Goals of the procedure include restoration of femoral bone stock and provision of an interface for mechanical interlocking of cement in cemented femoral revision.
Avoid distal fixation long stems that do not preserve or restore femoral bone stock, where the required length of scratch fit cannot be achieved, and where the canal diameter exceeds 18 mm (in which uncemented stems have a high incidence of thigh pain).
The procedure is applicable when fixation with a coned, fluted stem cannot be achieved because of the reverse cone shape of the natural femur below the isthmus.
Preoperative planning should include exclusion of infection, analysis of bone deficiencies, and templating the femoral component and canal plug.
Surgical technique should include the following: (1) positioning; (2) surgical approach; (3) explantation, debridement, and graft preparation, including removal of the femoral component, further femoral exposure, cement and membrane removal, preparation of the graft, and femoral preparation for impaction; (4) cementation; (5) stem insertion and closure; (6) strut grafts; and (7) postoperative management.
Ensure the distal plug is secure.
Preserve and use well-fixed distal cement if not contraindicated.
Choose the correct of size of phantom.
Impact the phantoms to a point at which hand removal is not possible.
Avoid calcar mesh until after trial reduction.
Apply calcar mesh with phantom in situ.
Tight packing should be done to the point of appearing like cortical bone.
Use large bone chips for greater torsional stability.
Leave the phantom in situ until the last possible moment.
Contraindications include complete loss of the proximal femur exceeding 10 cm, very elderly patients with poor bone stock recovery, and medically unfit patients in whom a relatively short revision operation is desirable.
Common errors include failure of containment by closure of cortical defects, failure to use supportive cerclage wires where indicated, overfilling the distal canal, failure to identify a calcar or diaphyseal fracture, and failure to build up the calcar and thereby reduce rotational stability.
Ensure that sufficient allograft is available.
Insertion of the stem too early with cement in a low-viscosity state may lead to reduced stem stability.
In 2005 approximately 7500 (15%) of all total hip arthroplasties carried out in the United Kingdom were revision procedures. The reasons for this are numerous and include a demographic shift toward an increasingly aged population and a trend for performing joint replacements in younger patients. Significant potential drawbacks to revision operations exist. They have higher complication rates, have longer mean operative times with an associated greater blood loss, and are a significant financial burden. When faced with the difficult situation of a revision procedure with significant femoral bone stock loss, the technique of femoral impaction grafting provides a method of reconstructing stable, supportive periprosthetic anatomy.
The material most frequently used in impaction grafting is allograft human femoral bone, harvested from the femoral heads of donors at the time of primary total hip replacement. Bone harvested from cadaveric femoral heads also is used; however, cadaveric bone is most frequently used in the production of freeze-dried bone. Because of the possibility of disease transmission with allograft materials, stringent guidelines regarding their use are in place. The use of a xenograft bone substitute has not met with success. Before long the situation may arise when the demand for allograft femoral heads will exceed supply; this fact has triggered renewed interest in cadaveric allograft and xenograft bone substitutes.
In the majority of units the donated femoral heads are not sterilized. The UK Blood Transfusion Services have produced strict guidelines on donor selection and microbiologic screening. Before donation patients are screened for communicable diseases, after which a quarantine period of 6 months is followed by further serologic screening. Fewer than 10 cases of human immunodeficiency virus (HIV) and hepatitis C virus transmission by unprocessed bone allograft have been described in the literature. The quarantine period reduces the risk because it may reveal a donor who was HIV positive at the time of donation but had not yet seroconverted. Currently, in unprocessed bone the risk of HIV transmission is one in 1 billion and hepatitis C is one in 2 million.
The formal sterilization of donor bone with gamma irradiation or ethylene oxide is used by some surgeons. This can be combined with a freeze drying process. Although gamma irradiation at a dose of 25 kGy has not been found to affect the mechanical properties of bone, higher doses (60 kGy) have been shown to significantly reduce the compressive failure stress and elastic modulus of human cancellous bone. However, when bone is freeze dried and then subjected to 25 kGy its ultimate strength is reduced by 42.5%. It also is recognized that lipids altered by the irradiation process may have a cytotoxic effect on osteoblasts and could therefore potentially interfere with bony ingrowth and graft survival.
