 3). To increase the rigidity of fixation, the authors recommend attaching a diaphyseal-engaging stem extension to the metaphyseal sleeve (Figs. 4-A and 4-B). Potential disadvantages of sleeve and cone use include the expense and the potential of removal difficulties33. Clinical results with metaphyseal sleeve use are currently short term but favorable. Jones et al., in a study with a mean follow-up of forty-nine months, described a combined series of thirty knee revisions, in which press-fit diaphyseal-engaging stems and metaphyseal sleeves were used34. All of the implants showed bone apposition and positive remodeling of bone adjacent to the metaphyseal sleeves on follow-up radiographs with no mechanical failures.
Condyle-Replacing Hinged Prosthesis
In massive Type-III defects with loss of collateral ligamentous support, use of a condyle-replacing hinged prosthesis should be considered, especially in low-demand, elderly subjects. The operative procedure is relatively simple and efficient and allows rapid rehabilitation and weight-bearing (Figs. 5-A and 5-B). The most prominent disadvantage of this treatment option is that there are few remaining reconstructive options if this method should fail.
Numerous treatment options are available to manage bone loss associated with revision total knee arthroplasty. Selection of the treatment method is based on many factors including the defect size and location in addition to the age, health, and ability of the patient to participate in the necessary postoperative rehabilitation.