Instability , the displacement of the articular components, is one of the most common causes of early prosthetic failure after total knee arthroplasty. Instability prompts revision arthroplasty on average 4 years after primary arthroplasty [15]. Asymmetric widening of the prosthetic joint space suggests ligamentous imbalance and varus–valgus instability [29]. Yercan et al. describe three categories of instability seen in total knee arthroplasty including flexion, extension, and global instability [30]. Extension instability can be symmetric or asymmetric with respect to the joint space. Symmetric instability is often the sequela of improper surgical technique such as excessive resection of the distal femur or proximal tibia. Failure to correct valgus or varus deformities or overcorrection of angular deformities results in asymmetric instability. Asymmetric instability is far more common than symmetric. Flexion instability results in an excessive joint space gap and is usually created by undersizing of the femoral component or an excessive tibial slope. Global instability results from a combination of both loose flexion and extension gaps. Causes are multiple, including implant migration, extension mechanism failure, and polyethylene wear that give way to loss of surrounding soft tissue integrity. Most patients with global instability require constrained total knee arthroplasty revisions [30]. Flexion instability in the anterior–posterior plane can result in acute posterior dislocation, which is more common in posterior-stabilized prostheses. Prevalence of dislocations ranges from 1 to 2% in the early posterior cruciate ligament-stabilized designs, though recent design improvements have decreased this rate to 0.15–0.5% [15]. While signs of instability can be seen at a higher rate on radiographs, instability occurs in less than 1–2% of patients after primary TKA [30].

Evaluation of TKA alignment is important because of the direct relationship between malalignment, loosening, and instability. Both implant alignment and bony alignment must be evaluated to distinguish ligamentous instability from implant malpositioning. This is generally done with weight-bearing radiographs. Anteroposterior unilateral weight-bearing radiographs are useful for determining polyethylene liner wear. Valgus and varus stress AP radiographs can help evaluate the integrity of the collateral ligaments and determine if any deformity can be manipulated and reduced. Lateral extension and flexion radiographs are useful in detecting tibial slope and posterior subluxation [15].

The mechanical axis should pass through the center or just medial to the center of the prosthetic knee with both components perpendicular to it. The femoral component should be within 4–11° of valgus, with 7° generally optimal [1, 14, 18, 31]. On the lateral view, the posterior flange of the femoral component should be parallel or nearly parallel to the long axis of the femur and the femoral component outline should match the outline of the original bone [10, 29]. Notching of the anterior femoral cortex can be seen when the femoral component is undersized, which predisposes to fracture. The posterior aspect of the anterior flange should be parallel to and flush with the anterior femoral cortex [29].

The tibial prosthesis should be aligned perpendicular to the tibial shaft on the AP view. Varus malalignment of the tibial component has been identified as a risk factor for prosthesis loosening [1]. On the lateral view, the position of the tibial component should be either central or posterior relative to the center of the tibial shaft. The plateau should be parallel to the ground or slope downward no more than 10° on the lateral view [10, 29]. Overhang of the tibial component can result in bursitis, especially anteriorly [29].

It has been reported that optimal TKA results are achieved when the joint line is altered 8 mm or less and the patellar height (as measured from the distal point on the femoral articular surface to the inferior pole) is 10–30 mm [10, 32]. The AP thickness of the patellar implant should not exceed the thickness of the original patella, as increased retinacular pressure may lead to pain and maltracking. Patellar tracking can be grossly assessed on tangential patellar views with the knee in 30–40° of flexion [29]. On this view, patellar tilt is assessed as the angle between a line along the anterior aspect of the femoral condyles and a line along the patellar component cement–bone interface.

Component malrotation can lead to rotational instability [22]. Berger and Rubash describe a method of evaluating component malrotation prior to revision surgery using CT. The rotation of the femoral component is evaluated using the posterior condylar angle, defined as the angle subtended by the posterior condylar line and the surgical epicondylar axis. The normal posterior condylar angle for men is 0.3° (+/− 1.2°) and 3.5° (+/− 1.2°) for women. The rotation of the tibial component is determined using the tibial tubercle orientation. This is defined as the angle between two lines: 1. a line drawn perpendicular to the horizontal posterior margin of the tibial tray that runs through the geometric center of the tibial tray and 2. a line drawn through the middle of the tibial tubercle that runs parallel to its axis. This is most easily calculated by creating a superimposed image of the tibial tray and tibial tubercle. The normal rotation value for the tibial component is 18° (+/− 2.6°) of internal rotation from the tip of the tibial tubercle. When femoral and tibial rotations were combined, patients without patellofemoral symptoms all had TKAs with mild degrees of combined external rotation (0–10°), while patients with patellofemoral problems all had TKAs with combined internal rotation. The degree of internal rotation correlated directly with the severity of patellofemoral complication [22].

