Impaction Bone Grafting
Repair and reconstruction of a failed total knee arthroplasty (TKA) should begin with a plan that includes conservative treatment of the bone and soft tissues, restoration of bone stock with autologous bone reamings and morselized allograft, and selection of uncemented implants that will stabilize the knee and support bone reconstitution and soft tissue healing. Failed TKA almost always is accompanied by bone loss and pathologic changes to the ligaments and quadriceps mechanism, but this degeneration need not condemn a functional joint. This chapter presents the surgical technique for joint reconstruction, from exposure to closure, and offers a detailed rationale for each step of bone preparation, implant selection and fixation, and bone grafting technique. A clinical series also is presented.
The most conservative method of managing bone defects in revision TKA is to conserve bone stock and soft tissue attachments, to fix the implants rigidly to bone using a diaphyseal-engaging stem and rim contact with the articular implant, and to fill the resulting defects with cancellous bone.
Exposure can damage soft tissue attachments, so tibial tubercle osteotomy often is the most conservative method to achieve effective exposure of the knee and adjacent structures.
Minimal resection of bone at the ends of the femur and tibia allows the implants to “close the gap” between the bone surfaces and allows moderately sized implants to achieve stability.
Stem fixation with a coarsely grit-blasted or porous-coated stem fixed rigidly in the medullary canal of the metaphyseal-diaphyseal junction imparts rigidity of fixation to the articular implant that has been abutted to the bone rim.
Local bone (resected fragments and medullary reamings) is used to fill metaphyseal defects and is important for healing and ossification of cancellous allograft.
Expose with a long incision but with minimal soft tissue stripping. Use a tibial tubercle osteotomy for exposure if necessary, so that the metaphyseal bone that supports the implants and sustains the morselized graft will remain strong and viable.
Clean the surfaces of the defects down to fresh, bleeding bone, but do not remove any bone from the floor of the defects.
Use the hard bone of the metaphyseal rim for axial support, not the central medullary bone of the defect. This is easier and conserves bone more effectively.
Pack the cancellous graft material loosely so that it will vascularize and ossify quickly.
Cement is the main pitfall in revision TKA, just as in revision total hip surgery. Do not cement either the stem or the metaphyseal components.
Avoid using major structural allograft. When major structure is needed, use porous-coated metal augments rather than the more expensive and fragile allograft material.
Do not pack the cancellous graft to use as a primary weight-bearing structure. Instead, rest the implant on the metaphyseal rim and stabilize it with the stem. Pack graft loosely into the resulting defects.
Failure of total knee arthroplasty (TKA) almost always is accompanied by bone loss and pathologic changes in the ligaments and quadriceps mechanism. Revision TKA should be designed to reverse this deterioration by reconstructing lost bone stock and managing the bone structures such that the ligament attachments remain intact and the bone structures remain viable to assist with healing and regeneration of bone stock. Even after multiple revision TKA procedures and use of constrained hinges, sufficient bone and soft tissue structures remain to allow stable reconstruction. Techniques that preserve and utilize these intact structures offer the potential to reverse deterioration; even if failure occurs, they improve rather than destroy bone stock and soft tissue structures about the knee and thus avoid destabilizing the knee and compromising the next procedure that may be necessary.
This concept of conserving and reconstructing bone support contrasts with many traditional revision techniques in which the prosthesis is cemented into remaining bone stock—a technique that has failure rates of 10% to 15% after 5 years. Fixing the prosthesis to remaining bone stock with the use of osteointegration techniques, bone graft material, and augments, along with tensioning and adjusting the soft tissue structures, establishes a stable and durable knee while creating an environment that encourages bone regeneration. These principles can replace high-risk and potentially destructive procedures such as use of massive allografts, cemented stems with pressure-injected cement, and fixed hinges that cause additional bone loss and damage to the remaining soft tissue support of the knee.
Reconstruction of bone stock depends on secure fixation of the prosthesis in the depleted metaphyseal bone stock and remaining intact diaphyseal cortical bone. Several techniques have been described to achieve secure fixation of the implants. Highly porous metal cones currently are popular but require sacrifice of metaphyseal bone to achieve fixation of the cone-shaped structure into the existing metaphyseal architecture. They also rely on cement fixation past the metaphyseal cone and therefore risk all of the bone of the metaphysis and some of the remaining diaphyseal bone stock as well.
