Gavin Pittman
Thomas K. Fehring
General Principles
Total knee arthroplasty (TKA) has become one of the most successful orthopedic procedures with at least 95% good or excellent results reported after 10 years of follow-up. Despite this success, a painless, well-functioning prosthetic knee joint cannot be uniformly guaranteed. Approximately 32,000 revision total knee replacements were performed in 2002 in the United States alone, and that number continues to rise annually.
Although there are various reasons for TKA revision surgery, the most prevalent causes of failure can be grouped into early and late mechanisms of failure based on the length of time from the index operation. Early failures most commonly result from postoperative infections, joint instability, aseptic component loosening, and patellofemoral issues. Some of these failures are attributed to poor surgical judgment or surgical technique during the initial arthroplasty. Common causes of late failures include infections from a hematogenous source or the sequelae of polyethylene wear. Late failures appear to be more closely related to host factors and material limitations.
As discussed in previous chapters regarding the assessment of the painful TKA, a systematic approach using the patient history, physical exam, laboratory analysis, and radiographic images is required to determine the mechanism of failure. Surgical exploration of a painful total knee without clearly defined treatable mechanism of failure is rarely successful. A firm understanding of the underlying pathology and methods of addressing it is mandatory before intervening. The purpose of this chapter is to help the surgeon establish a sound preoperative plan and to execute this plan in an efficient manner to provide the patient with a durable, successful revision total knee replacement.
Preoperative Planning
Accomplishing a successful revision total knee arthroplasty can be a challenging task even for the most seasoned surgeon. Despite the added difficulties of addressing infection, malalignment, bone loss, and instability, the primary goals of revision surgery should remain the same as those in primary knee arthroplasty. These goals include establishing proper alignment, balancing the extension space in the coronal plane, and creating a flexion space that equals the extension space. Preoperative planning allows the surgeon to visualize these goals prior to entering the operating room and provides the opportunity to determine what additional tools and devices will be necessary to facilitate the revision.
Although nuclear studies, MRI, and CT scanning may be beneficial during the assessment of the painful total knee, conventional radiographs continue to be the most useful references for planning a revision. Weight-bearing anteroposterior (AP), lateral, and Merchant views offer an informative and quick method for evaluating the prosthetic components and the surrounding bone. Previously obtained radiographs may be available for comparison to assess progressive changes in bone quality or component stability. Oblique flexion views of the distal femur improve the ability to detect and estimate the size of osteolytic lesions that may exist adjacent to the femoral prosthesis. Fluoroscopy may also be helpful in eliminating small degrees of obliquity seen in conventional radiographs that may obscure radiolucent lines at component interfaces. This is particularly useful in evaluating whether or not a cementless implant is bone ingrown.
The hip-knee-ankle radiograph allows assessment of lower extremity alignment and is used to determine the proper femoral and tibial mechanical axes. The distal femoral cut is made based on the femoral mechanical axis. A line is drawn from the middle of the femoral head to the
P.199
middle of the knee. A perpendicular from this line at the level of the distal femur denotes a proper distal femoral resection. The tibial mechanical axis is defined in a similar fashion, drawing a line from the middle of the talus to the middle of the knee. A line perpendicular to this line at the level of the proximal tibia denotes a proper proximal tibial resection. Deformities of the femur and tibia, such as bowing or fracture that may compromise the position of intramedullary alignment guides or stem insertion are detected with this radiograph. This information allows templating of the revision distal femoral and tibial resections to recreate a neutral lower extremity alignment and alerts the surgeon to potential pitfalls if intramedullary alignment guides are used.
Exposure of the Revision Knee
A thoughtfully designed surgical approach is necessary to avoid potentially devastating healing problems and to obtain adequate access to the knee. These issues should be addressed during the initial history and exam of the patient. Inquiries regarding delayed wound healing, wound drainage, stiffness, or any other complications after primary knee surgery should be made. The examination should clearly document the location of previous skin incisions, presence of sinus tracts, mobility of the soft tissues and patella, active and passive range of motion, and whether an extensor lag is present. Investigating these issues preoperatively may prevent unexpected difficulties in the operating room.
The surgical incision must be of adequate length to fully expose the knee joint safely without applying excessive tension on the skin edges. It has been shown that a midline incision is less disruptive to the anterior arterial network of the knee. However, when multiple prior incisions are present, the most lateral skin incision should be used. Since the blood supply to the skin over the knee comes from medial to lateral, a medial incision in the presence of a previous lateral incision compromises the blood flow to the skin between the incisions. When a medial and lateral incision is encountered, the lateral incision usually should be used and extended in both directions. A flap is created, which must be done in the subfascial plane to preserve the dermal plexus' contribution to vascularity of the flap. Frequently it is necessary to incorporate or cross a previous incision. Any scar that must be intersected should be done in a perpendicular fashion. Similarly, the incorporation of a previous incision must avoid acute angles to minimize the risk of skin necrosis and a resultant narrow bridge of skin tissue. Adherent skin must be mobilized in the subfascial plane to identify the proper location of the capsular incision and allow retraction of the underlying tissues.