Although not widely used in the United Kingdom, ethylene oxide gas sterilization is used in allograft bone preparation. Whether bone prepared in this way has a diminished propensity to incorporate is still unclear. What is known is that ethylene oxide is potentially carcinogenic and can cause an inflammatory reaction in the recipient.
The two principal methods of graft preservation are (1) fresh frozen with storage at −80° C and (2) freeze drying with subsequent storage at room temperature. The process of freeze drying bone alters its mechanical properties, making this form of graft brittle and less suitable for impaction grafting. Care needs to be taken when interpreting comparison studies between these two processing techniques because the preparation of fresh-frozen graft can have a significant effect on its mechanical properties.
It is well recognised that achieving high initial stability of impacted graft is of paramount importance and that early subsidence of greater than 10 mm is a clear marker of failure. The properties of impacted bone graft have been likened to those of an aggregate in soil mechanics, and as such the Mohr-Coulomb theory may apply. According to this theory of the strength of granular materials, the shear strength of a granular aggregate such as bone graft depends on the internal friction, expressed as the angle at which the aggregate will slide, and the interlocking of the particles, expressed as a stress:
τ = I + σ tan φ
where τ is the shear strength of the aggregate, I is the interlocking between particles, σ is the normal stress applied, and φ is the angle of internal friction. The resistance to shear can therefore been seen as proportional to the interdigitation between particles and the force directly compressing the aggregate.
The optimal size of allograft bone chips is a topic of much contention, and it probably differs depending on the site of impaction grafting. Some groups have shown that large chips (8 to 10 mm) have produced 25% greater stability than small chips (2 mm) in the acetabular model. In vitro studies examining the bone chip size on the femoral side have been done and have shown improved stability with chips greater than 4.5 mm in size.
A well-graded mixture of different-sized particles may produce an aggregate with a higher shear resistance than an ungraded mixture. By doing this, I (the interlocking between particles), is greatly increased, which in turn reduces the significance of the normal stress applied and the angle of internal friction.
The washing of bone chips is perceived to be an important step in graft preparation ( Fig. 35-1 ). The removal of fat reduces particle lubrication and therefore increases frictional resistance. This may allow improved compaction of the graft, which will allow greater interdigitation. In addition, by removing fat and marrow contents, porosity may be increased and osteoblastic inhibitory factors could be removed, both of which could potentially allow improved ingrowth. Washing also acts to reduce the immunogenic load and risk of disease transmission to the recipient. Some surgeons believe that washing the graft is unnecessary and by doing so the cohesive forces between particles actually may be reduced, having a deleterious effect on the graft’s mechanical properties.
The technique of impaction grafting has not been successful when used with uncemented technology. This is probably because of inadequate initial stability, which would then result in a fibrous layer being formed before bone has the opportunity to remodel.
To improve graft incorporation, some researchers have considered a variety of graft additives, including bone morphogenic protein-7 or osteogenic protein, bisphosphonates, and demineralized bone matrix. Although early ingrowth with these additives has been shown, so has an increased resorption, which may lead to reduced mechanical stability.
Because of the dependency on donor material, the use of graft expanders has been studied. A number of materials have been examined, including bioglass particles. The soil mechanics theory suggests that the addition of small particles would improve shear strength; this has been demonstrated in in vitro models for which a tri-calcium phosphate and hydroxyapatite mixture was added to human allograft. However, the biologic aspects of graft expanders are not yet fully understood, and a denser graft may reduce both cement penetration and later ingrowth, both of which could lead to early failure.
