I. Causes of Implant Failure
A. Osteolysis
1. Wear rate—Many factors can affect the wear rate of ultra-high molecular weight polyethylene (UHMWPE) in total knee arthroplasty (TKA):
a. Sterilization and manufacturing method
b. Presence of third-body debris
c. Motion between the modular tibial insert and metal tray (resulting in backside wear)
d. Roughness of the femoral component counterface
e. Alignment and stability of the knee arthroplasty
f. Biomechanical demands or activity level of the patient
2. Differences between wear mechanisms in total hip arthroplasties (THAs) and TKAs
a. The hip is a congruent joint with a relatively large contact area that produces lower contact stress. At low contact stress, surface wear mechanisms (abrasion and adhesion) predominate.
b. As a result of the higher contact stresses and moving contact area in the knee, alternating tensile and compressive stresses are created in a tibial UHMWPE insert that can lead to fatigue (delamination and pitting) wear mechanisms.
c. Surface wear mechanisms produce smaller particles than do fatigue wear mechanisms. Smaller particles can elicit more of an osteolytic response than larger particles.
d. Wear particles are generally smaller in THAs than in TKAs.
e. Osteolysis appears to be more common in THAs than in TKAs.
*Michael D. Ries, MD, or the department with which he is affiliated has received royalties from Smith & Nephew.
f. When osteolysis does occur in TKAs, it can lead to the development of expansile bone defects.
3. Biologic response
a. Different patients may respond differently to wear debris, but smaller wear particles (<10 μm) are more readily absorbed by macrophages, which then release cytokines.
b. The cytokines signal osteoclasts to resorb bone, resulting in well-demarcated cystic lesions in the periarticular bone (
Figure 1).
c. Femoral osteolysis is difficult to detect on an AP radiograph because the lesions are typically located in the posterior condyles and obscured by the femoral component, whereas tibial lesions are usually more readily visible. Oblique radiographs are often helpful for detecting lesions in the femur.
[Figure 1. AP radiograph of the knee demonstrates particulate wear debris-induced osteolysis resulting in a well-demarcated uncontained lateral tibial defect (arrows).]
B. Loosening
1. Type of fixation
a. Both cementless and cemented TKAs have resulted in satisfactory outcomes.
i. Early loosening is a more common complication of cementless TKA.
ii. Cement fixation is currently used in the United States for the vast majority of primary TKAs.
b. Late failure of TKA results from wear more often than from loosening, although the two mechanisms can be related. Osteolysis that results in loss of bony support for the prosthetic components or disrupts the bone-implant interface may lead to mechanical loosening.
2. Alignment
a. Limb malalignment causes asymmetric loading of the knee, which also can result in early loosening. Loosening appears to occur more frequently with varus malalignment than with valgus malalignment.
b. Tibial loosening typically presents as a change in implant position or alignment associated with varus or valgus subsidence of the component. Pain occurs more during weight-bearing activity than at rest, and tenderness is localized to the tissues in proximity to the loose component.
C. Arthrofibrosis
1. The process by which pathologic scar tissue forms after TKA and restricts functional range of motion is relatively poorly understood.
2. Arthrofibrosis may develop in patients who have normal intraoperative range of motion. However, passive flexion, extension, or both can become restricted and painful.
3. The response to both nonsurgical and surgical treatment is often unsatisfactory.
4. Arthrofibrotic scar contains dense fibrous tissue with abundant fibroblasts.
5. Heterotopic bone is frequently found in patients with arthrofibrosis.
6. Stiffness may result from inadequate postsurgical pain management or rehabilitation or from a biologic process that causes rapid proliferation of scar tissue.
7. Genetic factors may also play a role, although it is difficult to predict which patients are at increased risk for arthrofibrosis after TKA.
D. Instability
1. Mediolateral instability
a. Gross instability caused by loss of collateral ligament support may result from intraoperative collateral ligament laceration or postsurgical trauma.