The capsular incision is performed along the medial border of the quadriceps tendon, the patella, and the patellar tendon with the knee in flexion. Less invasive approaches, such as the midvastus or subvastus incisions, generally do not provide sufficient exposure for revision surgery. Adhesions, synovium, and scars beneath the quadriceps tendon and throughout the medial and lateral synovial recesses are removed with the knee extended. Care must be taken while releasing scar over the epicondyles to prevent damage to the collateral ligaments. The medial joint capsule is released from the underlying tibial metaphysis, maintaining an intact sleeve around the medial border of the proximal tibia to the posterior midline. The tibia is externally rotated while flexing the knee until anterior subluxation of the proximal tibia ensues. With this patellar inversion technique, no attempt to evert the patella is made until the femoral and tibial components have been removed. Attempts to evert the patella early risks avulsion of the patellar tendon and should be discouraged in revision surgery.
When revision surgery is performed on a stiff knee, more extensile exposures may on occasion be necessary. The quadriceps snip is the most commonly used method to relax the extensor mechanism. It is accomplished by completely incising the quadriceps tendon at a 45-degree angle from distal medial to proximal lateral in line with the fibers of the vastus lateralis 4 to 6 cm above the patella. The quadriceps V-Y turndown is typically reserved for the near ankylosed knee. This technique divides the quadriceps tendon at its junction with the rectus muscle, at a 45-degree angle, in a distal and lateral direction. Although effective for exposure, this approach is frequently accompanied by extensor lag postoperatively. A tibial tubercle osteotomy provides excellent exposure and is preferable to the V-Y turndown in extremely stiff knees. However, persistent anterior knee pain and nonunion have been described with this technique. The routine use of extensile exposure is unnecessary and should be discouraged. More than 90% of revisions can be performed with the patellar inversion technique previously described.
Component Removal
Depending on the stability and type of fixation, the removal of total knee components can be a time-consuming and frustrating process. However, the benefits of a safe and orderly extraction must be considered. Avoidance of condylar bone loss or fracture can greatly facilitate a successful reconstruction. Familiarity with implant removal instruments and their proper use assists in making this step more efficient and less complex.
The order in which the implants are removed decreases the risk of complications and enhances the ability to remove the next component. The tibial polyethylene insert is removed first. Knowledge of implant-specific removal instruments for modular inserts is necessary. Most modular inserts can be removed with an osteotome that will disengage the plastic from the underlying tibial tray. Initial removal of the tibial insert will increase space in the knee that is needed to remove the other components.
The femoral component is generally removed second. The implant/cement interface for cemented implants is disrupted with thin osteotomes or ultrasonic devices. Working at this interface from both the medial and lateral sides limits bone loss. The implant/bone interface for cementless implants is most easily divided with thin osteotomes, power saws, or thin high-speed cutting burrs. The anterior condylar, distal, and chamfer interfaces are usually
P.200
readily accessible. The posterior condylar interface is more difficult to reach and may require angled osteotomes or a Gigli saw for disruption if this portion of the component is well fixed. Once loosened, the femoral implant can be removed by gently tapping the anterior flange with a metal punch and mallet in a distal direction. Vigorous disimpaction should be discouraged as this indicates that the fixation has not been sufficiently disrupted to proceed with removal.
Removal of the femoral component provides access to the tibial tray. The same instruments used to disrupt the femoral interfaces are also used here. However, cemented implants with a roughened surface may require a saw or ultrasonic device to separate the cement from the implant. After the component is loosened, the knee is hyperflexed and the tibia is anteriorly translated. Osteotomes can then be stacked under the tray to elevate it out of the proximal tibia.
Uncemented stems without biologic fixation typically are removed with their respective implant during the extraction process outlined above. However, well-fixed cemented stems and roughened stems secured biologically to bone can be very difficult to remove. If the stems cannot be disimpacted with the attached implant, the condylar portion should be disengaged or cut away from the stem. This provides direct access to the stem. Cemented stems can then be removed by breaking the cement/stem interface with a high-speed burr or ultrasonic device. Biologically fixed stems can be removed with trephines designed to remove cementless total hip arthroplasty femoral stems or a narrow high-speed burr. Rarely a femoral window or tibial tubercle osteotomy is required for the removal of well-fixed cemented or cementless stems.