The initial stability of the bone, impaction graft, cement, and prosthesis construct is of paramount importance; this is what is responsible for short-term success. However, unless ingrowth occurs, forming a bond between graft and host, any initial success is certain to be short lived. This process has been shown to occur in histologic studies of human retrievals. The time taken for graft incorporation is unclear, although some authors have suggested about a period of approximately 1 year. The evidence for this comes from positron emission tomography studies that demonstrated the initial increase in blood flow and bone remodeling reduces to preoperative levels after this time. However, scintigraphic analysis has demonstrated this process takes up to 3 years.
Gie et al examined the proximal femur from a patient who had died 3.5 years after impaction femoral grafting and cemented revision. Two large cortical femoral defects in Gruen zones 2 and 4 of the femur had been covered with wire mesh stabilized with cerclage wires; impaction femoral grafting had then been performed. The macroscopic specimens showed clear healing of the lateral femoral cortex underlying the mesh, back to a normal thickness of cortical bone. The histology demonstrated three zones: regenerated cortical zone, interface zone, and deep layer zone. The layer closest to the implant contained necrotic bone entombed in cement. The intermediate zone showed direct contact between methylmethacrylate and osteoid, with scattered foreign body–type giant cells, an appearance similar to that described by Charnley in the femur after primary cemented hip arthroplasty. No direct contact between viable mineralized bone and cement was apparent. The outer zone contained histologically normal cortex and fatty bone marrow with a few islands of dead bone. More than 90% of the total surface area of the sections of new cortical bone contained filled osteocyte lacunae. No continuous fibrous membrane was present between cement and new bone.
Nelissen et al took biopsies from the medial femoral neck during surgery to remove painful cerclage wires in four patients 11 to 27 months after revision with impaction grafting. These cases show features similar to those reported by Ling et al. The inner zone contained cement and trabeculae of partially necrotic bone, fibrosis, and occasional lymphocytes, consistent with bone graft undergoing remodeling. Several areas of viable mineralized bone were found directly adjacent to bone cement. The middle zone contained occasional particles of bone cement as well as viable trabecular bone in all cases. The outer zone uniformly consisted of viable cortex. These studies demonstrate what can be achieved when the operation is done well.
Linder histologically examined six femora obtained postmortem and eight femoral biopsies 3 months to 8 years after impaction grafting with a cemented stem. He found a viable cortical shell around the grafted area in all cases. One patient showed complete bony restitution; the others had varying amounts of remaining graft in the neomedullary cavity. These particles were embedded in dense fibrous tissue, forming a composite capable of carrying load. No time-dependent histologic deterioration was found in any of the cases. Radiographically he found that cortical healing and trabecular remodeling corresponded to viable bone.
HISTORY OF IMPACTION GRAFTING AT EXETER
Impaction femoral grafting and cementing of the femoral component into the newly created medullary tube was introduced to clinical practice at the Princess Elizabeth Orthopaedic Hospital in Exeter, United Kingdom, in April 1987. The senior author (G.A.G.), who at the time was in a fellowship training position, was faced with a patient with a multiply revised hip in which proximal bone stock was poor and the endosteal surface of the femur had the appearance of an expanded smooth tube with no possibility of achieving any interdigitation of cement with host bone.
Being aware that Professor Robin Ling had previously impacted a femoral canal with milled bone in 1985 and placed a femoral component without cement that subsequently subsided significantly but where the bone graft had incorporated, and of Professor Slooff and colleagues’ experience with impaction bone grafting in the socket, the senior author decided to impact bone chips into the femur, creating a new medullary canal and cementing in a new component. The bone was packed distally on to a canal plug with femoral canal sizers. A trial stem two sizes larger than the definitive stem to be used was hammered firmly into the femur to form a new femoral canal. This stem was repeatedly removed, more bone graft added, and again hammered into the femur until a stable graft layer was achieved. The canal was then sucked dry of blood, low-viscosity cement was sucked down the femur, the cement was pressurized, and the definitive component was inserted. The newly created medullary tube ensured no contact between the cement and host bone. This is the first documented case of a femoral stem revision being performed in this way. The patient had an excellent clinical and radiological outcome until her death 5.5 years later from causes unrelated to her hip surgery.