A technique that allows the porous augment to seat directly against the existing metaphyseal rim and controls tilting and micromotion offers a technically easy alternative that conserves bone. This “rim and stem” technique does not require contouring and sculpting of the metaphyseal bone stock and compresses the bone stock rather than requiring resection and reaming of the metaphysis. It has the added advantage of preserving the intraosseous blood supply that is so important for metaphyseal bone stock viability and regeneration. Various porous-coated augments are available to position the joint surface and correctly balance the knee in flexion and extension ( Fig. 22A.1 ). Whereas these buildups can provide excellent support for the component, toggle control of the entire implant is gained by the use of a noncemented stem with sharp flutes or a roughened surface to engage the diaphyseal isthmus. This simple but effective rim and stem technique offers a mechanism for anchoring the prosthesis even to severely damaged cancellous and cortical bone about the knee.
Reports of reconstruction of major bone deficiencies using morselized allograft combined with rim and stem fixation of the prosthesis have been encouraging. Durability of this construct has been excellent, and repeat revision as a result of failed fixation is rare. Use of a nonporous stem in the diaphyseal bone stock depends on reliable bone stock contact medially, laterally, and posteriorly on the femoral component and at least medially and laterally on the tibial component. This rim and stem method of fixation has been shown in biomechanical testing to be highly effective in dealing with major bone deficiency. A stem that engages the diaphysis and a metaphyseal component that rests on a portion of the metaphyseal rim provide excellent resistance to migration. Although this type of fixation usually can be achieved with effective use of metaphyseal augments, occasionally adequate contact of porous metal surface against viable bone cannot be achieved. In these cases, a porous-coated stem is needed to achieve long-term fixation of the implant into reliable bone stock. Whereas fully porous-coated stems are commonplace in revision total hip arthroplasty, their use is not yet extensively reported in revision TKA.
Soft tissue balancing and reconstruction are closely related to the problems of prosthesis fixation and bone reconstruction. Maintaining the soft tissue attachments to the remaining damaged bone structure is a crucial feature for successful revision TKA. When unconstrained implants are used, correct alignment is the first step and is the basis for achieving a balanced, stable knee after TKA. It also is a key factor in minimal resection and fixation of the implants to existing bone stock. Finding reliable landmarks for positioning and alignment of the implants is challenging in revision arthroplasty. In most cases, the medullary canal requires reaming, and as the reamer becomes fixed firmly in the medullary canal of the femur and tibia, it is the most reliable instrument for aligning the femoral and tibial components ( Figs. 22A.2 and 22A.3 ). With basic alignment established and bone remnants (along with their soft tissue attachments) meticulously preserved, the knee is ready to be reconstructed and balanced.
Rationale for Morselized Allograft in Bone Defects
Cancellous and cortical bone loss after failure of TKA may be restored and the bone stock improved by block allograft, but the morselized grafting technique obviates the more expensive, prolonged, and complicated preparation of a solid allograft. When compacted, morselized allograft can have considerable structural integrity, but it is most readily vascularized and ossified if it is pushed loosely into the defective area and allowed early protected load bearing. Use of morselized allograft or demineralized bone material was developed in conjunction with diaphyseal stems and metaphyseal augments that allow rigid fixation and adjustment of joint surface position and also allow the graft material to be loaded safely soon after surgery.
The ready availability and clinical success of morselized allograft has led to its increasing use for reconstruction of bone defects. Over the years, a considerable body of knowledge has been produced to define how this material functions. Granulated allograft bone comprising pieces smaller than 0.5 cm often is resorbed and removed by the inflammatory process. Bone pieces that are between 0.5 and 1 cm in diameter are the most effective because they resist resorption and allow rapid vascularization. Larger segments are much more difficult to construct and fit to the defect, and they are slower to ossify and incorporate. Rapid healing and ossification occurs routinely when morselized bone grafting is done with appropriate technique, provided that the implants are stable and the allograft is surrounded by viable bone ( Figs. 22A.4 to 22A.7 ). Although morselized cancellous allograft is osteoconductive rather than osteoinductive, it serves as an effective scaffolding for new bone formation. Demineralized bone added to the morselized allograft provides the osteoinductive stimulus and probably augments healing of failed massive implants in which large defects are encountered. The technique used with morselized allograft is important to the rate and completeness of healing in large defects. Loose (rather than tight) packing improves vascularization and hastens bone formation. Addition of osteoinductive proteins encourages bone formation deep in the graft. Implantation that encourages early load bearing improves bone formation and maturation.
