Arthroplasty of the Knee

Varus-valgus constrained prostheses

The original constrained condylar knee (CCK) was developed from the posterior-substituting design by enlarging the central post of the tibial polyethylene insert, constraining it against the medial and lateral walls of a deepened central box of the femoral component ( Fig. 7.9 ). Varus-valgus stability is controlled by this mechanism with a small amount of varus-valgus toggle allowed. This type of prosthesis otherwise functions as a PS design and is used in patients with instability that might otherwise require a hinged prosthesis. It cannot be used for recurvatum deformity because it does not control hyperextension. Originally designed with cemented intramedullary stems on femoral and tibial components, the design evolved to include modular press-fit or cemented intramedullary stems on the tibial and femoral components.

The CCK design has been used extensively for revision arthroplasty when instability is present and for difficult primary arthroplasties in patients with extreme valgus deformity and medial collateral ligament insufficiency. Although no loosening was reported at 8 years in a group of 28 CCK knees implanted for severe valgus deformities in an older patient group (average age 73), the added constraint of the CCK design raises the concern of whether it might incur increased rates of loosening, particularly when used in younger patients. Progressive bone-cement radiolucencies have been reported in 16% of patients at an average of 44 months after arthroplasty with the total condylar prosthesis III, the precursor of the CCK, and nonprogressive radiolucent lines were found in 16% of 148 knees with CCK implants without stem extensions used for correction of significant deformities. Reported failure rates are low, ranging from no failures to 2.5% failures at 4-year follow-up. Most total knee systems include a variation of a varus-valgus constrained design. More recent reports have also shown similar results after primary TKA using a CCK type of implant. Maynard et al. reported 127 primary TKAs using a CCK type of implant and found that at an average 110-month follow-up the revision rate was 0.8%, with a 10-year survivorship of 97.6%. These reports seem to indicate that for complicated primary TKAs with balancing issues that cannot be resolved with standard implants, a CCK implant can be a valid option for consideration. To combat higher forces from being imparted to the tibial implant, a mobile-bearing option can be added to the tibial side of the implant (see next section).

Mobile-bearing prostheses

Mobile-bearing knee designs have seen an increase in popularity, and requests have been made to the United States Food and Drug Association (FDA) to down-classify these devices for clearance purposes. The meniscal-bearing version of the low contact stress (LCS) prosthesis developed by Buechel and others incorporated many of the features of the earlier Oxford knee. Individual polyethylene menisci articulate with the femoral component above and with a polished tibial baseplate below. The LCS design has additional dovetailed arcuate grooves on the tibial baseplate that control the anteroposterior course of the menisci. The femoral component has a decreasing radius of curvature posteriorly. This modification of the Oxford design decreases the posterior excursion of the menisci in flexion, helping to decrease the incidence of posterior extrusion of the menisci.

The LCS total knee system also includes a rotating platform design with congruent tibiofemoral geometry in extension similar to other current deep-dish designs; however, the tibial polyethylene is additionally free to rotate within the stem of the tibial baseplate. This design has had rare rotational dislocations of the tibial inserts because of inadequate flexion-extension gap balancing, but it has exhibited excellent longevity. Callaghan et al. reported a 100% prosthesis survival rate in 82 patients at a minimum of 9-year follow-up of the cemented rotating platform LCS design. In a later follow-up study, Callaghan et al. reported the status of 53 knees in 37 of these patients who were still living at a minimum follow-up of 15 years. None of the knees had required revision because of loosening, osteolysis, or wear; three knees had required reoperation (two for periprosthetic fractures and one for infection), but none of the components was revised as part of the reoperations. Buechel, one of the developers of the LCS design, reported a 98% 20-year survivorship with this design and a similar survivorship at 18 years with the cementless rotating platform design. A recent meta-analysis comparing outcomes with fixed-bearing and mobile-bearing TKA found no clinically significant differences in patient-specific or clinical outcome parameters. One reason these implant designs are not offered by many manufacturers is their designation as a class III device by the FDA. In vivo fluorokinematic studies have shown that the bearing rotation in the transverse plane may be nonphysiologic in some patients. A meta-analysis found that there was moderate to low-quality evidence that CR mobile-bearing TKA was as good as fixed-bearing TKA.

Potential advantages of mobile-bearing knees include lower contact stresses at the articulating surfaces, rotational motion of the tibial polyethylene during gait, and self-alignment of the tibial polyethylene compensating for small rotational malalignment of the tibial baseplate during implantation. Recent studies have found a higher revision rate in both the short-term and mid-term follow-up period with a mobile-bearing insert compared with fixed-bearing TKA. Whether mobile-bearing designs will outperform fixed-bearing designs has yet to be determined and may be specific to individual manufacturer design as well as proper surgical technique.

Unicompartmental prostheses

There are some reported results of unicompartmental knee arthroplasty (UKA) that are as good as those of TKA, but controversy remains about indications and the use of unicompartmental prostheses in patients with high body mass indices (BMIs); some reports show higher revision rates at 2-year follow-up in these patients. Many surgeons advocate the use of UKA for arthritis limited to only one knee compartment ( Fig. 7.10 ). These prostheses replace the articular surface of either the medial or the lateral femoral condyle and the adjacent tibial plateau surface. The current trend toward minimally invasive surgery has rekindled enthusiasm for these devices despite the fact that most studies, with some notable exceptions, have shown a slightly worse survivorship for UKA compared with TKA but is better than the revision rates for a revision TKA.

Marmor introduced a unicompartmental replacement in the early 1970s, obtaining better results with replacement of the lateral compartment than of the medial compartment. The Marmor prosthesis was anatomically shaped with a flat, all-polyethylene tibial component. Squire et al. observed 87.5% 15-year survivorship using this prosthesis. Subsequent unicompartmental prostheses with metal-backed tibial components and thin polyethylene occasionally exhibited rapid polyethylene wear.

Meniscal-bearing unicompartmental knee replacements allow translational motion at the polyethylene tibial baseplate interface similar to meniscal-bearing TKA and are enthusiastically supported by some authors, with reports of a 96% 10-year survivorship of the unicompartmental Oxford meniscal knee arthroplasty. In the Swedish National Registry, the Oxford unicompartmental knee fared worse, having a 7% revision rate by 6 years.

Sparing of the cruciate ligaments, the opposite tibiofemoral compartment, and the patellofemoral joint in UKA is purported to result in more normal knee kinematics and to allow easy revision to a tricompartmental prosthesis at a later time. A more normal knee possibly can be obtained with UKA with quicker rehabilitation time and greater range of motion than with TKA. The second purported advantage, bone stock preservation, is more controversial. Revision of UKA to tricompartmental prostheses, requiring special components, bone grafting, or cement with screw augmentation to fill osseous defects, was necessary in 76% of patients reported by Padgett, Stern, and Insall and in 45% reported by Barrett and Scott. McAuley, Engh, and Ammeen reported a 26% use of local autograft, whereas 21% required wedge augmentation. They commented that the UKA revisions to TKA were simpler than typical revision TKA. An emphasis on more limited tibial resection with many newer designs may lessen the incidence of significant bony defects at the time of revision.

Mobile-bearing medial compartment UKA designs have the advantage of a congruent femoral polyethylene articular surface that allows the medial compartment to roll back and maintain more physiologic motion. Indications typically include a functioning ACL and limited flexion contracture with a flexion arc of greater than 100 degrees. To ensure bearing stability, a gap-balancing approach typically is used. The results have been good, and recent reports other than those of the design surgeons have been good to excellent.

Hinged implants

The Kinematic Rotating Hinge (Howmedica, Rutherford, NJ) ( Fig. 7.11 ) has been a widely used linked, hinged knee replacement. Two polyethylene and cobalt chrome bearings allow flexion-extension and axial rotation. The rotating hinge type of implant offers constraint in the sagittal and coronal planes while allowing free rotation in the transverse plane to limit the transfer of forces to the implant-bone interfaces and allow substitution of soft-tissue constraint in the coronal plane because of insufficient collateral support. The implant provides a block to extension as well, which prevents recurvatum. This implant often is necessary in salvage revision TKA but can be used in a primary TKA when significant deformity and loss of soft-tissue support does not allow a stable knee to be obtained or a flexion gap is created that might “jump” a constrained condylar type of implant. An early report found outcomes with the Kinematic Rotating Hinge no better than those with the earlier GUEPAR prosthesis with respect to infection, loosening, and patellar complications. A more recent study of hinged prostheses found a much lower complication rate at 4-year follow-up of the S-ROM hinged prosthesis (Joint Medical Products/Johnson & Johnson, Stamford, CT). This type of prosthesis is used in patients with severe ligamentous insufficiency, severe flexion or extension gap mismatch, recurvatum deformity, neuromuscular disease, and limb salvage procedures. Use of a hinged implant for primary TKA should be reserved for patients with these problems because of the tendency for worse outcomes and more complications than with other types of implants.