Extraction of failed components is an often overlooked step of the revision. Use of proper tools and patience is warranted to avoid making the surgery more complex by removing excessive bone or initiating fractures.
Joint Alignment and Ligament Balance
With the failed components removed, adequate space should be available within the joint to observe bone quality and deficiency and begin preparing the femur and tibia for reimplantation. The hip-knee-ankle radiograph may be reassessed to determine whether an intramedullary guide may be used and to review the resection level and angle for the distal femoral and proximal tibial resections.
In most cases, the revision distal femoral cut may be referenced from an intramedullary rod that is secured within the diaphyseal cortex. If the femoral canal is not amenable to an intramedullary guide, an extramedullary guide or computer assistance may be used. The former is less accurate than intramedullary guidance, and the latter adds complexity to the case. Regardless, a flat cut is made with an oscillating saw through the cutting guide at an angle perpendicular to the mechanical axis. The amount of bone removed should be minimal to prevent elevating the joint line excessively. Exceptions to this include a significant preoperative flexion contracture from an inadequate primary distal femoral resection or if nonsupportive bone is encountered that will eventually require augmentation. The estimated level of the joint line is 2 to 2.5 cm distal to the epicondyles, 1 cm distal to the inferior pole of the patella with the knee in extension or more simply when the patella rests in the proper position in the trochlear groove in full extension.
The proximal tibial resection provides the foundation on which the revision is built. Therefore, the importance of creating an accurate cut that is perpendicular to the mechanical axis cannot be overstated. This resection can be referenced from intramedullary or extramedullary alignment guides. Intramedullary guides may be influenced by bowing within the proximal tibia or by the presence of sclerotic bone within the intramedullary canal.
Although several techniques exist, we prefer using the classic method of balancing the knee if the collateral ligaments are not attenuated. After the components are provisionally sized, a coronally symmetric extension gap is created with the appropriate medial and lateral releases. The height of this gap is measured with the knee in extension. The knee is then flexed and the flexion gap is measured and then tensioned. The appropriate AP cut guide is placed on the distal femur and is rotated until parallel with the tibial cut. A stylus is used anteriorly to prevent notching the distal femur. The size of the AP cutting block can then be manipulated to match the height of the flexion gap with the previously measured extension gap. Other methods used to establish rotation of the femoral component, such as the posterior condylar axis or trochlear axis, are difficult to reference in the revision situation because of previous bony resections and bone loss. While we prefer to use the classic method described above to set femoral rotation and balance the flexion gap, the epicondylar axis can be used to assess femoral rotation when collateral ligaments are severely attenuated.
After the flexion and extension gaps are balanced, the femoral box and chamfer cuts are made. If the use of augments has been calculated into the gap balancing, augments should be added to the femoral cutting blocks to avoid unnecessary bone resection. The necessity of stems, augments, and degree of articular constraint can now be determined.
Managing Bone Loss: Augments, Allograft, and Custom Implants
Some degree of bone loss is to be expected when performing a revision total knee replacement. Significant deficiencies may be seen when aseptically loosened components have been neglected or the primary total knee failed owing to osteolysis or infection. Although bone deficiencies cannot be comprehensively evaluated before the failed components have been removed, assessment of the preoperative radiographs should alert the surgeon to existing defects and allow preparation of a surgical strategy to manage these defects. Bone loss can be categorized into cavitary defects, which have an intact surrounding cortical rim, or segmental defects, with no surrounding cortex. Depending on the severity of the bone loss, cement, cement and screws, bone graft, modular metallic augments, and custom implants can be used to fill the deficiencies and recreate the joint line.
Cement is frequently used to fill small contained defects of both the femur and the tibia. When cement is used to
P.201
fill more significant defects, screws partially embedded in host bone can be added to strengthen the cement construct. This technique can be extremely effective in moderate-sized deficiencies, especially in older patients.
Bone graft has been used to address a wide variety of osseous deficiencies around the knee from small cavitary lesions to extensive segmental defects. Because autograft is usually in short supply, allograft bone is generally used in revision total knee arthroplasty. Although a thorough discussion of bone graft basic science is beyond the scope of this chapter, familiarity with the types of allograft, sterilization methods, storage process, disease transmission, and incorporation with the host is required before its use.
Particulate allograft is typically used to fill contained cavitary defects when sufficient host bone is present to provide structural support for the revision components. If healthy host bone is encountered at the base of the lesion, vascularization of the allograft and replenishment of bone stock can be expected.
Structural allografts are most commonly used to manage segmental defects that are too large to be managed with metallic augments. This scenario is commonly seen in cases of advanced osteolysis or infection. Femoral head allografts have been used successfully to fill large cavitary defects of both the distal femur and proximal tibia in these situations and provide a structurally sound surface on which stemmed cemented revision components are used.