Bone formation appears to begin early, and it progresses slowly through the first 18 to 24 months. Findings from biopsy specimens taken from patients undergoing this treatment program suggest that the graft is fully mature by 3 years after surgery. Most of the bone in the mature grafted areas is a combination of entombed allograft trabeculae and new lamellar bone. This suggests that bone graft healing is similar to the mechanism of fracture callus formation and maturation. When carefully planned and executed, this technique of stabilizing the implants and grafting the defects produces reliable bone to support the implant and to be available in the event that another revision operation is required.
Exposure of the knee plays an important role in maintaining soft tissue attachments to the deficient bone of the femur and tibia and avoiding damage of the blood supply to the remaining bone. This is accomplished by avoiding exposure procedures that damage ligamentous attachments to the femoral and tibial metaphysis. If exposure is difficult, the surgeon should not do epicondylar osteotomy, soft tissue stripping from the tibial metaphysis, or transection of the quadriceps mechanism. Instead, patients whose knees cannot be exposed with minimal soft tissue elevation should have tibial tubercle osteotomy combined with splitting of the quadriceps between the vastus intermedius and vastus medialis to achieve the broadest exposure with the least damage to the muscular and ligamentous structures of the knee. This procedure is necessary in 10% to 25% of revisions. In the process of exposing the knee, soft tissue stripping should be minimal. The broad attachment of the medial quadriceps retinaculum and capsular ligaments of the medial tibial flare are left intact, and the epicondylar areas of the femur are left undisturbed so that the soft tissue sleeve around the knee can be tensioned adequately with the spacer effect of the implants. This leaves the bone structures viable and provides the support for regrowth of bone stock and ultimate support of the implants.
Once the knee is completely exposed, the failed prostheses are removed using techniques that minimize stripping of soft tissues. The cement and soft tissue membranes are removed down to viable cortical bone. In most patients with repeatedly failed TKA, no cancellous bone remains in the distal femur and upper tibia; however, the cortical rim of the femur at the epicondylar level almost always is present (except in cases of revision of tumor prostheses), and there often is dense cortical-cancellous bone in the central portion of the remaining metaphysis. This bone can serve as a reliable and durable support for the femoral component if the revision system is designed to achieve rigid toggle control with the stem in the medullary canal and augmentation modules on the metaphyseal bone surface to provide rim support distally and posteriorly. The tibial rim often is deficient except for a small remnant, but this remaining rim and the remains of the fibular head can provide adequate axial support of the tibial tray. Careful preservation of metaphyseal bone stock and its blood supply is the key to successful bone reconstruction and the use of morselized allografting technique in revision TKA. It not only will improve bone stock but also will allow reconstruction of the knee without resorting to constrained hinges, because it leaves most ligament attachments intact.
The first step in preparing the bone for fixation is to ream the medullary canals of the femur and tibia. As the reaming is performed, each reamer is cleaned carefully; the bone shavings are removed and preserved and will be used to add osteoinductive stimulus to the morselized cancellous allograft filling the cavitary defects in the bone. Each canal should be reamed carefully with reamers of increasing size until a tight fit is achieved in the diaphysis of each bone. Reaming to a depth of 200 mm usually is necessary to achieve correct alignment. A convenient method to align the sawcuts for the tibial tray is to use the reamer itself as the alignment instrument. Once it is firmly fixed in the medullary canal, the cutting guide is applied over the shaft of the reamer, and the bone surfaces are trimmed, removing minimal amounts of the rim of the femur and tibia (see Figs. 22A.2 and 22A.3 ). This technique of maintaining as much metaphyseal rim as possible ensures maximum preservation of bone stock and soft tissue attachments. Once alignment has been achieved with deep reaming into the isthmus, this channel can be used to accommodate a shorter stem by reaming to greater diameter in a stepwise fashion, following the same track that was made by the deeper reaming process but reaming less deeply.