Knee replacement systems

Different types of prostheses are necessary for varying amounts of arthritic involvement, deformity, laxity, and bone loss. Prostheses used range from unicompartmental designs for single-compartment disease with minimal deformity to hinged prostheses for severe deformity and or ligamentous deficiencies and for salvage procedures. Many surgeons advocate the use of PCL-retaining prostheses for mild deformity and PCL-substituting designs for more severe deformity while for many surgeons the choice is based on training and experience. Knee prosthesis manufacturers have developed newer systems that offer either PCL retention or PCL substitution through modular tibial polyethylene inserts and PCL-substituting and PCL-retaining femoral components that require similar bone cuts. These prostheses typically use shared operative instrumentation and allow an intraoperative change from PCL retention to PCL substitution or even a constrained condylar design. If balancing of the PCL is difficult, the arthroplasty can be converted to a PCL-substituting design with relative ease in most cases. Many prosthesis designs also include a tibial polyethylene component with significant dishing (or increased AP congruency or constraint) in the sagittal plane for optional use instead of the PS design when the PCL is incompetent. Modular stems and metal augments and constrained condylar components are typically available in most systems.

Many other factors are important in prosthesis design and selection, including prosthesis fixation, the handling of the patellofemoral articulation, modularity, and polyethylene issues. These are discussed in subsequent sections of this chapter. It is the surgeon’s responsibility to understand the indications, contraindications, expected functional outcome, and longevity for each prosthesis type and for specific prostheses. Every surgeon should be familiar with the options and instrumentation of his or her choice to ensure that all bases are covered in the operating room. Long-term follow-up studies will continue to improve our understanding of appropriate indications for the variety of available knee prostheses.

Functional anatomy and kinematics

Knee motion during normal gait has been studied by many investigators, who have found it to be much more complex than simple flexion and extension. Knee motion during gait occurs in all three anatomic planes around the long axis of the limb ( Fig. 7.12 ). Knee flexion, which occurs around a varying transverse axis ( Fig. 7.13 ), is a function of the articular geometry of the knee and the ligamentous restraints. Dennis et al. described the flexion axis as varying in a helical fashion in a normal knee, with an average of 2 mm of posterior translation of the medial femoral condyle on the tibia during flexion compared with 21 mm of translation of the lateral femoral condyle. This observation was acquired by dynamic fluoroscopy coupled with three-dimensional CT scans of the studied knees. The axis became more variable after sectioning of the anterior cruciate ligament, with an average 5 mm of medial condylar translation and 17 mm of lateral condylar posterior translation in flexion. This pattern of medially based pivoting of the knee explains the observed external rotation of the tibia on the femur during extension, known as the “screw-home mechanism,” and internal rotation of the tibia during knee flexion. The inability of many early knee prosthesis designs to accommodate these complex knee motions and their attendant stresses was an unforeseen shortcoming. Many current prosthesis designs attempt to reproduce normal knee kinematics closely, whereas others settle for an approximation of normal motion, placing other concerns, such as polyethylene contact stresses, ahead of accurate reproduction of knee kinematics.

The use of gait laboratories, biomechanical models, and fluoroscopic analyses to study normal subjects and patients before and after knee arthroplasty has become an important tool in prosthesis design and functional evaluation of TKA patients ( Fig. 7.14 ). In kinematic studies of the knee during selected activities of daily living (ADLs), normal gait required 67 degrees of flexion during the swing phase, 83 degrees for stair climbing, 90 degrees for descending stairs, and 93 degrees to rise from a chair. Computer modeling now can be used to predict the effects of prosthetic designs on motion and how they respond to malpositioning during surgery, aiding in the development of designs that are more forgiving and provide more physiologic motion and kinematics . Variation in implant positioning in the transverse plane has been investigated by Milhalko and Williams using a dynamic kinematic model, and significant variations in internal and external rotation during a simulated deep-knee bend have been described. This enforces the need for forgiveness in implant designs so that the knee is not constrained in any one way that may increase the implant-bone interface stresses.

Role of the posterior cruciate ligament in total knee arthroplasty

Since the concurrent development of PCL-retaining and PCL-substituting prostheses, the relative merits of each design have been debated. Each design boasts multiple series with comparable excellent 10- to 15-year results. Studies of bilateral TKA with a PCL-retaining prosthesis on one side and a PCL-substituting prosthesis on the other side have failed to show significant subjective performance or patient satisfaction differences. A closer look at the differences in these designs illustrates, however, many of the factors involved in successful arthroplasty.

PCL retention achieves an increased potential range of motion by effective femoral rollback. In vivo kinematic analysis has shown that with a CR design there may be a “roll forward” positioning of the medial femoral condyle on the polyethylene insert with flexion, which may limit flexion. The designs now compensate for this and have evolved from a relatively flat articulation to a higher constrained anterior aspect in most CR designs to prevent the medial femoral condyle from sliding forward in an excessive manner. PCL substitution achieves femoral rollback by a tibial post and femoral cam mechanism. Compared with the original total condylar design, both of the modern designs of PCL retaining and substituting attain greater flexion (see Fig. 7.3 ). In multiple studies comparing PCL-retaining and PCL-substituting prostheses, the average flexion attained at long-term follow-up has been similar. When the PCL is retained, it frequently needs to be partially released or recessed to allow adequate flexion, especially in the varus-deformed knee because it is a more medial anatomic structure and may be involved in the coronal plane deformity. More recently, deep-dish designs with increased sagittal plane conformity have been studied (with PCL recession and with PCL sacrifice). The flexion with these more conforming devices is similar to that with the PCL-retaining and PCL-substituting devices with which they have been compared.

In PCL-substituting designs, posterior displacement in flexion is produced by the tibial post contacting the femoral cam, with the resultant stress borne by the prosthetic construct and ultimately transferred to the bone-cement interface ( Fig. 7.15 ). Originally, this situation led many authors to suggest that PCL-substituting designs would have higher failure rates than PCL-retaining devices because of loosening. The loosening rates in most reported studies show equal survivorships of PS and CR TKA. A recent study from the Mayo Clinic, however, compared 5389 CR TKAs to 2728 PS TKAs and found 15-year survivorships of 90% and 77%, respectively, a statistically significant difference. Higher mid-term and late-term revision rates in PS TKA have also been reported in other studies. This finding appears to support the theory that higher transfer of stress to the implant interface of some PS designs may decrease their longevity.

Early gait analysis studies have found that individuals with PCL-retaining prostheses have a more symmetric gait, especially during stair climbing, than do individuals with either PCL-sacrificing or PCL-substituting designs. They showed decreased knee flexion during stair climbing and a tendency to lean forward in a quadriceps-sparing posture in patients with PCL-sacrificing and PCL-substituting designs. They postulated that these observations may indicate inadequate rollback of these designs or possibly the loss of a proprioceptive role of the PCL. These observations have been cited as reasons to retain the PCL. Later gait analysis contradicts the conclusions of these earlier studies, however, after comparing PCL-substituting knees with normal controls. These earlier observations are refuted further by in vivo studies using fluoroscopy during single-stance deep knee bends to show a paradoxic forward translation of the femorotibial contact point during weight-bearing flexion in some PCL-retaining knees, whereas PCL-substituting knees studied showed more uniform femoral rollback.

The patellofemoral joint functions with a larger extensor lever arm when femoral rollback, as a function of PCL retention or PCL substitution, moves the tibial tubercle more anteriorly. The patellofemoral joint also is affected by joint line elevation, the extent to which the new prosthetic joint line is raised relative to the native joint line. PCL-retaining designs do not tolerate much alteration in the level of the preoperative joint line while balancing the flexion and extension gaps, whereas PCL-substituting designs frequently balance with some mild elevation of the joint line in extension to aid in the balancing of the increased flexion gap that occurs when the PCL is sacrificed. The PCL functions as a secondary stabilizer in the coronal plane, and its release often necessitates less collateral ligament balancing to obtain a symmetric flexion and extension gap during surgery ( Video 7.1, Video 7.2 ). In a cadaver study, Mihalko and Krackow showed that release of the PCL may increase the flexion gap 4 to 6 mm while increasing the extension gap less than 2 mm, but it should be pointed out that this was in a cadaver study with nonarthritic specimens. Figgie et al. suggested that joint line elevation may alter patellofemoral mechanics and result in postoperative pain and subluxation.

PCL-substituting femoral components have a cutout for a cam mechanism that begins just below the trochlea of the patellofemoral joint. Additional bone is removed from the femur when PCL-substituting designs are used to accommodate this box-and-cam mechanism. Additionally, the degree of flexion at which the patella contacts this “box” varies among different PS designs. The patella and hypertrophic synovium on the undersurface of the quadriceps tendon can bind in this mechanism. This clinical entity, termed patellar clunk syndrome, is a potential complication of PCL-substituting designs. Many posterior stabilized implant designs now offer a longer trochlear groove to combat the build-up of synovium from a shorter trochlear-groove-to-box length, but this continues to be reported in the literature, albeit less frequently, despite this design change.