[
Figure 2. Flexion lateral radiograph of a posterior cruciateretaining (CR) TKA demonstrates anterior subluxation of the femur on the tibia, or "paradoxical motion," which is often associated with flexion instability.]
b. Stress radiographs can help identify or confirm collateral ligament disruption.
c. Loss of collateral ligament support requires use of an implant constrained to varus and valgus stress. If ligament disruption is identified intraoperatively, however, then primary repair and postsurgical bracing without use of a fully constrained implant can provide satisfactory results.
2. Flexion instability
a. Soft-tissue laxity can develop after surgery despite appropriate mediolateral and flexion-extension gap balancing.
b. As with patients in whom arthrofibrosis develops, it is difficult to predict which patients are at increased risk of developing postsurgical attenuation of periarticular soft-tissue constraints.
c. Both posterior cruciate-retaining (CR) TKAs and posterior cruciate-substituting (PS) TKAs require sacrifice of the anterior cruciate ligament, which can result in flexion instability, despite intact collateral ligaments.
d. Patients with symptomatic flexion instability usually report vague pain and swelling after activity and have laxity to varus and valgus stress in flexion. Radiographs typically demonstrate "paradoxical motion," or anterior subluxation of the femur on the tibia in flexion, rather than rollback (Figure 2).
e. Symptoms may be controlled with activity restrictions, bracing, nonsteroidal anti-inflammatory drugs, and muscle strengthening exercises.
i. If these measures fail, revision TKA is appropriate.
ii. Conversion of a CR to a PS TKA is beneficial in most revision situations, but more reliable results are achieved with revision to a constrained TKA if the flexion-extension gap balance is not restored using less articular constraint.
E. Infection—For a discussion of infection associated with TKA, see chapter 105.
II. Evaluation of the Painful TKA
A. Overview
1. The source of pain after TKA may be difficult to determine. The workup should include evaluation for infection, neurogenic pain, and mechanical sources of pain.
2. Evaluation should include a thorough history and physical examination, laboratory studies, and plain radiographs. Additional nuclear medicine studies or specialized imaging also may be necessary.
3. A history of pain that developed immediately after surgery and then persisted (no pain-free interval) and pain with rest as well as weight bearing suggests an inflammatory and/or neurogenic source of pain.
4. Pain during weight-bearing activity or knee motion is consistent with a mechanical source of pain.
B. Infection and neurogenic pain
1. Infection is usually associated with an elevated erythrocyte sedimentation rate and C-reactive protein level and can be detected by aspiration. However, false-negative and false-positive results can occur, and additional imaging studies are often necessary.
2. Pain associated with localized warmth and swelling that occurs more after activity and is relieved with rest is less consistent with infection and more typical of soft-tissue inflammation resulting from postsurgical rehabilitation of the soft tissues during exercise.
3. Pain that improves with analgesic medications for neuropathic pain (gabapentin, pregabalin, and trycyclics) or local trigger point or epidural injections supports the diagnosis of neurogenic pain.
C. Mechanical causes of pain
1. Overview
a. Mechanical causes of early pain after TKA include patellar maltracking, patellar clunk, tibiofemoral instability, and periprosthetic fracture.
b. Patellar problems are usually evident on physical examination since the location of pain is restricted to the patellofemoral joint.
c. Patellar clunk is a rare complication of a PS TKA and occurs when a fibrous nodule at the inferior pole of the patella catches in the trochlear groove during knee extension. Symptoms are relieved by open or arthroscopic excision of the fibrous nodule.
d. Patellar maltracking and subluxation may result from dehiscence of the medial retinacular arthrotomy or femoral or tibial component internal rotation.
e. Rotational orientation of the femoral component may be assessed to some extent on an axial view of the patella, but it is better quantitated using a CT scan. Symptomatic patellar subluxation or maltracking resulting from internal rotation of the femoral or tibial components requires revision of the malaligned components (
Figure 3).