As the femur is reamed, care is taken to avoid penetration of the anterior cortex. Often the bone is soft, and holding the reamer posteriorly to position the final component relative to the posterior femoral condyles and the anterior femoral cortex can damage the anterior femoral cortex. Therefore, the reamer should be allowed to follow the track of the femur to a depth of 200 mm. This depth can then be used to allow the cutting guide to be applied and the surface of the femur resected at approximately a 5-degree valgus angle. The track in the distal femur can be enlarged at 1-mm increments to prepare for a shorter stem and avoid anterior perforation. In some cases, a small reamer (e.g., size 12 mm or 14 mm) fits correctly in the diaphyseal cortex of the femur but would cause perforation of the anterior cortex if the reamer were tilted posteriorly enough to place the femoral component in its correct anterior-posterior position. In these cases, the femur can be prepared with a larger reamer to accommodate a 100- or 150-mm stem placed at an angle that positions the femoral component correctly with the distal femoral condylar structures; use of the shorter stem avoids penetration of the anterior cortex of the femur. Curved stems are available to avoid anterior femoral penetration, but their curvature dictates the rotational position of the final component and may lead to internal or external rotational malposition of the femoral component, resulting in ligament balancing and patellar tracking problems. A module that displaces the femoral component posteriorly allows the surgeon to use a straight stem to take advantage of its universal rotational alignment characteristics while placing the femoral component posteriorly into its correct anterior-posterior position (see Fig. 22A.1 ).
Component Sizing and Positioning
Sizing and positioning of the femoral and tibial components are the next steps in reconstruction. It should be reemphasized that the rims of the remaining bone stock must be resected sparingly with minimal soft tissue stripped from the bone during exposure. It is especially important to maintain the attachment of ligaments, even if compromised, to the epicondylar areas of the femur. This procedure leaves a soft tissue envelope that is attached securely to bone stock, and the surgeon can tension this envelope by choosing the sizes of prostheses and augmentation modules that impart stability to the knee in flexion and extension.
The correct size of the femoral component can be estimated by choosing a trial component that covers the medial and lateral dimensions of the remaining femoral bone stock. In most patients, some posterior bone stock remains in the femoral condyles, and minimal posterior resection allows the posterior surfaces of the femoral component to engage the posterior femoral condylar bone surfaces. This is a crucial stage of the procedure. The surgeon must ensure that the final implant will seat firmly on distal and posterior femoral bone stock or that the femoral stem will fit tightly in the diaphysis. If the stem is porous-coated or has a roughened surface such that it can osteointegrate with the diaphyseal canal, then posterior bone support is not necessary for long-term rotational stability. Femoral augments are available to ensure that posterior bone contact is achieved when needed.
The trial femoral implant then is attached to the trial modular stem and the implant is inserted into the femur. The need for a posterior offset module is assessed at this time, and the femoral trial component is placed so that the anterior surface of the implant is flush with the anterior cortex of the femur. Rotational positioning of the femoral component should be guided by the epicondylar axis ( Fig. 22A.8 ). The epicondyles are found and marked, an epicondylar line is drawn, and this line is validated by placing the patella anteriorly to be sure that it lies in the midline between the two bone landmarks that have been chosen to represent the epicondyles. Trial fitting of the femoral component begins with the largest distal femoral modular augmentation module if distal femoral bone loss has been severe, or with thinner distal modules if bone loss is mild or moderate. This technique allows the most efficient stepwise establishment of ligament balance in both flexion and extension. Posterior buildups are assessed as well.
To ensure minimal resection of posterior femoral bone stock, the trial is first inserted without augments, and its engagement with the posterior femoral surface is assessed. If no posterior bone stock contact is achieved with the nonaugmented femoral component, then augments are applied until adequate posterior surface contact is achieved medially and laterally. This often results in asymmetric sizes of the medial and lateral posterior augments. The trial device can be used to mark the posterior femoral bone for resection, and then a saw is used to finish the minimal posterior femoral rim resection. In many cases, the trials themselves can be used to make this resection, being driven down to compact and prepare these posterior surfaces while rotational alignment is firmly held during the impaction process. The combination of stem and rim fixation of the femoral component almost always results in excellent, rigid fixation of the trial component in correct rotational position and acceptable distal and posterior position relative to the ligaments and remaining patella and quadriceps structures (see Fig. 22A.1 ). This usually leaves large, partially contained defects that can be filled with morselized graft and demineralized bone paste.