Many authors argue that it is difficult to balance a diseased or contracted PCL in the presence of a significant varus deformity in a reproducible fashion. Although intraoperative tests of PCL balance have been devised by advocates of PCL retention, other investigators have stated that it is difficult, even in a laboratory setting, to reproduce near-normal PCL strain and function in a PCL-retaining knee arthroplasty. To have near-normal strain, the PCL needs to be balanced to an accuracy of approximately 1 to 2 mm. A PCL that is too tight in flexion can limit the extent of flexion attained postoperatively and lead to excessive femoral rollback, which multiple retrieval studies have shown to accelerate posterior tibial polyethylene wear. Some authors have suggested that attaining reliable balance of the PCL requires experience and that surgeons who perform fewer than 20 TKAs a year should use PCL-substituting prostheses. Late rupture of the PCL is also thought to be a cause of late instability in PCL-retaining designs. It should be pointed out, however, that a posterior stabilized implant does not recreate the normal kinematics of the knee either, and there are proponents for both designs, which have had similar 15- to 20-year survivorships reported in the literature. The argument that the PCL in an osteoarthritic knee is involved in the osteoarthritic process is one used by proponents of PCL-sacrificing and PCL-substituting techniques. The mechanoreceptors in knees with osteoarthritis (OA) have been shown to be decreased but still present. PCLs retrieved at necropsy from CR TKAs ( Fig. 7.16 ) have been shown to be similar to those from osteoarthritic knees with PCL-sacrificing TKAs, suggesting that the mechanoreceptors are functioning and may contribute to proprioception after TKA.

Another argument in favor of PCL substitution is that significant deformity can be more reliably corrected with its use. Extensive collateral ligament release on the concave side of a fixed knee deformity may not be effective without release of the contracted PCL, which acts as a tether. Similarly, if the collateral ligament on the convex side of a deformity is significantly stretched or attenuated, opposite collateral ligament release is effective only in achieving varus-valgus balance to the extent that is allowed by the intact PCL. The tethering effect of the PCL on soft-tissue balancing of the varus or valgus knee also has been shown in cadaver studies. In a series of patients with preoperative fixed varus or valgus deformities of 15 degrees or more associated with flexion contractures, knees treated with PCL retention had less postoperative flexion, more severe residual flexion contractures, and less correction of the mechanical axis than knees with PCL substitution. In another large series of knees treated with PCL retention, however, no correlation was found between preoperative deformity and postoperative outcome.

Polyethylene wear is affected by prosthesis design and by its in vivo kinematics. The tibial articular surface of PCL-retaining prostheses is typically less conforming to the femoral component in the sagittal plane to allow femoral rollback. This less-conforming geometry in the sagittal plane is responsible for higher tibial polyethylene contact stresses in PCL-retaining prostheses ( Fig. 7.17 ). Several authors have suggested that these greater contact stresses are responsible in part for accelerated polyethylene wear. This wear can be compounded by an excessively tight PCL that may increase the polyethylene contact stress as it becomes tight in flexion. In the extreme, a PCL that is tight in flexion can cause the femoral condyles to override the posterior edge of the tibial polyethylene, causing extremely high polyethylene contact stresses. This mechanism of accelerated posterior wear has been proposed after study of retrieved polyethylene specimens by various authors, who expressed concern that paradoxic anterior tibial translation in flexion in a poorly functioning PCL-retaining knee may lead to early polyethylene wear. Conversely, the tibial post on many PCL-substituting designs has been shown to be a site of wear and occasional breakage, particularly when the femoral component can impinge on the post anteriorly in hyperextension. This condition is accentuated when the femoral component is implanted in a flexed position, when the tibial component is implanted with a greater posterior slope, and when the knee hyperextends. Other design features where the cam climbs the post in flexion and transfers load at the upper aspect of the post may also contribute to post wear and breakage (see Fig. 7.26D ).

Axial and rotational alignment of the knee

Numerous studies have shown a correlation between long-term success of TKA and restoration of near-normal limb alignment. Suboptimal alignment of total knee prostheses has been implicated in long-term difficulties, including tibiofemoral instability, patellofemoral instability, patellar fracture, stiffness, accelerated polyethylene wear, and implant loosening. The use of accurate instrumentation and an understanding of the basic principles inherent to the instruments are necessary to implant reproducibly well-aligned prostheses. Computer-assisted navigation and robotic techniques are now being used by some surgeons to try to improve the reproducibility of component alignment and to improve functional outcomes.

Normally, the anatomic axes of the femur and the tibia form a valgus angle of 6 ± 2 degrees. The mechanical axis of the lower limb is defined as the line drawn on a standing long-leg anteroposterior radiograph from the center of the femoral head to the center of the talar dome ( Fig. 7.18 ). This mechanical axis typically should project through the center of the knee joint, described as a “neutral” mechanical axis. When the mechanical axis lies to the lateral side of the knee center, the knee is in mechanical valgus alignment. In mechanical varus alignment, the mechanical axis of the limb lies to the medial side of the knee center. The amount of varus or valgus deformity can be determined on an anteroposterior long-standing radiograph by first drawing the mechanical axis of the femur, a line from the center of the femoral head to the center of the intercondylar notch, and extending this line distally. The mechanical axis of the tibia runs from the center of the tibial plateau to the center of the tibial plafond, discounting any bowing of the tibia. The angle formed between these separate mechanical axes of the femur and tibia determines the varus or valgus deviation from the neutral mechanical axis. By determining the tibial mechanical axis using the center of the tibial plateau and the femoral mechanical axis using the center of the intercondylar notch, any medial or lateral subluxation through the knee joint is disregarded. Insall argued that rotation affects the mechanical axis of the femur apparent on an anteroposterior radiograph, lessening the value of these preoperative and postoperative measurements.

In a normal knee, the tibial articular surface is in approximately 3 degrees of varus with respect to the mechanical axis and the femoral articular surface is in a corresponding 9 degrees of valgus. Multiple studies have shown that tibial components placed in more than 5 degrees of varus tend to fail by subsiding into more varus. Consequently, tibial components generally are implanted perpendicular to the mechanical axis of the tibia in the coronal plane, with varying amounts of posterior tilt in the sagittal plane, depending on the articular design of the component to be implanted. The femoral component usually is implanted in 5 to 7 degrees of valgus, the amount necessary to reestablish a neutral mechanical axis of the femur ( Fig. 7.19 ). Most implant systems offer various options, usually from 5 to 7 degrees of valgus for the distal femoral resection, but the proper angle can be calculated from the standing hip-to-knee radiograph by measuring the angle between the mechanical axis of the femur (line from the center of the femoral head to the center of the distal femur) and a line drawn from the entry point of the intramedullary rod to the diaphyseal line that the rod travels. Once the distal femoral cut is made, the use of an intramedullary goniometer to measure the actual cut distal femoral surface to the intramedullary femoral angle has been shown to be more accurate.

Recently, kinematic alignment of the lower extremity for TKA surgery has been described and advocated by many surgeons. Although this method has been used in the past, advocates point out that today we face less severe deformity correction than we did 20 years ago and that a kinematic alignment with a 3-degree varus joint line at the tibia and an increased distal valgus cut on the femur may improve functional results after TKA. Proponents also point to the ease of balancing the varus knee when using kinematic alignment. Opponents of the kinematic alignment method point to the fact that if outliers on the tibia reach 4 or more degrees varus, the rate of short-term and mid-term failures is much higher. We continue to recommend that the normal mechanical axis of the femur and the tibia be established and the joint line maintained parallel to the ground.

Rotational alignment of total knee components is difficult to discern radiographically, making the assessment of rotation primarily an intraoperative determination. The rotation of the femoral component has effects not only on balancing of the flexion space but also on patellofemoral tracking. Because the proximal tibial cut is made perpendicular to the mechanical axis of the limb instead of in the anatomically correct 3 degrees of varus, rotation of the femoral component also must be altered from its anatomic position to create a symmetric flexion space ( Fig. 7.20 ). To create this rectangular flexion space, with equal tension on the medial and lateral collateral ligaments, the femoral component is externally rotated an average of 3 degrees relative to the posterior condylar axis or perpendicular to the anteroposterior axis. In an average male femur, this technique rotationally places the femoral component with the posterior condylar surfaces parallel to the epicondylar axis. This technique fails when the posterior aspect of the native femoral condyle has significant wear, or when the lateral femoral condyle is hypoplastic, as is frequently seen in knees with valgus deformity. In these instances, the surgeon can use the epicondylar axis or the anteroposterior axis popularized by Whiteside (SEE Technique 7.1 for details). The epicondylar axis has been shown in multiple studies to be difficult to determine in vivo when comparing different observers and when comparing the measured axis with one determined by CT. Each of these techniques of determining femoral component rotation is based on the geometry of the femur primarily, with subsequent ligamentous releases to create symmetric flexion and extension gaps.