2. Flexion instability
a. History and physical examination
i. A history of pain and effusion that occurs after activity and is relieved with rest is consistent with flexion instability.
ii. Flexion instability caused by intact but attenuated soft-tissue constraints can be detected by varus and valgus stress testing in both flexion and extension.
b. Radiographic evaluation
i. Flexion instability is more common with CR TKAs than with PS TKAs and is associated with paradoxical motion, or roll forward of the femoral component, which can be seen on flexion lateral radiographs as anterior subluxation of the distal femur on the tibia (Figure 2).
ii. Complete dislocation of a PS knee presents with gross instability in flexion on physical examination and posterior displacement of the tibia on the femur (
Figure 4).
3. Loosening and wear
a. History and examination
i. Pain that develops late after TKA is more often associated with loosening or UHMWPE wear, although late hematogenous infection can occur and should be included in the differential diagnosis.
ii. Tenderness is usually localized to the area over the loose component.
[Figure 3. Internal rotation of the femoral component. A, Radiograph demonstrates considerable patellar tilt associated with an internally rotated femoral component. B, Intraoperative view of the same knee demonstrates that the femoral component is internally rotated relative to the epicondylar axis (white line).]
b. Radiographic evaluation
i. Wear can be seen radiographically as asymmetric height of the tibial plateaus, although rotation and flexion of the knee can alter the projected height of the joint space, making radiographic measurements of wear inaccurate.
ii. Loosening occurs when there is subsidence or displacement of the component or a symptomatic complete or progressive radiolucency.
III. Classification of Bone Defects
A. Assessment of tibial and femoral bone loss
1. Presurgical planning
a. Bone loss should be assessed during the preoperative planning process as well as during the revision surgery.
b. Preoperative evaluation often underestimates the extent of bone loss; nonetheless, appropriate materials must be available for reconstruction during revision TKA.
2. Intraoperative evaluation—Iyntraoperative assessment is the most accurate method of assessing remaining bone stock and determining the most appropriate method of reconstruction.
3. Classification of defects
a. Bone loss may be classified by defect size, location, depth, and the presence or absence of an intact peripheral rim of bone upon which to place a prosthesis or contain bone graft.
[Figure 4. Lateral radiograph demonstrates dislocation of a PS femoral component over the tibial post.]
b. The Anderson Orthopaedic Research Institute bone defect classification system (
Table 1) provides some guidelines for management of bone defects.
i. The classification is applied independently to the femur and tibia (
Figure 5).
ii. It is based on the amount of metaphyseal bone that remains following implant removal.
[Figure 5. Anderson Orthopaedic Research Institute classification of bone defects. A, Type 1 femoral defect. B, Type 1 tibial defect. C, Type 2 femoral defect. D, Type 2 tibial defect. E, Type 3 femoral defect. F, Type 3 tibial defect.]
iii. It does not specify whether the defects are contained or uncontained, an important consideration in the utilization of particulate graft, which is more readily impacted into contained defects.
B. Patellar bone loss
1. Patellar bone loss is not part of most defect classification systems, but it must be dealt with regularly in revision surgery.
2. Bone deficiency is usually central, resulting in a concave defect.
3. The amount and vascularity of remaining patellar bone determines the feasibility of placing a new patellar implant.
IV. Surgical Treatment
A. Medial parapatellar approach
1. Most revision TKAs can be adequately exposed through a medial parapatellar arthrotomy.
2. Exposure can be facilitated by mobilizing the extensor mechanism by removing retropatellar adhesions and performing lateral retinacular release and subperiosteal dissection of the proximal medial tibia. This permits more tibial external rotation. However, particularly for knees with limited presurgical motion, more extensile exposure is required to avoid patellar tendon avulsion.
[Table 1. Anderson Orthopaedic Research Institute Bone Defect Classification System]
[
Figure 6. Exposures for revision TKA. A, Diagram of a rectus snip. The rectus tendon is incised obliquely to decrease the tethering effect of the extensor mechanism and permit exposure of the knee with less tension on the patellar ligament insertion at the tibial tubercle. B, Diagram of a V-Y turndown. The proximal end of a rectus snip is extended distally and laterally. This provides wide exposure and permits lengthening of the extensor mechanism but extensor lag may occur after surgery.]