Segmental defects that extend proximal to the collateral insertions on the femoral epicondyles impair ligamentous stability as well as component stability. This situation may be encountered with a comminuted supracondylar periprosthetic fracture or severe cases of osteolysis. Partial or full distal femoral allografts (depending on the bone defect location), may be cemented to stemmed revision components using a back table technique. The allograft-prosthetic composite is then secured to the host femur using a diaphyseal engaging stem. Another alternative in this situation is a salvage system or tumor prosthesis which uses a rotating hinge device. This is particularly useful in the elderly, limited-demand population.
Concerns regarding structural allograft resorption, nonunion, fracture, and disease transmission have encouraged the use of metallic devices to fill segmental osseous defects. Modular metallic augments use the predictable properties of metal to manage segmental deficiencies adjacent to both tibial and femoral revision prostheses. These devices allow the surgeon to re-establish the joint line, assist in component alignment, and adjust soft tissue balance to improve knee kinematics. The augments are available in various sizes and shapes, with corresponding trials, creating a versatile system that has greatly reduced the necessity for structural bone grafts and custom implants.
Tibial augmentation devices are available in full wedge, hemiwedge, and block geometries, depending on the manufacturer. They usually are applied to fill segmental medial, lateral, or combined defects of 5 to 20 mm in depth. The location and size of the defect determines which type of augment is appropriate to preserve supportive host bone. Although joint line elevation may be accomplished with the use of a thicker insert, this has the disadvantage of placing increasing stress on the locking mechanism of the tibial insert to the baseplate. Combined medial and lateral block augments will elevate the joint line without applying undue stress on the locking mechanism. A longer stem should be considered in this situation, as the augments will effectively shorten stem penetration into host tibia.
Femoral modular augments are available in the block geometry in various thicknesses and may be applied to the medial and/or lateral condyles distally and posteriorly. Distal augments fill segmental defects below the epicondyles or depress the joint line to its anatomic location. As in the tibia, a longer stem may be required to effectively engage the host femur when distal augmentation is employed. Posterior augments are beneficial in decreasing the flexion gap, which is usually greater than the extension gap in the revision situation. Laterally placed posterior augmentation assists patellar tracking by externally rotating the femoral component when posterolateral bone is deficient.
Significant bony anatomic deficiencies occasionally require tumor prostheses or custom implants. Modern tumor prostheses may have cemented or press-fit intramedullary fixation to the host bone and are equipped with modular diaphyseal segments. These features provide a versatile implant that has intraoperative flexibility. However, most tumor prostheses use a rotating hinge articulation that increases stresses placed on the bone/implant interface and may not be ideal for more active patients.
Use of Stemmed Components
As previously discussed, some degree of bone loss is to be expected when performing a revision total knee replacement. The use of stems on revision components is designed to transfer stress away from the damaged periarticular bone to the shaft. Contemporary revision total knee systems are equipped with numerous stem options: variable length stems designed to engage the metaphyseal or diaphyseal bone, cemented or press-fit interfaces, and straight or offset stems. Although the use of stems in revision total knee arthroplasty is routine, the appropriate use of the stem options available remains controversial.
Several biomechanical studies comparing cemented versus noncemented stems in cadaveric tibias reveal significantly less tibial tray micromotion in the cemented group. Similarly, retrospective clinical studies consistently show higher rates of radiolucent lines adjacent to noncemented stems at 18 months to five years postop. The only retrospective comparison revealed significantly greater radiographic stability when cemented stems were used.
Uncemented diaphyseal engaging stems have become popular because they are simple to use and help guarantee acceptable implant alignment in most circumstances. However, in some cases, diaphyseal stem engagement may compromise tibial component position owing to the limitations of a canal-filling stem. An unrecognized valgus tibial bow will malalign the tibial component into valgus when a canal-filling diaphyseal engaging stem is used. Also, the frequent anteromedial location of the shaft in reference to the plateau will cause the baseplate to overhang medially when a canal-filling stem is used. Although an offset tibial stem may decrease such malposition, a relatively narrow
P.202
metaphyseal engaging cemented straight or offset stem may be adjusted within the cement mantle to fully accommodate these anatomic variants.
Canal-filling stems can have a similar negative effect on the alignment of the femoral prosthesis. Because of the anterior femoral bow, a diaphyseal engaging stem may contact the anterior endosteal surface of the canal, causing flexion of the component. Alternatively, if the canal-filling stem slides past the anterior endosteal surface, the femoral component may translate anteriorly, both overstuffing the patellofemoral joint and increasing the flexion gap. As in the tibia, a narrow metaphyseal engaging cemented stem may be positioned more posteriorly within the cement mantle of the distal femur to decrease the flexion gap without being biased by the femoral bow.