Knowledge of each of these techniques is necessary because arthritic deformity or previous surgery may obscure one or more of these landmarks. In revision TKA, the epicondylar axis usually is the only native landmark left to ensure proper femoral component rotation.

Two primary techniques are used to align the tibial component rotationally. The first technique aligns the center of the tibial tray with the junction of the medial third of the tibial tubercle with the lateral two thirds. The second technique places the knee through a range of motion with trial components in place, allowing the tibia to align with the flexion axis of the femur. This second technique tends to align the tibial component rotationally with the rotation of the femoral component, lowering the chance of a rotation mismatch that could lead to accelerated polyethylene wear, although combined internal rotation of both components may lead to patellofemoral maltracking, as shown by Berger et al., and a higher incidence of patellofemoral pain.

Proponents of rotating platform designs claim that the rotational freedom of the tibial polyethylene allows self-correction of minor malrotation of the tibial tray. Although this factor may improve the congruency of the tibiofemoral articulation, tracking of the patella may not be improved and may be related to the existing and proposed femoral rotational alignment.

Patellofemoral joint biomechanics and functional anatomy

The primary function of the patella is to optimize the lever arm of the extensor mechanism around the knee, improving the efficiency of quadriceps contraction. The quadriceps and patellar tendons insert anteriorly on the patella, with the thickness of the patella displacing their respective force vectors away from the center of rotation of the knee ( Fig. 7.21 ). This displacement or lengthening of the extensor lever arm changes throughout the arc of knee motion. The length of the lever arm varies as a function of the geometry of the trochlea, the varying patellofemoral contact areas, and the varying center of rotation of the knee. The extensor lever arm is greatest at 20 to 30 degrees of flexion, and the quadriceps force required for knee extension increases significantly in the last 20 degrees of extension as less of the patella is in contact with the trochlear groove.

Patellofemoral stability is maintained by a combination of the articular surface geometry and soft-tissue restraints. The Q angle is the angle between the extended anatomic axis of the femur and the line between the center of the patella and the tibial tubercle ( Fig. 7.22 ). The quadriceps acts primarily in line with the anatomic axis of the femur, with the exception of the vastus medialis obliquus, which acts to medialize the patella in terminal extension. Limbs with larger Q angles have a greater tendency for lateral patellar subluxation. Because the patella does not contact the trochlea until early flexion, lateral subluxation of the patella in this range is resisted primarily by the vastus medialis obliquus fibers. As the angle of flexion increases, the bony and subsequent prosthetic constraints play a dominant role in preventing subluxation. In most current femoral component designs, the lateral flange of the trochlea has been made more prominent, producing a more anatomic reconstruction. Many designs add a built-in trochlear groove angle of up to 7 degrees to enhance patellar mechanics and tracking ( Fig. 7.23 ). Trochlear enhancements and attention to femoral component rotation, reproduction of preoperative patellar thickness, and maintenance of joint line height have improved patellofemoral stability and have decreased the rate of lateral patellar retinacular release significantly. The application of these principles is discussed further in the section on surgical technique.

As a consequence of its role in transmitting the force of contraction of the quadriceps muscle to the patellar tendon around a variably flexed knee, the patella experiences a joint reaction force as the trochlea opposes its posterior displacement. This joint reaction force depends on the angle of knee flexion and the magnitude of the forces transmitted to the patella from the quadriceps and patellar tendons. During standing, the joint reaction force increases with increasing knee flexion as the force vectors of the quadriceps and patellar tendons become more parallel to the joint reaction force. Multiple investigators have calculated patellofemoral joint reaction forces of two to five times body weight during activities of daily living. However, during squatting with knee flexion up to 120 degrees, the joint reaction force may be seven to eight times body weight. These forces in a normal knee are resisted by thick articular cartilage, but they may exceed the yield strength of polyethylene, especially in the case of edge loading, which may lead to deformation of polyethylene patellar components over time.

Many authors have described variations in the area of contact between the patella and the trochlea during knee flexion ( Fig. 7.24 ). The inferior articular surface of the patella first contacts the trochlea in approximately 20 degrees of knee flexion. The midportion of the patella articulates with the trochlea in approximately 60 degrees of flexion, and the superior portion of the patella articulates at 90 degrees of flexion. In extreme flexion, beyond 120 degrees, the patella articulates only medially and laterally with the femoral condyles, and the quadriceps tendon articulates with the trochlea. A third articulating facet often is present on the medial aspect of the patella that articulates with the lateral aspect of the medial femoral condyle at more than 90 degrees of flexion.

As discussed in the earlier section on knee kinematics, the normal tibia internally rotates with respect to the femur during flexion with greater posterior translation of the lateral femoral contact point on the tibia relative to the medial femoral contact point. The net effect of this internal rotation of the tibia during flexion is to centralize the tibial tubercle in flexion or diminish the Q angle. These relationships may be altered in TKA with nonanatomic patellofemoral geometry, malrotation of the femoral and tibial components, elevation of the joint line relative to the tibial tubercle, and patella infera from patellar tendon contracture. Dennis et al. and Harman et al. noted that with multiple designs of TKA tested during fluorokinematic analyses, at least 19% had a reverse rotational pattern with deep knee bend. This indicates a need for less constraint during deep flexion to ensure that excessive implant-bone interface stresses are avoided and that the femoral component is supported in the deep flexed state without edge loading of the polyethylene insert (see Fig. 7.3D ).

Changes in the patellar area of contact with flexion have a significant effect on the prosthetic patellofemoral joint. Eccentric loading of the patellofemoral joint leads to shear forces within the patellar component and at the prosthesis-bone interface ( Fig. 7.25 ). Even if the mediolateral geometry of the patellofemoral articulation is perfectly conforming, the inferior-to-superior migration of the area of contact on the patella with increasing knee flexion leads to eccentric forces on the polyethylene patellar component. These forces may lead to failure of metal-backed patellar components, localized polyethylene wear, or component loosening.

Polyethylene and bearing choices

Ultrahigh-molecular-weight polyethylene articular surfaces have been an integral part of TKA from its conception. Catastrophic wear leading to early failure and osteolysis, although seen less frequently than in total hip arthroplasty, historically has occurred more frequently in some TKA designs. Studies of polyethylene have provided information on its varying wear characteristics after different fabrication and sterilization processes and its limitations in TKA applications. Several manufacturers have converted to utilizing either highly crosslinked polyethylene or polyethylene with vitamin E to improve long-term wear. Early results do seem to show a resistance to fatigue-related damage and wear.

Compared with the perfectly conforming articulations of total hip arthroplasties, the tibiofemoral articulations in modern TKA are typically less conforming, with the femoral condyles having a decreasing radius of curvature posteriorly. PCL-retaining prostheses tend to have an even greater degree of sagittal plane nonconformity because the tibial surface remains relatively flat to allow femoral rollback without excessive PCL tension. This nonconformity creates areas of high contact stress within the polyethylene that is design specific (see Fig. 7.17 ). Retrieval studies by various authors document polyethylene wear in areas of high contact stress. Wear is pronounced in areas of unusual stress caused by prosthesis malalignment or ligamentous imbalance. Several authors have emphasized “double dishing” as a tibial polyethylene geometry that appears to avoid areas of high contact stress. Conformity in the coronal and sagittal planes should be enough to allow stability without constraining the transverse plane or without causing edge loading ( Fig. 7.26 ). This type of geometry also aids in condylar lift-off with designs that do not include coronal plane dishing of the individual tibial plateaus and corresponding femoral condyles.

A thinner tibial polyethylene also has been correlated with accelerated wear. Several studies have recommended a minimal polyethylene thickness of 8 mm to avoid the higher contact stresses that occur with thinner polyethylene. Retrieval studies showing accelerated wear in knees implanted with thin polyethylene have supported this recommendation. An average 10-year follow-up study showed no difference, however, in radiographic loosening, wear, or osteolysis in patients with a thin, one-piece, compression-molded, metal-backed polyethylene component (polyethylene thickness 4.4 mm) in one knee compared with a similar, although thicker (minimal thickness 6.4 mm), tibial component in the contralateral knee. Other reports have shown that survivorship decreases with less polyethylene thickness in some designs.

Retrieval data suggest that variations in polyethylene quality are partly responsible for reports of accelerated wear in the past. Landy and Walker found delamination only in areas of polyethylene that contained granular fusion defects. Whether ram extrusion with subsequent machining or direct compression molding of the components is the optimal polyethylene manufacturing process is debatable. Various manufacturers are attempting to improve the wear characteristics of polyethylene either by reprocessing bar stock purchased commercially to achieve a higher degree of uniformity within the polymer or by compression molding their own implants in an inert gas environment. Newer indications for TKA revision surgery have now been reported by Sharkey et al., and wear and osteolysis as reported more than a decade ago have fallen from the top of the list.