B. Extensile exposure
1. Extension of the medial parapatellar arthrotomy can be performed proximally with a rectus snip or V-Y quadriceps turndown, or distally with tibial tubercle osteotomy.
a. Rectus snip (Figure 6, A)
i. An oblique medial-to-lateral transection of the rectus tendon at the proximal portion of the arthrotomy (rectus snip) does not appear to compromise long-term knee function and can relieve some of the tethering effect of a contracted extensor mechanism.
ii. The exposure afforded with use of a rectus snip is not as extensile as with a V-Y quadriceps turndown or tibial tubercle osteotomy.
b. V-Y quadriceps turndown (Figure 6, B)
i. This permits lengthening the extensor mechanism, which may be helpful in the treatment of arthrofibrosis, but this results in extensor lag.
ii. Vascularity to the rectus tendon is substantially disrupted after V-Y quadriceps turndown, which can further contribute to extensor lag.
c. Tibial tubercle osteotomy
i. Elevation of an osteotomy containing only the tibial tubercle that is detached from the anterior compartment muscles is associated with a high rate of nonunion. However, use of a long osteotomy of the tibial tubercle in continuity with the tibial crest and attached anterior compartment muscles maintains vascularity of the osteotomized bone fragment and a distal soft-tissue tether to prevent proximal migration of the bone fragment. Reliable union rates have been reported using this osteotomy technique.
ii. Tibial tubercle osteotomy is best indicated for cases with adequate tibial bone stock, while those with severe osteolysis or osteoporosis of the proximal tibia in which fixation of the osteotomy would be compromised may be better treated using a proximal (rectus snip or V-Y quadriceps turndown) transection of the rectus tendon for more extensile exposure.
C. Management of bone defects
1. Reconstructive options—Bone defects can be reconstructed with morcellized or structural allograft, synthetic bone-graft substitutes, cement, porous-coated or cemented metal augments, or a combination of materials.
2. Contained defects
a. Contained defects can be filled with cement or bone graft.
b. Bone grafting can restore bone stock and may be more appropriate for younger patients who could require future revision surgery.
c. Since cancellous grafts do not provide structural support, but revascularize more rapidly than structural grafts, they are most appropriate for contained defects.
3. Noncontained defects
a. Noncontained defects imply loss of cortical, structurally supportive bone and should be managed by restoration of structural stability. This requires augmentation of the defect with metal augments or structural allograft.
b. Although a structural allograft can heal to host bone and provide mechanical support of the revision implant, it cannot be expected to fully revascularize and, with extensive revascularization, may eventually collapse.
4. Stem fixation
a. Use of stemmed components is necessary to provide additional implant stability if metaphyseal fixation is compromised by bone loss or to protect bone grafts from weight-bearing stresses during postsurgical healing.
b. Stems may be cemented or cementless and have variable lengths. The choice of stem length and fixation depends on many factors, including:
i. The mechanical stability of metaphyseal fixation achieved
ii. The quality of diaphyseal and metaphyseal bone stock
iii. Weight-bearing capability of bone grafts or augments used
iv. Biomechanical demands of the patient
v. Amount of implant constraint
D. Choice of implant constraint—Revision TKA may be performed using an unconstrained CR or PS implant, constrained PS prosthesis, or a fixed or rotating hinge mechanism.
1. Unconstrained CR implant
a. An unconstrained CR implant requires an intact posterior cruciate ligament and collateral ligament support.
b. This may be appropriate for revision of a failed unicompartmental arthroplasty with minimal bone loss and intact ligament supports.
2. Unconstrained PS implant—Use of an unconstrained PS implant requires intact collateral ligaments with balanced mediolateral soft tissues and flexion and extension spaces.