The judicious use of offset stems may limit component malposition when canal-filling stems are used. However, the intraoperative flexibility, as well as biomechanical and radiographic comparisons, prompts some surgeons to continue to use cemented metaphyseal engaging stems in most cases.
Articular Constraint
Articular constraint refers to the degree of stability afforded to the knee joint by prosthetic design. Additional constraint must be supplied by the implants if the soft tissues around the knee are insufficient to maintain joint stability. In the revision situation, the choices of constraint include posterior stabilized articulations, nonlinked constrained prostheses, and rotating hinge constrained designs. Because increased levels of constraint generate larger stresses on the implant/bone interface, the least amount of constraint that provides a stable joint should be selected. Although a thorough preoperative examination may alert the surgeon to potential instability issues, a final decision regarding the degree of articular constraint often cannot be made until bone defects are addressed and ligament balancing has been accomplished.
Most revision total knee replacements can be accomplished with posterior stabilized implants. These prostheses provide minimal constraint through a congruent tibiofemoral articular surface and a spine and cam mechanism that promotes femoral rollback and prevents posterior displacement of the tibia in flexion. This design requires functional collateral ligament support for varus and valgus stability in flexion and extension. Equalizing the flexion and extension gaps is necessary to prevent the spine from dislocating posterior to the cam if an excessive flexion space exists.
Nonlinked constrained implants are used when one or both collateral ligaments are insufficient, creating varus or valgus laxity or an excessive flexion gap that cannot be balanced with the extension gap. This design is equipped with a taller and thicker polyethylene spine or post that may be reinforced with an underlying metal pin. The post closely approximates the intercondylar box providing rotational, translational, and varus/valgus support to the knee joint. Stems should be used to dissipate forces transmitted to the implant/bone interface.
Rotating hinge constrained components are generally reserved for cases of severe bone loss or global instability where the flexion gap is so excessive that the condylar post of a nonlinked constrained device will be unable to prevent posterior displacement of the tibia under the femur. Because rotating hinge devices do not constrain rotation, less stress may be placed on the implant/bone interface than with nonlinked constrained devices. Generally, the use of rotating hinge constrained components is discouraged except in elderly patients with global instability, severe bone loss, and low functional demands because the large osseous resection necessary for implantation limits salvage options (including arthrodesis).
Special Considerations
Infection in Total Knee Arthroplasty
Infection is the cause of failure in approximately 1% to 2% of total knee replacements. Therefore a high index of suspicion for infection is warranted during the evaluation of all failed total knees. Prompt diagnosis not only may decrease the risk of morbidity and mortality but also increases the available treatment options.
An investigation for infection should occur during the evaluation of every patient with a painful total knee. Unfortunately, the presentation varies considerably depending on the length of time since surgery, the duration of the infection, the virulence of the offending organism, the host status, and use of antibiotics. Nevertheless, a history of continuous knee pain, swelling, warmth, problems with postoperative wound healing, drainage, or an active infection in another area of the body prompts further scrutiny. Radiographs are valuable for assessing the components and periarticular bone quality but rarely distinguish infectious from aseptic failures.
Because of inaccuracies of diagnostic tests, none can be used in isolation to reliably predict the presence of infection, but when used in concert, they allow detection of approximately 90% of infections. Serologic tests including white cell count, sedimentation rate, and C-reactive protein are sensitive but relatively nonspecific screening tests. Routine use of various radioisotope scans appears impractical owing to their low sensitivity. Aspiration of the knee may provide the most useful preoperative information if there is clinical suspicion for infection and the patient has not received antibiotics within 2 weeks. Although the sensitivity of culture results after aspiration is limited, analysis of the synovial fluid revealing >2,500 white blood count (WBC)/mm and >90% polymorphonuclear (PMN; leucocytes) is highly sensitive and specific for infection in the prosthetic knee joint. If an infectious cause for total knee failure remains questionable intraoperatively, frozen histologic analysis by an experienced pathologist should be performed with tissue taken from multiple sites in search of evidence of acute inflammation. Despite clinical scenarios, such as the presence of inflammatory arthropathies or indolent organisms, that continue to impair our ability to distinguish septic from aseptic failures, the judicious use of diagnostic tests improves the ability to identify infected total knee replacements.