Polyethylene “enhancements” do not always improve the survivorship of their associated total knee components. In the mid-1980s, carbon fiber–reinforced polyethylene was introduced with the hope of improving wear characteristics of standard ultrahigh-molecular-weight polyethylene ( Fig. 7.27 ). This polyethylene was available for a brief period before it was withdrawn from the market because of accelerated and catastrophic wear. A dark carbon staining of the synovium can be seen at revision arthroplasty in knees with carbon fiber–reinforced polyethylene. Another unsuccessful polyethylene modification was the process of heat-pressing the prosthetic articular surface after the insert had been milled in an attempt to create a very smooth articular surface. This process led to a physical transition zone 1 mm beneath the articular surface of the polyethylene, which is in a region of high subsurface stress concentration. This coincidence in some early porous-coated anatomic knees, along with an articular geometry that was characterized by high contact stresses, led to a high rate of failure because of polyethylene delamination ( Fig. 7.28 ), particularly with thin polyethylene.

The method of polyethylene implant sterilization can affect polyethylene properties, with evidence that gamma radiation in an oxygen environment causes detrimental effects that can hasten polyethylene wear. Over a period of years, a subsurface white band appears in polyethylene sterilized in this manner ( Fig. 7.29 ). This occurs even in prostheses that have not been implanted and represents an area of high oxidation and chain scission within the polyethylene. McGovern et al. reported a failure rate of 49% at 18 months after surgery in a series of UKAs sterilized by gamma radiation in an air environment and stored preoperatively for 4.4 years or more before implantation. They found an inverse relationship between the shelf life of the tibial components (after sterilization but before implantation) and the time to revision surgery. Removed components showed high degrees of oxidation, wear, and fragmentation. Alternative approaches to prevent accelerated oxidation include radiation sterilization and packaging in an inert gas environment and sterilization by ethylene oxide or gas plasma.

The introduction of highly crosslinked polyethylene produced by high-dose gamma irradiation with subsequent annealing has produced dramatic decreases in wear in simulated hip and knee studies. Many authors caution, however, that the wear mechanisms in the knee are different from the mechanisms in the hip, and that highly crosslinked polyethylene may not be beneficial in TKA applications, especially when a PS implant is used with a post-cam bearing surface that is under higher shear than the tibiofemoral articulation. Some manufacturers have introduced the use of vitamin E–stabilized polyethylene in their knee replacement lines to correct the possible issue related to highly crosslinked polyethylene. Although mechanical studies have shown the benefits of vitamin E polyethylene with decreased wear properties in vitro, it remains to be seen whether this will translate into improved wear in vivo in the long term. The fact that an updated publication has indicated that wear and osteolysis have dropped from the top of the list may mean that we are on the right track with updated polyethylene formulas and sterilization techniques.

Some surgeons advocate alternative bearing surfaces in primary TKA. One type of bearing surface available is oxidized zirconium on the femoral components as a means to reduce polyethylene wear. This technology incorporates a zirconium oxide ceramic coating on a zirconium metal alloy femoral component. Developers claim that this surface is more scratch resistant than cobalt chrome, lessening wear debris from the polyethylene tibial articular surface. Ezzet et al. showed a 42% wear volume reduction in a knee simulator at 5 million cycles using this device. Other alternative bearing surfaces on the femoral side of the knee are now being introduced to decrease polyethylene wear. These include solid ceramic femoral components (still under FDA investigational trials) and zirconium or titanium nitride types of coatings; however, not all zirconium or titanium nitride coatings are the same since monolayer types of coating can be susceptible to mechanical ablation, and the surface can peel off the substrate metal layer through an eggshell effect wherein the underlying substrate is much softer than the thin coating. Using a multilayer coating as one implant manufacturer does with a top layer of zirconium nitride can combat this eggshell effect (Aesculap Inc.). Other manufacturers utilize a transformed layer of the surface alloy as with some titanium nitride or oxidized zirconium options.

The use of a modular metal-alloy baseplate backing for tibial polyethylene inserts became standard in the early 1980s. Multiple studies stated theoretical advantages of metal backing, including more even distribution of weight-bearing stresses to the underlying fixation interface and cancellous bone and a reduction in the potential polyethylene deformity caused by creep. A multisurgeon, multiprosthesis study of 9200 knee arthroplasties showed a 98% 5-year survivorship of knees with metal backing of the tibial polyethylene compared with 94% survivorship of knees with all-polyethylene tibial components. A Hospital for Special Surgery study reported a 97% 7-year survival of all-polyethylene, PS tibial components and a 99% 7-year survival of metal-backed components. In a more recent review of the literature concerning cruciate condylar TKAs, there were no significant differences, however, in prosthesis survival or periprosthetic lucencies between metal-backed tibial polyethylene and all-polyethylene tibial components. Similar survival rates of 98% have been reported for all-polyethylene and metal-backed tibial components in patients older than 80 years old at the time of surgery, suggesting that one-piece, all-polyethylene, cemented tibial components of sufficient thickness may be appropriate for use in low-demand, elderly patients.

Component fixation

Prosthetic fixation in TKA with polymethyl methacrylate (PMMA) has consistently shown long-term durability. Cementless fixation with bone ingrowth has been less reliable in long-term studies, however, with a few notable exceptions. The more successful cementless TKA designs typically have multiple attributes to attain baseplate stability while ingrowth occurs ( Figs. 7.30 and 7.31 ).

Retrieval analysis of some failed cementless implants has shown little, if any, bony ingrowth into tibial trays removed at the time of component revision. The bony ingrowth that did occur tended to be centered around fixation screws. Other reports of bony ingrowth have been more favorable. In 13 Miller-Galante prostheses removed for reasons other than loosening or infection, the average area of bone ingrowth was 27% of the available porous surface. They found a propensity for bone ingrowth in the region of fixation screws and pegs and in the anterior half of the tray. They postulated that the area of bony ingrowth necessary for stable fixation may be significantly less than 100%.

Clinically, many of the early cementless TKA systems had poor survival rates because of associated failure of metal-backed patellar components. Even discounting this factor, however, a 72% 10-year survivorship was reported with a cementless press-fit condylar design (Johnson & Johnson, Raynham, MA) compared with a 94% 10-year survivorship with similar cemented TKA. These results are being improved on as newer designs are now showing similar survivorship in the short term. A recent study on a medial pivot CR-designed implant in 54 patients showed 100% survival at 9 years in both cemented and cementless baseplates. A recent review of the literature also concluded that newer designs have had excellent results but cautioned that longer follow-up studies were necessary.

There are some notable exceptions in earlier press-fit designs with excellent long-term success. The Ortholoc (Wright Medical, Arlington, TN) prosthesis was reported to have one loosening in 184 knees followed for a minimum of 15 years, and the cementless LCS rotating platform knee was reported to have a survivorship of 98% at 18 years. Because of the importance of adequate initial fixation of cementless implants, stems with keels and multiple screws have been incorporated into the design of modern cementless tibial baseplates (see Figs. 7.30 and 7.31 ). A purported advantage of a biologic interface over cemented fixation is its durability, reported to result in better long-term pain relief; however, excellent long-term durability of cement fixation has been reported. Despite claims that cementless fixation may be more durable over time, most authors believe cemented fixation has produced more uniformly reliable long-term fixation with less osteolysis in multiple prosthesis designs. This belief is supported by the Swedish National Registry data, which showed that cementless designs had a 1.4 times higher rate of revision than did cemented designs.

Modular stems that engage the diaphyseal bone on the femur and the tibia are used for revision when there is bone loss and when components with varus-valgus constraint are used. These stems typically are press-fit in cementless reconstructions and either cemented or press-fit when the articular portion of the prosthesis is cemented. Because of deformity and anatomic variations, the stems occasionally need to be offset to achieve alignment.

Newer three-dimensional printed ingrowth surfaces are being used, and there has been a resurgence in popularity of cementless fixation for primary knee replacement. With the success of newer designs and ingrowth surfaces noted in primary total hip replacement, adopting these into modern knee implants is a logical next step as biologic interface should become more robust over a longer time period.

Total knee arthroplasty

The primary indication for TKA is to relieve pain caused by severe arthritis, with or without significant deformity. Other sources of knee and leg pain must be sought and systematically excluded. These include radicular pain from spinal disease, referred pain from the ipsilateral hip pathology, peripheral vascular disease, meniscal pathology, and bursitis of the knee. Radiographic findings must correlate with a clear clinical impression of knee arthritis. Before surgery is considered, conservative treatment measures should be exhausted, including physical therapy, antiinflammatory medications, intraarticular injections, activity modifications, and the use of a cane for ambulation. Patients who do not have complete cartilage space loss before surgery tend to be less satisfied with their clinical result after TKA.

Because knee replacement has a finite expected survival that is adversely affected by activity level, it generally is indicated in older patients with more sedentary lifestyles. It also is clearly indicated in younger patients who have a significant functional impairment from OA or from other pathologic causes such as systemic arthritis with multiple joint involvement or osteonecrosis with subchondral collapse of a femoral condyle. Severe pain from chondrocalcinosis and pseudogout in an elderly patient is an occasional indication for arthroplasty in the absence of complete cartilage space loss. Occasionally, severe patellofemoral arthritis in an elderly patient may justify TKA because the expected outcome of arthroplasty is better than that of patellectomy or patellofemoral replacement in these patients.