3. Constrained PS or hinge
a. If one or both collateral ligaments are deficient or adequate soft-tissue balance cannot be achieved, then use of additional prosthetic constraint is appropriate.
b. A constrained PS prosthesis includes a wide tibial post that fits tightly into the femoral component box. This results in constraint to varus-valgus motion and rotation.
c. A rotating hinge prosthesis contains an axle that links the femoral and tibial components and provides stability to varus and valgus stress, but permits rotation.
d. Since the constrained PS relies on the UHMWPE post to provide constraint while the hinge uses a metal axle, the hinge is generally considered to be more rigidly constrained and indicated for cases with complete loss of collateral ligament support.
e. The hinge mechanism also limits hyperextension; thus, a hinge may be a better choice than a constrained PS device if the extensor mechanism is deficient, because damage to the tibial post can occur if a constrained PS prosthesis is hyperextended.
f. Most hinge implants require more femoral bone removal than a constrained PS prosthesis to accommodate the hinge mechanism.
V. Complications
A. Pain
1. Etiology and prognosis
a. Activity-related pain after revision TKA can be expected for 6 months or more after surgery.
b. Pain associated with soft-tissue inflammation should gradually diminish during this time.
c. Chronic neurogenic pain that is not consistent with a mechanical source or occult infection may occur after TKA.
d. Patients with greater presurgical pain appear to have an increased risk of developing chronic postoperative pain.
2. Treatment
a. Persistent neurogenic pain should be treated with a multimodal pain management approach, local or epidural injections, and manipulation if associated with stiffness. However, the response to treatment is often poor and requires a long-term pain management program.
b. Surgical treatment of a chronically painful TKA with no mechanical source or evidence of infection is usually associated with poor outcome.
c. If arthrofibrosis is also present, motion can be improved by surgery, but particularly for chronically painful TKAs without limited motion, revision is unlikely to lessen the pain.
B. Stiffness
1. Early rehabilitation including passive- and active-assisted range of motion is important to avoid limited motion after TKA.
2. If motion remains restricted, manipulation or occasionally revision surgery for arthrofibrosis may be necessary.
3. Arthroscopic resection of arthrofibrotic scar and open debridement with tibial insert exchange have been associated with variable results.
4. Modest gains in range of motion can be obtained with revision TKA and wide resection of periarticular arthrofibrotic scar, although pain may still persist.
5. Infection must be ruled out.
C. Infection
1. Superficial infection that clearly does not involve the knee joint can be treated with antibiotics alone, while intra-articular infection requires prompt surgical management.
2. Early infection can be treated with debridement and retention of the components, while late or chronic infection requires one- or two-stage exchange of the prosthetic components (see chapter 105).
D. Skin necrosis
1. Risk factors
a. Skin necrosis after TKA can rapidly lead to infection of the prosthetic components. Prior scars should be incorporated into the skin incision when possible to minimize the risk of skin necrosis after revision TKA.
b. Vascularity of the skin over the knee affects the rate of healing postoperatively and the risk of necrosis.
i. The lateral skin edge is more hypoxic than the medial edge. This suggests that when multiple prior scars are present, the most vertical lateral incision should be used to minimize skin hypoxia.
ii. The choice of scars to be used for the skin incision also depends on the length, orientation, and proximity of the scar to the knee.
2. Treatment
a. Skin tension can affect its vascularity. Knee flexion further increases skin tension and reduces skin oxygen tension.
b. Particularly for patients with multiple risk factors for developing wound complications, avoidance or delayed used of constant passive motion and early range-of-motion exercises may be beneficial in reducing the development of skin necrosis.
c. If skin necrosis does occur, early treatment will minimize the risk of deep infection of the prosthetic components.
d. Necrosis of the proximal wound including the area over the patella may be treated by local wound care and skin grafting. However, necrosis over the tibial tubercle or patellar tendon requires muscle flap coverage to prevent extensor mechanism disruption and deep infection.
E. Extensor mechanism disruption
1. Disruption of the extensor mechanism after TKA can occur from patellar tendon tear or avulsion, patellar fracture, or quadriceps tendon tear.
2. Primary repair of a chronic extensor mechanism disruption without autogenous or allograft soft-tissue augmentation is associated with a high risk of failure.