P.203
The treatment of an infected total knee depends on the timing and duration of the infection, the virulence of the organism, and the patient's overall health. Fluid and tissue cultures must be obtained to guide antibiotic therapy. Acute infections, whether in the early postoperative period or hematogenous in origin, are typically managed with attempted prosthetic retention. Open surgical debridement, radical synovectomy, and tibial insert exchange followed by a short period of intravenous (IV) antibiotics and chronic suppression are recommended in the proper situations. Required criteria include the following: no evidence of osteomyelitis, component loosening, or sinus tract formation; an organism of low virulence that is susceptible to antibiotics; and symptoms that have been present for a short period of time (i.e., <4 weeks postop or within 48 hours of a hematogenous infection). The presence of resistant organisms or an immunocompromised host appear to provide less favorable results with prosthetic retention.
Exchange arthroplasty is the preferred treatment for infections that have been present for >4 weeks or involve more virulent organisms. This process involves resection of the infected components, a thorough joint debridement, and prosthetic reimplantation.
Delayed exchange arthroplasty has been preferred by most surgeons in North America. Eradication of infection consistently approaches 90% in the literature with a two-stage approach. During the first stage, after component removal and debridement, a cement spacer that is impregnated with heat-stable antibiotics is used to discourage soft tissue contracture while delivering high local doses of antibiotics to the infected knee. Some surgeons currently fashion an articulating spacer to minimize bone loss and knee stiffness while improving patient function between stages. This technique also facilitates exposure at reimplantation without any apparent increase in reinfection rates. Systemic antibiotics usually are discontinued about 6 weeks after the spacer is placed. An aspiration prior to reimplantation may be performed and serologic tests for inflammation (erythrocyte sedimentation rate [ESR], C-reactive protein [CRP]) are followed for a gradual return to normalcy prior to reimplantation. Reimplantation usually is performed 2 to 3 months after the first stage. If serologic tests do not normalize, or if frozen section analysis at the time of reimplantation suggests continued active infection, a repeat debridement with articulating spacer exchange is performed and reimplantation is delayed.
Functional results after revision TKA for infection have improved with more rapid diagnosis and the availability of modern revision systems. However, results have been tempered by increasing rates of bacterial antibiotic resistance as well as more immunocompromised hosts. Failure to eradicate total knee infections after repeated attempts at reimplantation leaves the surgeon with few options. Resection arthroplasty is reserved for patients with poor baseline functional requirements. Arthrodesis remains a viable option in patients with irreparable extensor mechanism disruption, multiple recurrent infections, or an inadequate soft tissue envelope provided the contralateral limb and ipsilateral ankle and hip are functional. Above-knee amputation is generally reserved for situations where other reconstructive or salvage efforts are deemed futile.
Periprosthetic Fractures
Based on the Mayo Clinic Joint Registry, the prevalence of periprosthetic fractures adjacent to primary TKAs is 2.3% and rises to 6.3% after revision TKAs. The number of patients with this problem appears to be increasing as the volume of total knees being implanted climbs and patients live longer and more active lives. Distal femur fractures are seen more frequently than patellar or proximal tibia fractures. Issues relevant to the treatment of these periprosthetic fractures include the anatomic location, bone quality, functional requirements of the patient, and whether the fracture occurred intraoperatively or postoperatively.
Intraoperative distal femur fractures typically involve the condyles and occur during removal of failed components or during the insertion of a posterior stabilized implant into an incompletely prepared intercondylar box. Although screws may provide sufficient fixation for a nondisplaced fracture, transferring stress away from the condyle to the shaft through the use of a stemmed component is warranted when fractures are unstable.
Most postoperative fractures of the distal femur are traumatic in origin. Previous notching of the distal femur may contribute to torsional failure. Treatment is based on fracture displacement and component stability. Nonoperative treatment using a cast or brace should be considered for nondisplaced fractures when the femoral component remains stable, especially while treating a patient with low functional demands. Close radiographic follow-up is required to ensure that the fracture remains well aligned.
Displaced fractures with stable implants are treated with rigid internal fixation or intramedullary nails. Although several types of plates have adequate reported results, fixed-angle devices with proximal and distal locking screws are an attractive option. Fixation in the osteopenic bone that is frequently encountered is improved significantly when locking screws are used. The precontoured distal femoral plate may also assist with indirect reduction of the fracture and can be placed through a smaller incision to produce minimal disturbance to the fracture hematoma. Modern retrograde intramedullary nails also provide rigid fixation. However, distal fixation can be a problem in some fracture patterns. The ability of a specific posterior stabilized femoral component to accept a retrograde nail must be investigated preoperatively. Occasionally bone grafting or cement augmentation is required to enhance distal fixation when extensive comminution or osteopenia is present.
Distal femur fractures with loose femoral components require component revision in addition to fracture fixation. A stemmed component with adjuvant fixation is used. However, if comminution is severe, a rotating hinge tumor prosthesis can be used to reconstruct the knee.