Deformity can become the principal indication for arthroplasty in patients with moderate or severe arthritis and variable levels of pain when the progression of deformity begins to threaten the expected outcome of an anticipated arthroplasty. As a flexion contracture progresses beyond 20 degrees, gait is significantly hampered and difficulty with regaining extension may warrant surgical intervention. Similarly, as varus or valgus laxity becomes severe, a constrained condylar type of prosthesis may become necessary to prevent subsequent coronal plane instability. Intervening before this degree of laxity is present allows the use of a prosthesis that lacks coronal plane constraint and has a more favorable expected survivorship.

Absolute contraindications to TKA include recent or current knee sepsis; a remote source of ongoing infection; extensor mechanism discontinuity or severe dysfunction; recurvatum deformity secondary to neuromuscular weakness; and the presence of a painless, well-functioning knee arthrodesis. Relative contraindications are numerous and debatable and include medical conditions that compromise the patient’s ability to withstand anesthesia, the metabolic demands of surgery and wound healing, immunodeficiency, and the significant rehabilitation necessary to ensure a favorable functional outcome. A severely osteoarthritic ipsilateral hip joint also should be considered for arthroplasty before the symptomatic osteoarthritic knee, because rehabilitation is easier with a total hip arthroplasty and an osteoarthritic knee than with a TKA and an osteoarthritic hip joint. Other relative contraindications include significant atherosclerotic disease of the operative leg, skin conditions such as psoriasis within the operative field, venous stasis disease with recurrent cellulitis, neuropathic arthropathy, superobesity (BMI ≥ 45), recurrent urinary tract infections, and a history of osteomyelitis in the proximity of the knee. This list is not all inclusive, and any preoperative condition that can adversely affect the patient’s outcome can be considered a relative contraindication.

Outcome studies have now shown that patient optimization is key to ensuring the best chance of a good long-term outcome. Certain modifiable risk factors should be considered before elective TKA, including low vitamin D levels, metabolic syndrome (MetS), low albumin, neutropenia, super-obesity, and a BMI less than 20. The current AAOS Clinical Practice Guidelines on Knee Osteoarthritis Treatment states that delaying primary TKA for up to 8 months while a patient works to improve a modifiable risk factor does not appear to worsen the outcome. A recent review noted an increasing rise in infection and other complications as obesity classification increased from severe to super-obese. Although these issues need to be weighted by the surgeon, quality of life in super-obese individuals can be attained cost effectively and without a higher incidence of early aseptic loosening.

Unicondylar knee arthroplasty

UKA is being selected for increasing numbers of patients, particularly with minimally invasive techniques that allow overnight hospital stays or outpatient procedures. Ten-year follow-up studies of two designs of UKA, the Oxford mobile bearing knee (Biomet Orthopaedics, Warsaw, IN) and the Miller-Galante knee (Zimmer, Warsaw, IN), have shown survivorship approaching that of TKA. Newer techniques are now available, including robotic-arm–assisted surgical procedures. These procedures involve preoperative computed tomographic studies for determining appropriate component sizing and accurate positioning of the implants. Many of the studies of UKA were conducted in older patients, and many authors doubted that the survivorship of UKA into the second decade would parallel the survivorship of TKA. A long-term survivorship analysis of over 500 medial compartment Oxford meniscal-bearing unicompartmental arthroplasties, however, found a 10-year survivorship of 94% and a 20-year survivorship of 91%, indicating that the implant remains durable into the second decade.

UKA currently is advocated for different reasons in two patient populations. The first patient group for whom UKA has been advocated comprises elderly, thin individuals with unicompartmental disease who would otherwise undergo TKA. The suggested benefits of UKA over TKA are a shorter rehabilitation time; a greater average postoperative range of motion; and preservation of the proprioceptive function of the cruciate ligaments, which give a more natural-feeling knee. The procedure can be done with a shorter hospital stay and with less blood loss. The argument in this patient group is that a UKA is a less invasive procedure that has a good likelihood of lasting the patient’s lifetime. Berger et al. reported a 10-year survivorship of 98% with a cemented UKA design in older patients using stringent selection criteria. UKA should not be considered in an elderly patient who has evidence of arthritis in more than one compartment of the knee unless there are medical contraindications to TKA.

The second group currently considered for UKA comprises younger individuals with unicompartmental disease in whom UKA is used as a “first” arthroplasty, usually instead of high tibial osteotomy (HTO) in patients with isolated medial compartment arthritis. This indication is becoming more prevalent with the increasing popularity of minimally invasive surgery and the increasing demands of patients and higher BMIs of the general population. Although this is a frequently stated indication for UKA, few studies that have been published to date have reported results in this patient group. One study described an 11-year survivorship of 92% in patients younger than 60 years old, with another 22% showing progression of the unresurfaced compartment, although not requiring revision at the time of follow-up.

There has been continued discussion concerning whether the patellofemoral compartment should be a determining factor in the decision to perform a UKA. One recent study pointed out that patellofemoral congruence is improved after UKA and that this may be the reason for reports of good outcomes after UKA done for patellofemoral osteoarthritis (OA). The type of bearing used in medial compartment arthroplasty also has been debated, with many advocating the use of a mobile-bearing UKA. A recent study found that the difficulty of revision surgery is comparable, regardless of the bearing used. In this series mobile-bearing UKA required more medial augments for conversion to primary TKA.

It has been suggested that UKA is a bone-sparing operation that would allow an uncomplicated revision later, but earlier studies of failed UKA did not show this anticipated benefit, with significant bone grafting, tibial wedges, or long-stem components necessary in nearly half of revisions and major osseous defects in 76% of knees. With more contemporary UKAs, the need for structural grafts is rare and results of revision approach those of primary TKA. In a matched retrospective review, the results of revision of a failed UKA to TKA were slightly worse than conversion of a previous HTO to TKA. The selection of UKA or HTO in this patient population remains unclear because many studies have cited difficulties with exposure and slightly less satisfactory clinical outcomes with TKA after previous HTO compared with primary TKA.

Another argument favoring TKA over UKA is the unfamiliarity of many surgeons with UKA. According to Stern, Becker, and Insall, only 6% of patients needing arthroplasty have none of the contraindications to UKA. Because the success of the procedure is dictated by the technical performance of the operation, surgeons who rarely perform UKA may have difficulty reproducing the reported results from large reconstructive centers. Gioe and Bowman described an 89% 10-year survivorship for UKA performed in a community hospital setting compared with a 95% survivorship of TKA done in the same time period . A recent review of patient-reported outcomes after both UKA and TKA from a large European registry found no significant differences between the two groups; however, some registries have reported higher short-term and mid-term revision rates with UKA. Another report pointed out that superior outcomes typically have been reported by high-volume centers and surgeons, which raises concerns about poorer outcomes with low-volume surgeons and centers. A recent systematic review also pointed out slightly better results in patients with UKA but confirmed higher revision rates. Another review reported no detrimental effects of preexisting patellofemoral disease on 10-year outcomes.

Although the certain indications for UKA are debatable, the contraindications are fairly well defined: inflammatory arthritis, a flexion contracture of 15 degrees or more, a preoperative arc of motion of less than 90 degrees, angular deformity of more than 10 degrees from the mechanical axis for varus knees or 5 degrees for valgus knees, significant cartilaginous erosion in the weight-bearing areas of the opposite compartment, and anterior cruciate ligament deficiency. Obesity also has been cited as a relative contraindication to UKA.

Patellofemoral arthroplasty

Although historically controversial, new interest in patellofemoral arthroplasty over the past few years has been fueled by contemporary implant designs that have produced improved clinical outcomes. First-generation designs failed because of narrow trochlear grooves and high constraint, which often produced maltracking, patellar catching, or persistent anterior knee pain. Despite improvements in the current designs, the most common reason for failure of the second-generation implants is progression of tibiofemoral arthritis, making careful patient selection the key to a successful outcome. The use of custom-designed implants to allow the least amount of bone loss has been reported. These implants, although more expensive than the off-the-shelf sizing that most manufacturers offer, are being reported to have excellent outcomes and short-term to mid-term survivorship.

The ideal candidate for patellofemoral arthroplasty is a patient who is younger than 65 years of age and has debilitating, isolated patellofemoral arthritis with minimal coronal deformity; pain during daily activities is localized to the patellofemoral joint and has not responded to nonsteroidal antiinflammatory medications or injection. Patellofemoral arthroplasty is recommended for isolated patellofemoral OA to provide a conservative, bone-sparing alternative to TKA, which may not have as good patient satisfaction in young, active patients. A recent report indicated that conversion of a patellofemoral arthroplasty is comparable to performing a primary TKA. Parratte et al. compared 21 patellofemoral arthroplasty conversions to TKA with 21 primary and revision TKAs. Although there were more complications in the patellofemoral arthroplasty conversion group, this was not comparable to the revision cohort. To date there are no published studies showing that outcomes of patellofemoral arthroplasty are age dependent. All of these considerations should be taken into account before offering a patellofemoral arthroplasty over a TKA.