3. Semitendinosus or fascia lata autograft can be used to augment the primary repair.
4. Extensor mechanism allograft using Achilles tendon (tendon with calcaneal bone block in continuity) or patellar tendon (tibial tuberosity, patellar tendon, patella, and quadriceps tendon in continuity) has been reported most frequently for extensor mechanism disruption after TKA, although failure caused by intraoperative undertensioning and graft attenuation can occur.
5. The graft should be sutured in maximal tension with the knee in full extension and postsurgical range of motion restricted for 6 weeks after surgery to minimize problems associated with graft attenuation.
6. The medial gastrocnemius muscle and tendon can be used to reconstruct the extensor mechanism and to provide soft-tissue coverage, particularly in the setting of infection or wound necrosis. The distal tendinous portion is harvested along with the medial muscle belly and retracted proximally over the anteromedial aspect of the knee, allowing attachment to the remaining extensor mechanism.
F. Neurovascular problems
1. Although the tibial nerve, artery, and vein lie in close proximity to the posterior aspect of the knee and are vulnerable to direct trauma during surgical dissection, neurovascular injury is a rare complication of revision TKA.
2. Patients with preexisting vascular disease are more likely to have an increased risk of vascular injury.
3. Tourniquet use may further increase the risk of injury, so for patients with preexisting vascular disease, avoiding or minimizing tourniquet use may decrease the incidence of vascular complications.
4. The peroneal nerve is located peripheral to the center of the knee and is subject to injury from traction, particularly during correction of valgus and flexion deformity. If loss of peroneal nerve function is identified postsurgically, treatment by placing the knee in flexion can diminish tension on the nerve and may facilitate recovery.
5. For patients with chronic peroneal nerve dysfunction, late surgical exploration and decompression of the nerve may also be beneficial.
VI. Salvage Procedures
A. Arthrodesis
1. Arthrodesis is a viable salvage option to permit ambulatory function after failed revision TKA. However, bone loss usually results in considerable shortening and leg-length inequality, and the loss of range of motion is a significant functional impairment for many elderly patients or those with ipsilateral hip or ankle problems or back pain.
2. Arthrodesis techniques include intramedullary rodding with either a long rod from hip to ankle or with modular compression devices inserted through the knee; dual plating; and external fixation. Although each of these techniques can result in successful arthrodesis, more reliable results have been reported with either intramedullary rodding or dual plating.
B. Amputation
1. Above-knee amputation is considered by many patients as too disfiguring to accept.
2. Amputation does, however, provide a single definitive treatment of most complex failed revision TKAs.
3. Use of a well-fitting prosthesis may permit more comfortable sitting activity than arthrodesis and comparable ambulatory function.
C. Revision TKA
1. Bone loss, lack of ligament support, inadequate soft-tissue coverage, and extensor mechanism problems are more common to reconstruction of a failed revision TKA than to a failed primary TKA.
2. If successful revision cannot be performed or early failure occurs after a second revision, then bone stock may not be adequate to perform a fusion, and amputation becomes necessary.
VII. Clinical Results/Outcomes
A. Comparison of revision TKA to primary TKA
1. Pain relief and knee function after revision TKA are generally less favorable than results of primary TKA. However, patients who have revision TKA are older than the population of primary TKA patients and may have more complex medical or orthopaedic impairments that limit their overall functional ability.
2. Although less predictable than primary TKA, satisfactory improvement in pain and function can be expected with most revision TKAs.
B. Factors influencing outcome
1. Presurgical diagnosis and extent of reconstructive surgery required
2. Knee range of motion
3. Extensor mechanism function
4. Collateral ligament support
5. Quality of skin and soft tissues
6. Remaining bone stock
Top Testing Facts
1. Contact stresses in a TKA are higher than in a THA, which can cause fatigue wear of UHMWPE.
2. Early failure after TKA usually results from infection, malalignment, instability, and arthrofibrosis, while late failure more typically occurs from wear and loosening.
3. Pain during weight-bearing activity after TKA suggests a mechanical cause such as loosening or instability, while pain that occurs both at rest and with weight bearing suggests an inflammatory source such as infection or neurogenic pain.