The next 110 cases of impaction femoral grafting were performed by Exeter surgeons in the same way, but the method needed improvement to place the femoral component both in the central axis and in neutral alignment and to make the impaction process more consistent. In conjunction with Slooff’s team in Nijmegen, The Netherlands, impaction instruments were designed. These consisted of a threaded plug into which a guide wire could be inserted and fixed and cannulated self-centering distal impactors and stem-shaped proximal impactors (phantoms). Similar instruments were developed independently in the United States.
These instrumentation systems are based on the use of a collarless, double-tapered, polished stem that is cemented into a neomedullary canal formed by the tight impaction of morselized allograft in the proximal femur.
The outcomes presented in this chapter were obtained with the use of this type of stem and may not apply to the use of other femoral component designs used with impaction grafting because of different loading conditions found with stems of alternative designs.
Femoral impaction grafting is indicated for any patient in whom revision of the femoral stem is required and where, after removal of the femoral component, a smooth endosteal surface of the femur remains with little cancellous bone to provide adequate fixation for a cemented femoral component. The technique also is indicated in any revision situation for which restoring bone stock is desirable and also where the intramedullary canal diameter is 18 mm or greater because an uncemented stem in this situation has a high incidence of thigh pain. It is further indicated, even in the hands of surgeons dedicated to uncemented stems, when the required length of scratch fit cannot be achieved—such as in the setting of significant bone loss to below the isthmus of the femur.
The technique is applicable to patients of any age but is most useful in the younger patient for whom restoring bone stock is desirable.
No contraindications exist, although the authors advise that the procedure be performed in two stages in the presence of infection although some surgeons use this technique as a single-stage procedure in infected cases. Where complete loss of the proximal femur exceeds 10 cm, reconstruction with impaction grafting becomes extremely complex and other methods of stem fixation are recommended.
Although the technique works in patients of all ages, it may not be indicated in the very old or in medically unfit patients, for whom a relatively short revision operation is desirable and distal fixation is achievable with an uncemented stem ( Table 35-1 ).
|Indications||Restoration of bone stock in femoral revision hip arthroplasty Provision of an interface for mechanical interlocking of cement in cemented femoral revision when removal of the previous prosthesis has left a smooth endosteal surface To avoid the use, in young patients, of distal fixation long stems that do not preserve or restore femoral bone stock When the required length of scratch fit cannot be achieved with a fully coated cementless stem When the intramedullary canal diameter is greater than 18 mm, at which uncemented stems have a high incidence of thigh pain When fixation with a coned, fluted stem cannot be achieved because of the reverse cone shape of the natural femur below the isthmus|
|Contraindications||When complete loss of the proximal femur exceeds 10 cm Very elderly patients in whom bone stock recovery is not an issue Medically unfit patients for whom a relatively short revision operation is desirable|
|Relative contraindications||Age limit and life expectancy; physiologic age rather than chronologic age should be considered|
Exclusion of Infection
Screening patients for infection is carried out along conventional lines. When infection is suspected or antiinflammatory markers are raised, aspiration of the joint is carried out before the definitive revision procedure. Antibiotic powder is added to the graft at the subsequent impaction grafting procedure; the authors’ usual practice is to add 1 gm of vancomycin for each femoral head of allograft.
Analysis of Bone Deficiencies
Prerevision radiographs are analysed in detail. Anteroposterior pelvis, anteroposterior femur extending to well below the tip of the existing implant, and lateral radiographs are taken to detect endosteal and cortical femoral bone deficiencies. Donor allograft femoral heads or condyles and strut grafts, if necessary, are ordered from a bone bank. Femoral reconstruction metal meshes must be available to reconstruct the femoral tube when preoperative radiographs indicate cortical deficiencies exist or where part of the proximal femur is lost ( Fig. 35-2 ).