Good results have been reported after patellofemoral arthroplasty in patients with posttraumatic arthritis, primary patellofemoral OA, and patellofemoral dysplasia without malalignment. In patients with posttraumatic arthritis, patellofemoral arthroplasty may be considered as an alternative to patellectomy. Primary patellofemoral arthritis includes Outerbridge type IV chondromalacia of the patella or trochlea or both. Note that progression of tibiofemoral arthritis is more frequent with primary OA than with posttraumatic arthritis or patellofemoral dysplasia. Malalignment is most often determined using the quadriceps angle (Q angle). Angles of more than 15 degrees in men and 20 degrees in women are considered abnormal. Any condition that increases the Q angle increases the lateral displacement forces on the patella and may lead to subluxation or dislocation. Patellofemoral arthroplasty alone cannot correct patellar malalignment and/or instability. Malalignment and / or instability of the patellofemoral joint is not an indication for the procedure. Mild patellar tilt or subluxation can be corrected at the time of patellofemoral arthroplasty with lateral retinacular release, medialization of the patellar component, and possibly partial lateral facetectomy. Malalignment should be corrected before or during patellofemoral arthroplasty. No particular patellar or trochlear wear pattern has been determined to be a contraindication to patellofemoral arthroplasty, unlike various tubercle osteotomy procedures. Lateral and inferior patellar facet lesions in younger patients can be treated with anterior medialization of the tibial tubercle. A recent analysis showed that patients with trochlear dysplasia tend to have an internally rotated placement of the trochlear groove and that patellofemoral arthroplasty can compensate for this pathologic alignment.

Progression of tibiofemoral arthritis is the most common reason for revision to TKA, emphasizing that tibiofemoral arthrosis is a principal contraindication to patellofemoral arthroplasty. Inflammatory arthropathies involve the entire joint and currently are a contraindication to patellofemoral arthroplasty because of progressive tibiofemoral arthritis and painful synovitis. This includes chondrocalcinosis, which can be indicative of an inflammatory arthropathy and can lead to altered joint mechanics because of abnormal menisci.

Patellofemoral arthroplasty also is not indicated in patients with severe coronal deformity of the knee (valgus of more than 8 degrees or varus of more than 5 degrees) unless the deformity is corrected by osteotomy before arthroplasty. Flexion of 120 degrees in the sagittal plane, with less than 10 degrees of flexion contracture, is recommended as long as the flexion contracture is not caused by OA in the medial or lateral compartment of the knee. Knee joint stiffness should be carefully assessed because this patient population has a high rate of previous surgery that increases the frequency of arthrofibrosis and patellar height abnormalities. Patients with patella baja from quadriceps muscle atrophy or patellar tendon scarring are not good candidates for patellofemoral arthroplasty. Although few data exist correlating the outcome of patellofemoral arthroplasty with BMI, currently it is not recommended in obese patients because of concerns about overloading of the implant. A recent study showed a higher rate of revision to TKA in obese patients (BMI > 30) than in nonobese patients, whereas primary diagnosis, age, or sex did not significantly affect the revision rate.

Reported results of patellofemoral arthroplasty indicate that it provides excellent pain relief and functional improvement and is a reliable alternative to TKA for the treatment of patellofemoral arthritis in carefully selected patients ( Fig. 7.32 ). Good-to-excellent 3- to 17-year results have been reported in 66% to 100% of patients ( Table 7.1 ); less mean blood loss, shorter hospital stays, and better functional outcomes have been reported in patients with patellofemoral arthroplasty compared to those in patients with TKA. Recent studies, however, have found that revision rates after conversion of a patellofemoral arthroplasty to a TKA are higher than revision rates for a primary TKA. These findings should be considered when making decisions concerning the possible effectiveness of patellofemoral replacement (see Table 7.1 ).

Indications and considerations for patellar resurfacing in primary total knee arthroplasty

The role of universal patellar resurfacing in TKA is somewhat controversial, with some advocating it because of clinical series indicating that knee scores after patellar resurfacing are slightly better because of less residual peripatellar pain and improved quadriceps strength. In a large retrospective study, patellofemoral complications occurred in 4% of patients with patellar resurfacing compared with 12% of patients in whom the patella was unresurfaced. Significant residual anterior knee pain was the most common complication in the unresurfaced group. A 5-year prospective, randomized study of a single knee design found that 25% of patients with unresurfaced patellas complained of anterior knee pain, whereas only 5% of patients with patellar resurfacing complained of anterior knee pain. Secondary resurfacing of the patella for residual anterior knee pain after TKA has been studied by various authors who found that pain relief after secondary resurfacing was inferior to what would be expected with primary resurfacing and found a higher rate of complications, including patellar fracture and postoperative stiffness.

Other authors have advocated selective resurfacing of the patella. The major argument in favor of selective resurfacing of the patella is that complications of resurfaced patellae account for most of the reoperations after TKA in many series. Also, with selective resurfacing of the patella, using a femoral component that incorporates an anatomically shaped femoral trochlea, essentially equal knee scores have been reported for resurfaced and unresurfaced groups. Prospective studies comparing TKA with and without patellar resurfacing have found no significant differences in patient preferences, functional scores, anterior knee pain, or revision rates. However, these reports have found that those with an unresurfaced patella who have secondary procedures for resurfacing are not always satisfied after secondary resurfacing. This has led some to suggest that anterior knee pain after TKA is related more to component design and proper alignment in the transverse plane than to patellar retention or resurfacing. A recent report from the Australian registry on selective resurfacing found a higher rate of revision compared to those who routinely resurface the patella.

The desirability of resurfacing continues to be debated, and the results of selective patellar resurfacing seem to be dependent on the design of the trochlear groove, with a native patella articulating within an anatomic trochlear groove giving results similar to those of TKA with resurfacing of the patella. Suggested indications for leaving the patella unresurfaced are a primary diagnosis of OA, satisfactory patellar cartilage with no eburnated bone, congruent patellofemoral tracking, a normal anatomic patellar shape, and no evidence of crystalline or inflammatory arthropathy. A report using a similar implant with a resurfaced patella on one knee and an unresurfaced patella on the contralateral side found that patients could not discern a difference.

Patient weight also seems to be a factor, with lighter patients tending to do well with unresurfaced patellae. This may be one factor in the trend to routinely leave the native patella seen in the European literature. Some have suggested that resurfacing the patella in “super-obese” patients (BMI ≥ 50) may overload the sesamoid bone and be a generator of pain, but no clinical series has been published showing a difference in pain scores in this patient population and lighter-weight patients. A report of the Australian registry indicated that PS TKA designs with unresurfaced patellae had higher revision rates after surgery than those with resurfaced patellae. They found that the minimally constrained (CR) designs with resurfaced patellae had the lowest rate of revision followed by PS designs with resurfaced patellae. In addition, onlay resurfacing fared better than inlay patellar designed buttons. Another report determined that resurfacing was cost effective based on the number of revisions reported after unresurfaced TKA.

Indications and considerations for simultaneous bilateral total knee arthroplasty

Numerous studies in the literature have documented the safety and cost effectiveness of simultaneous bilateral TKA compared with separate staged procedures. With respect to cost, simultaneous bilateral procedures can reduce hospital charges by 58% compared with staged procedures because of overall decreases in operative time and total length of hospital stay. Lane et al. questioned, however, if this is a true savings because 89% of their patients with bilateral TKA required an additional rehabilitation hospital stay, whereas only 45% of their patients with unilateral TKA required rehabilitation hospital stays. Other outcomes, as measured by infection rate, knee scores, and radiographic criteria, have been similar between the two groups.

Controversy continues regarding the relative incidences of complications in simultaneous and staged procedures. Various studies have shown total blood loss to be equal in the two groups, whereas others have shown significantly more blood loss with simultaneous procedures. A greater degree of postoperative thrombocytopenia the second day after surgery and more frequent deep vein thrombosis (DVT) and pulmonary embolism (PE) also have been reported after simultaneous procedures, but many other authors reported similar or lower rates of DVT and PE after simultaneous bilateral TKA than after staged procedures.

Fat embolism is a risk of TKA when intramedullary stems or alignment devices are used, and the risk of clinically significant fat embolism syndrome probably is increased with simultaneous bilateral TKA. Dorr et al. found a 12% prevalence of fat embolism syndrome with simultaneous bilateral TKA, as documented by neurologic changes with hypoxemia. Other authors found no differences in the occurrence of clinically significant fat embolism between the two groups. Venting of the intramedullary canal with fluted intramedullary alignment rods and a slightly enlarged entrance hole for intramedullary alignment rod insertion have been recommended to decrease the risk of fat embolism syndrome.