4. If exposure during revision TKA is difficult and the patellar ligament may avulse from its insertion, the exposure can be extended proximally with a rectus snip or V-Y turndown, or distally with a long tibial tubercle osteotomy.
5. Cancellous bone grafts heal and revascularize more effectively than solid structural bone grafts and are most appropriate for contained cavitary defects.
6. Revision TKA for arthrofibrosis can be expected to result in modest gains in range of motion, but pain may not be improved.
7. If vascularity to the skin over the knee is compromised, range of motion should be restricted for a few days after surgery to minimize skin hypoxia.
8. Reconstruction of a disrupted extensor mechanism after TKA requires soft-tissue augmentation with autogenous or allograft tissue in addition to primary repair.
9. When using an extensor mechanism allograft for extensor mechanism disruption, the graft should be sutured in maximal tension with the knee in full extension.
10. If loss of peroneal function occurs after revision TKA, the knee should be positioned in flexion to minimize tension on the nerve.
Bibliography
Babis GC, Trousdale RT, Pagnano MW, Morrey BF: Poor outcomes of isolated tibial insert exchange and arthrolysis for the management of stiffness following total knee arthroplasty. J Bone Joint Surg Am2001;83:1534-1536.
Barrack RL, Nakamura SJ, Hopkins SG, Rosenzweig S: Early failure of cementless mobile-bearing total knee arthroplasty. J Arthroplasty 2004;19:101-106.
Berend ME, Ritter MA, Meding JB, et al: The Chetranjan Ranawat Award: Tibial component failure mechanisms in total knee arthroplasty. Clin Orthop Relat Res 2004;428:26-34.
Busfield BT, Huffman R, Nahai F, Hoffman W, Ries MD: Extended medial gastrocnemius rotational flap for treatment of chronic knee extensor mechanism deficiency in patients with and without total knee arthroplasty. Clin Orthop Relat Res 2004;428:190-197.
Chockalingam S, Scott G: The outcome of cemented vs. cementless fixation of a femoral component in total knee replacement (TKR) with the identification of radiological signs for the prediction of failure.Knee 2000;7:233-238.
Engh GA, Ammeen DJ: Bone loss with revision total knee arthroplasty: Defect classification and alternatives for reconstruction. Instr Course Lect 1999;48:167-175.
Kim J, Nelson CL, Lotke PA: Stiffness after total knee arthroplasty: Prevalence of the complication and outcomes of revision. J Bone Joint Surg Am 2004;86:1479-1484.
Leopold SS, Greidanus N, Paprosky WG, Berger RA, Rosenber AG: High rate of failure of allograft reconstruction of the extensor mechanism after total knee arthroplasty. J Bone Joint Surg Am1999;81:1574-1579.
Mitts K, Muldoon MP, Gladden M, Padgett DE: Instability after total knee arthroplasty with the Miller-Galante II total knee: 5- to 7-year follow-up. J Arthroplasty 2001;16:422-427.
Ries MD, Badalamente M: Arthrofibrosis after total knee arthroplasty. Clin Orthop Relat Res 2000;380:177-183.
Schmalzried TP, Callaghan JJ: Current concepts review: Wear in total hip and knee replacements. J Bone Joint Surg Am 1999;81:115-136.
Togawa D, Goldberg VM, Rimnac CM, Greenfield EM: Polyethylene and titanium particles induce osteolysis by similar, lymphocyte-independent, mechanisms. J Orthop Res 2005; 23:376-383.
Vessely MB, Whaley AL, Harmsen WS, Schleck CD, Berry DJ: The Chetranjan Ranawat Award: Long-term survivorship and failure modes of 1000 cemented condylar total knee arthroplasties. Clin Orthop Relat Res 2006;452:28-34.
Werner FW, Ayers DC, Maletsky LP, Rullkoetter PJ: The effect of valgus/varus malalignment on load distribution in total knee arthroplasty. J Biomech 2005;38:349-355.