One drawback in using CT for the assessment of component malrotation is the potential risk for inter and intraobserver variability [33, 34]. In a recent study be Servien et al. CT was used to assess for tibial component rotation in unicompartmental knee arthroplasty [35].

While some authors find patellofemoral problems the most common postoperative complications associated with TKA and the most common reason for revision arthroplasty , others suggest patellofemoral problems closely follow infection and aseptic loosening as cause for revision surgery [15]. Patellofemoral complications range from patellar fractures, extensor mechanism rupture, patellar component failure, instability/maltracking, and soft tissue impingement syndromes. Patellar tilt and patellar subluxation are commonly seen on tangential (sunrise) views. These findings are often due to a tight lateral retinaculum, though a search should still be made for radiographic clues indicating component malrotation, valgus alignment, or oversizing of either the femoral or tibial component in the AP dimension—all of which can also lead to patellar tilt, subluxation, and even dislocation (Fig. 4.2). As Berger and Rubash studied, excessive internal rotation can result from incorrect positioning of either the femoral or tibial component or both. Incidence of patellar instability after TKA can be up to 12%, ranging from 1 to 12% in one study [22]. Patellar tilt and subluxation also tend to result in more rapid polyethylene wear, which can lead to particle disease and even metallosis if the components are metal backed [29].

The polyethylene portion of the patellar component has been reported to come loose from its metal backing. The dense synovial linear opacities of metallosis may be apparent if this occurs [36]. The radiolucent polyethylene component often is displaced inferiorly into the region of Hoffa’s fat pad but may be difficult to identify on routine radiographs due to its similar density to soft tissue. While metal-backed patellar prostheses were first used in the 1980s, more recent designs are entirely made of polyethylene with several peripheral pegs for cement or uncemented fixation. These components have a relatively low incidence for loosening of less than 2%. If the patellar resurfacing component is displaced for a substantial amount of time, biological remodeling, also called stress contouring, of the retropatellar surface will occur in the form of eroding and morphological changes of the subchondral bone plate as it adapts to the trochlear shape [37]. Adequate visualization may require soft tissue radiographic techniques, CT, or arthrography [36] (see Fig. 4.3). Displacement of the metal backing and polyethylene together, which results from fracture of fixation pegs [10, 29], is easily identified. A displaced patellar component may result in abrasion and rupture of the quadriceps or patellar tendons [10].

Patellar stress fractures occur with some frequency [10], as patellar resurfacing results in a thinned, possibly devascularized patella combined with stress risers via the peg holes [36] (Fig. 4.4). Fractures can be vertical or transverse, but most are vertical without compromise of the extensor mechanism [15]. Patellar component fractures may also be seen. These occur almost exclusively in metal-backed prostheses [10]. Patellar fractures are ideally treated conservatively as surgical intervention can result in high complication rate and marginal outcomes. Fractures in conjunction with extensor mechanism ruptures or resurfacing component loosening usually require repair and surgical fixation [15].

Rupture of the quadriceps or patellar tendon results in abnormal position of the patella (low and high, respectively) and localized soft tissue swelling with obscuration of fat planes. A wavy or buckled appearance of the soft tissues in the region of the tendon is sometimes seen on radiographs. An abnormally low patella (patella baja or infera) can also occur with an intact quadriceps tendon after TKA, due to fibrosis and scar contracture in Hoffa’s fat pad. An abnormally high patella (patella alta) with an intact patellar tendon is much less likely [29].

Cross-sectional imaging with MRI or ultrasound is much more sensitive and specific for the detection of extensor mechanism tears and ruptures [15]. Dynamic evaluation with ultrasound is used at our institution in the detection of quadriceps tears in patients with TKA. The knee can be examined in full extension as well as in varying degrees of flexion. When there is suspicion for an extensor mechanism rupture on physical exam with an abnormal appearing patellar location on radiographs, discontinuity of the quadriceps or patella tendon is readily visible on ultrasound as there is minimal artifact to overcome from the metal prosthesis. MRI with metal artifact reduction techniques has also proved a reliable diagnostic modality for extensor mechanism ruptures.

Another potential complication of the patellofemoral extensor includes a soft tissue impingement syndrome called “patellar clunk” syndrome. In this scenario, a soft tissue fibrous nodule develops at the junction the posterior aspect of the quadriceps tendon and the proximal pole of the patella [20]. As the knee is extended from a fully flexed position, this nodule becomes entrapped within the intercondylar notch. Near the end of full extension, tension is placed on the fibrous nodule which causes it to “clunk out” of the intercondylar notch resulting in pain and sometimes a sense of instability. Some authors have found success in using MRI to demonstrate the soft tissue nodule at the junction of the patella and quadriceps tendon confirming the diagnosis. While possible causes for patellar clunk syndrome include surgical technique, patellar maltracking and prosthesis design and technical enhancements, such as deepening the femoral trochlea at the time of TKA, have dramatically reduced the incidence of this complication.