Periprosthetic fractures of the tibia plateau that occur intraoperatively may be secured with cancellous screws or a low-profile plate. A longer tibial stem is recommended to protect the fractured plateau. Metaphyseal fractures that occur adjacent to the stem are usually nondisplaced and vertical in orientation. They often occur during removal of a failed tibial component. When recognized, they should also be treated with a longer stem to bypass the defect. Nondisplaced intraoperative fractures that are distal to the
P.204
stem have been successfully treated with a brace or cast and limited weight bearing. However, displaced fractures distal to the stem require internal fixation for stability.
Postoperative periprosthetic fractures of the tibia plateau commonly lead to tibial implant failure. Prosthetic malalignment, osteopenia, osteolysis, and osseous necrosis have been implicated as predisposing factors. Revision of the tibial component with a stemmed implant and managing the bone loss with modular augments or bone graft are recommended.
Postoperative fractures adjacent to the stem of a stable tibial prosthesis are generally caused by a traumatic mechanism. These fractures are often nondisplaced and may be treated in a long-leg cast. Displaced fractures are best managed with a plate and screw construct. Metaphyseal fractures with an unstable prosthesis require revision to a longer stemmed prosthesis. Stable tibial components with a postoperative fracture distal to the stem often can be treated nonoperatively in a long-leg cast with protected weight bearing. If the tibial prosthesis is loose, revision surgery is indicated. However, delayed reconstruction after the fracture has healed in a cast may simplify the surgery.
Tibial tubercle fractures that occur in the postoperative setting may be the result of trauma or owing to the nonunion of a tubercle osteotomy. Nondisplaced fractures are amenable to casting in extension. Displaced fractures require reduction and internal fixation with screws or wires.
Intraoperative periprosthetic patella fractures usually occur during the removal of a failed patellar implant with deficient underlying bone. Postoperative fractures generally are the consequence of direct impact onto the knee or the result of indirect trauma from a forceful quadriceps contraction. Increased vulnerability to patella fractures occurs after excessive bone resection, overstuffing of the patellofemoral joint, or eccentric position of the component. The method of periprosthetic patella fracture management is controversial and depends on the location, extensor mechanism function, stability of the implant, and medical status of the patient.
Patella fractures with a stable implant and a functional extensor mechanism generally are treated nonoperatively with a brace or cast keeping the knee in extension while allowing full weight bearing. Occasionally a nonunited marginal fracture fragment may require excision for continued pain. Surgical repair is recommended for fractures associated with a significant extensor lag. When fractures with an intact extensor mechanism are accompanied by a loose implant, the decision to simply remove the component versus attempted revision should be based on the remaining patellar bone stock.
Stiffness Following Total Knee Replacement
Difficulties with obtaining adequate knee motion after TKA can be a frustrating experience for the patient as well as the surgeon. Although the normal knee range of motion is approximately 0 to 140 degrees, most functional activities can be accomplished with 95 to 100 degrees of flexion. Less excursion of the knee makes walking, stair use, and sitting problematic.
Arthrofibrosis is the most commonly identified cause of limited motion after a technically sound total knee replacement has been performed. Dense adhesions develop within the joint, resulting in limited flexion and extension. Aggressive range of motion protocols have been instituted at most hospitals to prevent prolonged immobilization, which can exacerbate this condition. However, it is difficult to predict which patients will require further intervention to maintain motion. A careful manipulation under general or regional anesthesia may restore knee motion to near its observed intraoperative level if performed prior to the maturation of adhesions in the first 6 weeks postoperatively. Once scar has matured, arthroscopic lysis of adhesions has been used effectively to disrupt fibrous bands around the fat pad and superior pole of the patella. Arthroscopic release of the posterior cruciate ligament also has provided significant improvements in motion when a cruciate-retaining prosthesis presents with a stiff knee.
Physical examination and radiographs of the knee may reveal technical problems with the primary surgery that may impede motion. Retained osteophytes, implant malalignment, improper component sizing, and imbalance of flexion and extension gaps can mechanically limit motion. Unlike arthrofibrosis, diminished motion was most likely present at the completion of the primary surgery in these situations. Manipulation and arthroscopic releases are rarely helpful under these circumstances. Revision surgery is necessary to improve these motion limitations. Since significant complications may accompany revision TKA, careful consideration must be taken before embarking on this effort to re-establish motion.
Retained osteophytes located on the posterior femoral condyles can impair both flexion and extension. A mechanical block caused by impingement of the osteophyte on the posterior tibia may limit flexion. Extension is impaired by the mass of the osteophyte, tenting the posterior capsule. Removal of the tibial insert is required to access the posterior knee prior to the cautious removal of the osteophytes with an osteotome.