The tibial trial component is inserted so that the stem fits snugly but not tightly in the diaphyseal medullary canal and the tibial plate abuts against the remaining tibial rim. Often the tibial rim is so deficient that the remaining portion of the fibular head must be used as a portion of the tibial support.
Prosthesis Insertion and Fixation
Fixation of the stem in the diaphysis is important in reconstruction of deficient bone stock. Stems with sharp flutes and slots to provide immediate stability have been used successfully for decades. Porous-coated stems or titanium stems with coarse grit-blasting can provide similar immediate scratch-fit fixation and have the additional advantage of long-term direct fixation of the stem to diaphyseal cortical bone. Whereas this technique is safe and reliable, it must be done with great caution and care to avoid diaphyseal fracture of the femur and tibia. One-half millimeter undersized reaming usually prepares the diaphysis for its stem, but one cannot be completely assured of tight fit without fracture unless the stem fit is carefully tested. Before final insertion, the implant should be inserted and tapped gently with the soft surface of the hammer. It should approach 1 to 1.5 cm short of full seating before becoming tight in the diaphyseal canal. With the diaphyseal stem firmly in the medullary canal, varus-valgus toggle should be eliminated, and the stem should offer modest resistance to being rotated around its long axis by friction against the surfaces of the bone. Once this type of snug fit has been achieved, the implant can be driven finally into place, seating on the remaining rim of the metaphysis of the femur and tibia. This is done under careful control of positioning, ensuring that rotational alignment is correct as the implant is seated. Every effort should be made to perform the final seating in one effort and to avoid extracting the implant for repositioning. Slotted stems are somewhat safer and easier to insert and achieve rigid fixation, but the slot in the diaphyseal stem significantly weakens the structure. If the implant cannot be readily inserted to the appropriate distance for final press-fitting, then very gentle, stepwise, careful reaming more deeply to near line-to-line fit can allow the final safe seating of the implant. Attempting to press-fit over a longitudinal distance greater than 2 cm invites catastrophic fracture or difficult extraction of the stem for repositioning.
Augments on the metaphyseal portion of the implants should be designed so that excessive bone resection is not necessary to fit the implants to the remaining bone stock. These may be porous-coated block augments that seat on the metaphyseal rim or cone-shaped devices that wedge into the deficient metaphyseal bone. With block augments that seat on the metaphyseal rim, reaming and coring out of large amounts of metaphyseal bone are not necessary. The porous surfaces are driven firmly against the remaining metaphyseal rim, and then the cavitary defects are grafted. Seating of the prosthesis on the remaining rim and use of a stem for stabilization create a stable construct around large cavitary defects in both femur and tibia.
Bone Grafting Procedure
Defects should be filled with non–load-bearing bone graft material that is both osteoinductive and osteoconductive. Available remnants of the patient’s bone should be used, and allograft tissue should be added to augment the native bone. Demineralized bone paste is mildly osteoinductive, is minimally immunogenic, and does not provide a nidus for bacterial growth. The reamings from the patient’s femur and tibia and any other fragments left over from the bone preparation are the first choice for filling cavitary defects in the bone. Between 40 and 50 cc of bone often is available from these local sources. If this is not sufficient, morselized allograft and demineralized allograft bone are the next best bone graft materials. Fresh-frozen allograft with morsels measuring 0.5 to 1.0 cm in diameter is soaked for 5 to 10 minutes in normal saline with the addition of polymixin 500,000 U, bacitracin 50,000 U, and vancomycin 1 g/L. The bone morsels are blotted to remove excess fluid, and 10 mL of powdered demineralized cancellous bone or demineralized bone paste is added to each 30 mL of frozen cancellous allograft. The bone defects are packed loosely with the mixture, and the implants are impacted to seat on the remnant of viable bone while the morselized bone graft is gently compacted ( Figs. 22A.9 and 22A.10 ). The morselized bone is not a weight-bearing structure and is meant to encourage bone regeneration in deficient metaphyseal bone. Actual weight bearing in the early postoperative period is accomplished through direct contact with the metaphyseal cortical rim and in the long stem through load sharing between the porous stem and the metaphyseal bone contact.