From the radiographs, the size, length, and offset of the stem required for revision is determined with the appropriate translucent templates or digital templating software ( Fig. 35-3 ).
The stem must bypass the most distal femoral defect, that is, a cortical defect or an endosteal lytic lesion involving 50% or more of the cortex seen on two views by at least one, and preferably two, cortical diameters. The femoral impaction grafting system of X-change revision instruments (Stryker Corp., Rutherford, NJ) allows implantation of all Exeter stems from 30 mm to 50 mm offset and from 125 mm to 260 mm in length ( Fig. 35-4 ).
The threaded femoral plug must be templated to lie at least 2 cm beyond the tip of the stem to be used at the revision operation ( Fig. 35-5 ). If significant limb shortening is present, the plug should be placed a little more distal in the femur in case trial reduction is too tight and the femoral component needs to be inserted deeper than expected.
If a well-fixed cement plug lies at least 2 cm distal to the most distal bony defect and the tip of the stem to be used, it can be left in situ and drilled; the threaded guidewire can then be inserted into the preexisting plug. Similarly, if a long column of well-fixed cement lies well distal to the lowest bony defect, it can be drilled and the guide wire can be secured into the cement.
Positioning of the Patient
The patient is securely positioned by the operating surgeon personally on his or her side in the lateral decubitus position. The lower limb is draped freely, giving exposure from the iliac crest to the knee ( Fig. 35-6 ).
The authors use a posterior approach exclusively for its versatility and extensibility and are not concerned about previous approaches to the hip. A posterolateral incision is made, incorporating or excising the previous scar if possible. Minimally invasive surgery has no place when bone quality is such that femoral impaction is required. Limited clearing of subcutaneous fat from the fascia lata is made in the line of the incision to facilitate closure. The fascial incision follows the line of the skin incision, with the gluteus maximus muscle being split in the line of its fibers.
Identification of the Sciatic Nerve
The sciatic nerve is identified by palpation and blunt dissection. Formal exposure of the sciatic nerve is unnecessary in every case, but the authors do advise dissecting out the nerve in the setting of previous acetabular fracture and when the posterior column of the pelvis is deficient and requires augmentation. When incising the external rotator muscles and hip capsule, electrocautery is used, giving due warning if the dissection approaches the sciatic nerve.
Aspiration of the Hip
The hip is aspirated before opening the joint capsule and the fluid obtained is sent for immediate microscopy and routine and enrichment culture. If microscopy reveals greater than 100 neutrophils per high powered field or if organisms are identified, the revision is abandoned, the hip carefully debrided, and a temporary hip spacer inserted. This involves the appropriate metal and polyethylene components with a total of 4 gm vancomycin and 1 gm gentamycin for each mix of cement. Frozen section of multiple tissue samples may be useful if any doubt remains about the possibility of the joint being infected, although this is not in routine use at the authors’ institution.
Exposure of the Joint
With the hip positioned in slight flexion, adduction, and internal rotation and the vastus lateralis muscle retracted anteriorly, the tendinous insertion of gluteus maximus is released from its femoral insertion. The gluteus minimus is lifted off the capsule and retracted anteriorly.
If a previous exposure by a posterior approach has been used, the external rotators and capsule are raised as a single layer, incising them with electrocautery close to the posterior aspect of the greater trochanter and obtaining as much length as possible to facilitate closure at the end of the procedure. This is best achieved by the limb being held in 45 degrees of flexion and internal rotation during capsular incision. If previous surgery involved an anterior approach, the external rotators and capsule are raised as separate flaps. A single straight longitudinal incision is made in the capsule to the upper border of the acetabulum with no posterior curve or T-shaped incision. This helps create a large, sturdy posterior flap for later repair, which helps reduce the risk of dislocation. Stay sutures are placed in the capsular flap and are used to reflect it posteriorly, protecting the sciatic nerve.