Component malalignment can also limit motion, particularly concerning the slope of the tibia baseplate in the sagittal plane. Excessive posterior slope may limit full extension and cause flexion instability. Anterior slope of the baseplate may cause hyperextension of the knee and limit flexion. Knowledge of the required slope for the specific implant in question is needed prior to its revision.
Improper femoral component sizing and the mismatch of flexion and extension gaps are closely related. A common error is underresection of the distal femur. This produces a tight extension gap and a resultant flexion contracture. Similarly, an insufficient posterior femoral resection will lead to oversizing the femoral component and create a tight flexion space. Flexion of the knee may also be impaired if insufficient patellar bone is resected, creating an “overstuffed” patellofemoral joint.
Surgical management of a stiff total knee should be undertaken cautiously because several mechanisms are not amenable to operative repair. A thorough evaluation is necessary to determine the appropriate treatment. Although surgical intervention has provided statistically significant improvements in motion, functional improvements are not consistently realized.
P.205
Polyethylene Wear and Osteolysis
Modern ultra–high-molecular-weight polyethylene provides a low-friction surface intended to articulate with a polished femoral component. However, even with a highly congruent articulation, small polyethylene particles are released from the tibial insert owing to the complex shearing and rotational motions of the knee joint, even during the normal gait cycle. Furthermore, tibial locking mechanisms in modular components allow some degree of insert micromotion, producing backside polyethylene wear. Several additional variables have been implicated in accelerating particle release: abrasive wear, method of polyethylene sterilization and shelf life, coronal alignment of implants, congruency of prosthetic articulations, patient size, and activity levels. Polyethylene wear produces various total knee problems ranging from aseptic synovitis to osteolytic defects that can impair component fixation and complicate reconstructive efforts. As total knee replacements are now implanted in young, active patients, long-term survivorship may become increasingly impaired by material limitations.
The clinical triad of effusion, pain, and progressive changes in coronal alignment of the knee characterizes accelerated polyethylene wear. The effusion is related to the inflammatory reaction induced by macrophage engulfment of the shed particles creating a boggy synovitis that may be accompanied by pain. The asymmetric wear may produce symptomatic instability. If neglected at this point, progressive osteolysis will ensue. Therefore, it is important to revise the patient presenting with these symptoms promptly.
Management of osteolysis depends on the extent and location of the lesions as well as component stability. Preoperative radiographs, including oblique flexion views, should alert the surgeon to loose components and allow an estimation of bone loss. Revision components with alternate levels of articular constraint, stems, modular augments, and allograft may be necessary to facilitate reconstruction. The damaged tibial insert is removed and a full synovectomy is performed. All components should be stressed to ensure stability. If stable, consideration of filling defects with particulate allograft or cement is warranted. Unstable components are removed and revised as described in the section above.
The continuum of problems associated with polyethylene wear highlights the need for routine postoperative surveillance of total knee patients. It has also led to innovative devices intended to reduce backside wear, which appears to correlate with osteolysis. All-polyethylene, nonmodular metal-backed, and rotating-platform tibial components are currently being investigated to determine their ability to limit osteolysis.
Suggested Readings
Barrack RL, Smith P, Munn B, et al. Comparison of surgical approaches in total knee arthroplasty. Clin Orthop. 1998;356:16–21.
Emerson RH, Head WC, Malinin TI. Reconstruction of patellar tendon rupture after total knee arthroplasty with an extensor mechanism allograft. Clin Orthop. 1990;260:154–161.
Fehring TK, Griffin WL. Revision of failed cementless total knee implants with cement. Clin Orthop. 1998;356:34–38.
Fehring TK, McAvoy G. Fluoroscopic evaluation of the painful total knee arthroplasty. Clin Orthop. 1996;331:226–233.
Fehring TK, Odum S, Griffin WL, et al. Patella inversion method of exposure in revision total knee arthroplasty. J Arthroplasty. 2002;17:101–104.
Fehring TK, Odum S, Griffin WL, et al. Early failures in total knee arthroplasty. Clin Orthop. 2001;392:315–318.
Felix NA, Stuart MJ, Hanssen AD. Periprosthetic fractures of the tibia associated with total knee arthroplasty. Clin Orthop. 1997;345:113–124.
Ritter MA, Keating M, Faris PM. Screw and cement fixation of large defects in total knee arthroplasty. A sequel. J Arthroplasty. 1993;8:63–65.
Spangehl MJ, Masri BA, O'Connell JX, et al. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg. 1999;81-A:672–683.
Springer BD, Hanssen AD, Sim FH, et al. The kinematic rotating hinge prosthesis for complex knee arthroplasty. Clin Orthop. 2001;392:283–291.
Whiteside LA, Ohl MD. Tibial tubercle osteotomy for exposure of the difficult total knee arthroplasty. Clin Orthop. 1990;260:6–9.