Ideally, a prosthetic joint component would carry stress and distribute it to the underlying bone in a manner identical to the original bone. However, the mechanical properties of the prosthetic components are different than the original bone, resulting in altered distribution of forces to underlying bone. Bone is formed and maintained along the lines of stress. Thus, bone resorption occurs in areas that no longer receive as much stress after joint replacement. This is called stress shielding . On radiographs, this is evident as rarefaction of trabeculae, or localized osteopenia. This must be differentiated from osteolysis, which causes focal complete destruction of bone. Progressive bone loss due to stress shielding is one of the primary causes of loosening and one of the limiting factors in the life span of a joint prosthesis. Stress shielding occurs in all knees in which the femoral component has an anterior femoral flange [37]. Stress shielding can also occur around the tibial tray, especially when there is a long-stem distal fixation as forces are diverted distally, away from the tibial plateau. It usually occurs within the first 2 years of the life of the prosthesis. Upon follow-up imaging, it is imperative to comparison with early postoperative radiographs to detect subtle progression of osteolysis and component loosening [20].

Polyethylene wear and particle-induced osteolysis remain a common cause for revision arthroplasty. Contributing factors to polyethylene wear are multitudinous, including increased patient’s weight and/or activity level, specific type of polyethylene composing the liner, configuration and alignment of the femoral condylar component, and irregularities in the surface of the femoral condylar component articulating with the polyethylene [39]. Delamination of the polyethylene generates intra-articular particulate debris, which may subsequently engender osteolysis. Wear can occur from both the articular side (topside wear) and between the metal tibial tray and polyethylene liner (backside wear) [20]. Wear should be suspected when radiographs show narrowing of prosthetic joint spaces on weight-bearing views. When wear is asymmetric, varus or valgus deformity or patellar tilt results. Polyethylene fragments may be shed into the joint. It is important to look for loose intra-articular, porous-coating beads on radiographs, because they can lead to an accelerated type of wear, called third-body wear . Annual weight-bearing films are recommended to detect subclinical wear in TKAs, especially for prostheses with metal backing [29]. Early detection may allow simple exchange of the polyethylene liner before irreversible damage to the metal tray occurs [40]. Mild liner wear often can be subtle and can be confounded by differences in patient positioning. Therefore careful evaluation with prior studies is very useful in detecting subclinical polyethylene liner wear.

Using ultrasound, it is possible to detect polyethylene wear directly by measuring the thickness of the polyethylene tibial tray [41]. The joint effusion and synovitis that can result from polyethylene wear are also detectable with ultrasound. The effusion appears completely black (hypoechoic), while synovitis is manifested as fronds or nodules of intermediate echogenicity projecting into the joint fluid. This is most readily visualized in the suprapatellar pouch [19]. It is also possible to directly evaluate the tibial tray with ultrasound, enabling detection of polyethylene wear and tray fractures [19].

Osteolysis is a general term that simply means destruction of bone . In the setting of joint replacement, the term is used specifically to denote bone destruction due to particulate debris, thus designated particle disease. Particles may be polyethylene, cement, or metal [38]. Debris of a critical size triggers an inflammatory reaction with macrophages and foreign body giant cells, which results in osteolysis. When severe, the bone loss from osteolysis can result in component loosening. Osteolysi s is one of the leading causes of revision arthroplasty [15, 39].

Osteolysis is manifested on radiographs and CT as focal periprosthetic areas of lucency due to loss of trabeculae (Fig. 4.5). Common anatomic regions include the femoral condyles near the collateral ligament attachments and about the periphery of components. The reduction in metal artifacts and the improved ability to reformat high-quality multiplanar images made possible by multidetector CT have resulted in CT becoming a valuable tool for the detection and quantification of osteolysis. Puri et al. showed helical CT with metal artifact minimization to be more sensitive than radiographs for identifying and quantifying osteolysis after total hip arthroplasty [42]. Work by Seitz et al. indicates that CT is similarly advantageous in the evaluation of osteolysis at the knee [8, 43]. On sonographic images, osteolysis can be appreciated as focal loss of the normal bright, hyperechoic line of cortical bone, with an underlying hypoechoic, cystlike erosion [19]. The MRI appearance of osteolysis has been described as focal periprosthetic intraosseous masses with low T1 signal and heterogeneous, predominantly low to intermediate T2 signal. With IV contrast, these masses show peripheral enhancement and some irregular internal enhancement [24]. Vessely and colleagues found that the extent of osteolysis was greater on MRI than on radiographs in nine of 11 patients. MRI demonstrated radiographically occult lesions in five of 11 [44]. Similar findings of radiographically occult lesions visible on MRI were also described by Mosher et al. [45].