Dissection of the tissues off the bone is carried distally to the level of the lesser trochanter. Releasing the psoas tendon from the lesser trochanter and the anterior capsule from the femoral neck helps deliver the femur out of the wound, facilitating the exposure, especially in obese patients and muscular men. This release should be carried out before the leg is significantly flexed and internally rotated because the anterior wall of the femur often is flimsy as a result of osteolysis and may fracture or avulse during dislocation of the joint. These releases also help reduce the risk of femoral fracture during the impaction grafting process because they reduce the tensile and torsional forces to which the femur is subjected. The hip is then dislocated with as little force as possible, aided by a bone hook around the prosthetic neck and a gentle lift.
Explantation, Debridement, and Graft Preparation
Removal of the Femoral Component
Once the proximal femur is adequately exposed with the aid of a femoral elevating retractor, cement is removed from over the shoulder of the prosthesis with a high-speed burr. The femoral component is then extracted, with care being taken to ensure that impingement on the greater trochanter does not occur. An instrument is applied to the neck of the prosthesis to control rotational forces that otherwise may result in a fracture of the proximal femur during component removal.
If an uncemented component is being removed, a single longitudinal femoral split or an extended trochanteric osteotomy may be required. This does not preclude the technique of impaction grafting and indeed has been used in the authors’ center in more than 20 cases without any aseptic loosenings or trochanteric nonunions to date. The osteotomy must be soundly repaired with cables.
Further Femoral Exposure
The proximal part of the greater trochanter must be sufficiently exposed to allow insertion of the guide wire down the medullary canal in the midline axis so that the neomedullary canal that is subsequently formed is in neutral alignment, avoiding either varus or valgus. This often requires opening the trochanteric overhang laterally by approximately 1 cm to accommodate the introduction of instruments in the correct alignment without risking fracture of the trochanter.
Cement and Membrane Removal
Cement and membrane removal must be complete in the area for impaction grafting. However, if the distal cement plug is greater than 2 cm beyond the tip of the stem to be used, it may be left in position and used to occlude the distal canal during reconstruction.
Several separate specimens of tissue and membrane from the interfaces are routinely sent for microbiologic examination.
Preparation of the Graft
The authors use allograft from fresh-frozen femoral heads, sourced from their center’s bone bank, which complies with the standards and procedures laid down by the UK Tissue Banking Standards Authority. ABO compatibility between graft donor and recipient is not necessary and rhesus compatibility is only important when the patient is a rhesus-negative woman of childbearing age. Even the most straightforward cases usually require two femoral heads; in more complex cases several heads will be required.
Graft preparation is critical to the success of the procedure. All soft tissue and cartilage must be removed from the bone. Two sizes of bone chips are required, 3- to 4-mm chips for packing the distal three quarters of the canal above the plug and 8- to 10-mm chips for the proximal quarter. Note that neither very fine milled bone nor bone slurry is suitable for impaction grafting because they do not have the mechanical properties required for adequate impaction. Their use will lead to failure.
Femoral Preparation for Impaction
In essence, the technique of femoral impaction grafting restores the femur to a state equivalent to that at the time of primary arthroplasty. The first step, if required, is cortical tube restoration with mesh followed by cancellous restoration with impaction grafting. The success of impaction grafting depends on adequate physical constraint of the graft material. The surgeon should therefore have a low threshold for prophylactic cerclage wiring, and any defects in the femoral diaphysis must be repaired before impaction grafting.
Malleable stainless steel meshes (Stryker Corp.) are secured with monofilament cerclage wires or cables to contain any cortical defects or perforations. Periprosthetic fractures are addressed in a similar fashion. These meshes are placed with as little soft tissue dissection as possible. Cortical strut allografts may be required to augment the diaphysis in certain situations. Uncontained defects of the proximal femur are reconstructed later in the procedure.
Distal Occlusion of the Femur
Before grafting the medullary canal must be occluded distally to contain the graft. The canal size is determined with canal sounds ( Fig. 35-7 ).