In considering patients for simultaneous bilateral TKA, comorbidities and physiologic age should be considered because significant cardiopulmonary disease may sway the surgeon toward unilateral procedures. An increased risk of cardiovascular and neurologic complications has been noted in patients older than 70 years undergoing simultaneous bilateral TKA. No increased risks of complications with bilateral TKA have been identified in patients with a BMI of 30 or more compared with those with a lower BMI. An analysis of over 4 million hospital discharges over a 14-year period compared unilateral, bilateral, and revision TKA procedures and found that bilateral TKA had higher complication and mortality rates than either unilateral or revision TKA. Before choosing staged or simultaneous TKA procedures, each patient should be carefully evaluated, considering his or her age, cardiac risk factors, and other comorbidities. The risks associated with both approaches should be thoroughly discussed with the patient before a choice is made ( Video 7.3 ). Newer reports have been mixed concerning whether bilateral versus staged TKA is cost effective, and it appears that this should not currently have a bearing on the decision-making process. Other reports have shown that bilateral primary TKA surgery is as safe as simultaneous TKA in the right patient population.

Considerations for outpatient knee joint arthroplasty

Many centers have now begun to offer outpatient surgery for both TKA and UKA. To safely accomplish this, a complete team setup is required, including office staff, operating room personnel, anesthesia, physical therapy, and ancillary providers. The use of tranexamic acid to reduce the need for blood transfusion also has helped provide a safer path to ambulatory surgery joint replacement. Pain modalities, including intraarticular injections with liposomal encased or plain bupivacaine with or without the addition of Toradol, morphine, and dexamethasone also have helped provide adequate pain relief in the immediate postoperative period to allow centers to successfully and safely perform knee joint arthroplasty procedures in the ambulatory or 23-hour setting (see section on pain management modalities). A visit to a center that has a successful program is helpful before implementing a short-stay/ambulatory arthroplasty program. Having a program that educates patients on the process so that they and their immediate family or caregivers understand what to expect can be beneficial. Using a family member as a “joint coach” has also been shown to be beneficial, with education being a key to success and patient satisfaction.

Functional and radiographic outcome measures

Over the past 3 decades the most popular knee rating systems have been those of the Hospital for Special Surgery and the Knee Society. The Knee Society released a knee rating system in 1989 and updated it in 2011 ( Box 7.1 ). Because of increased patient demands and expectations over the past 2 decades, this latest update has been tailored to incorporate patient-specific activities and patient-perceived expectations. The updated system now consists of preoperative and postoperative objective measurements recorded by the surgeon and patient-driven measures evaluated by patients concerning their perceptions of the most important and deleterious aspects of their knee arthritis and replacement surgery.

The first parts of the score include patient demographics and the patient’s Charnley functional score. The objective measures and knee score (out of 100 points depending on range of motion measures) include alignment and instability, which account for up to 50 points. The patient’s range of motion is considered by giving one point for each 5 degrees of total measured arc of motion, with deductions taken for flexion contracture and extension lag. The next part of the score takes into account patient-perceived measures including symptoms, satisfaction concerning pain and function during daily activities (40 points), and expectations after TKA concerning pain and daily and recreational activities (15 points). Functional activities (100 points) are assessed by the patient and include walking and standing (30 points), standard everyday activities (30 points), and advanced activities (25 points). The section on discretionary activities (15 points) allows the patient to pick three of his or her most important activities from a list and rate the level of difficulty he or she perceives in performing these activities. In this newest version of the rating system, the Knee Society has placed more importance on patient perceptions, possibly because patient- and surgeon-perceived outcomes have been reported to be significantly different. The preoperative and postoperative questionnaires are the same to allow direct comparison. Any surgeon can apply for a license through the Knee Society to use the Knee Society Scoring System.

Other activity-related scoring systems have been developed and validated. The Lower Extremity Activity Score (LEAS) was developed as a simple way to allow patients to report their highest level of possible activity before and after surgery, choosing from a list of activities that progress in the level of functional capacity. The LEAS was validated using the Western Ontario and McMasters Universities Osteoarthritis Index (WOMAC) and comparison to responses to pedometer readings from patients. The scale also has been shown to be accurate when filled out by next of kin, making it a unique measure of functional activity.

In 1989, the Knee Society introduced the Total Knee Arthroplasty Radiographic Evaluation and Scoring System ( Fig. 7.33 ) to standardize the radiographic parameters to be measured when reporting radiographic outcomes of TKA: component alignment, tibial surface coverage, radiolucencies, and a patellar problem list that includes angle of the prosthesis, eccentric component placement, subluxation, and dislocation. A score is tabulated for each component based on the width and extent of its associated radiolucencies. For a seven-zone tibial component, a nonprogressive score of 4 or less probably is insignificant, a score of 5 to 9 indicates a need for close follow-up for progression, and a score of 10 or more signifies possible or impending failure regardless of symptoms. Developers of total knee prostheses are requested to superimpose silhouettes of their designs on the Knee Society form and assign radiographic zones to be used by all authors in subsequent reports.

The Knee Injury and Osteoarthritis Outcome Score for Joint Replacement (KOOS JR) has gained popularity. It is a one-page questionnaire, with a seven-item instrument, that is easily administered. Since the questions represent “knee health” in patients, it can be used to determine pain, symptom severity, and activities of daily living (ADL). Movements or activities that are difficult for patients with advanced knee osteoarthritis before surgery are then compared to after TKA.

Prosthesis survival

Modern knee arthroplasty began in the early 1970s with the development of the total condylar knee prosthesis. Survivorship studies with this prosthesis are the standard with which modern knee replacement is compared. Long-term series have documented the longevity of the original total condylar prosthesis to be 95% at 15 years and 91% at 21 and 23 years. More recently, the reported 15- to 18-year survivorship of a cementless CR TKA was 98.6%, with 79% of patients reporting no pain.

Multiple studies of PCL-retaining and PCL-substituting designs have documented 10-year survivorship of 95% or greater, and most registry data agree with this figure. As discussed in the earlier section on component fixation, cementless fixation has had mixed results with respect to prosthesis survivorship in the past, but more modern, next-generation ingrowth surface technologies are now demonstrating similar mid-term results to cemented TKA. An update on why TKA fails was recently given by Sharkey et al. Of the 781 revisions performed at their institution over a 10-year period, the most common failure mechanisms were loosening (39.9%), infection (27.4%), instability (7.5%), periprosthetic fracture (4.7%), and arthrofibrosis (4.5%). Infection was the most common reason for failure in early revision (<2 years from primary surgery), and aseptic loosening was the most common reason in late revision. They reported that polyethylene wear was no longer the major cause of failure, which can be attributed to better forms of polyethylene and better designs. When they compared these results to their previous report, the percentage of revisions performed for polyethylene wear, instability, arthrofibrosis, malalignment, and extensor mechanism deficiency had all decreased.

Ambulatory and short-stay considerations

Many centers have now turned to outpatient surgery for both UKA and TKA. To safely accomplish this endeavor, one must have a complete team setup, including office staff, operating room personnel, anesthesiology, physical therapy, and ancillary providers. The use of tranexamic acid to reduce the need for blood transfusion also has helped to provide a safer path to ambulatory surgery for joint replacement. Pain modalities, including intraarticular injections with liposomal-encased or plain bupivacaine, with or without the addition of ketorolac, morphine, or dexamethasone, help to provide adequate pain relief in the immediate postoperative period (see section on Pain Management Strategies below). These measures allow centers to successfully and safely perform knee joint arthroplasty procedures in an ambulatory or a 23-hour setting. We established a 24/7 total joint hotline for patients with issues to call so that they feel more at ease and have access to a provider. Having a program that educates patients on the process so that they and their immediate families or caregivers understand what to expect can be beneficial. Before implementing an outpatient TKA it is helpful to first visit a center that has a successful program in place.

Anesthetic options

The selection of regional or general anesthesia for TKA is a complicated issue that is affected by comorbid medical conditions. The anesthesiologist has the ultimate responsibility for this selection, with input from the surgeon. Cardiovascular outcomes of regional and general anesthesia have not been proved to be significantly different, and perioperative mortality in patients with hip fractures is the same with both techniques. Cognitive function after surgery has been shown to be similar with regional and general anesthesia after the initial postoperative period. Most studies, as well as the AAOS Clinical Practice Guidelines, now advocate regional over general anesthesia for primary TKA because of better pain control and outcomes. A spinal anesthetic is preferred because it affords better pain control and has lower complications after primary TKA.

The effect of general versus epidural anesthesia on thromboembolic complications is controversial. A slight, but not statistically significant, decrease in overall DVT and PE rates has been reported in patients who have had epidural anesthesia compared with general anesthesia, whereas another randomized trial showed no difference in overall thromboembolic disease but did show a decrease in proximal thrombus formation with epidural anesthesia. Possible benefits of epidural anesthesia include vasodilation of the lower extremity, resulting in increased blood flow, hemodilution, and decreased blood viscosity. A fibrinolytic effect also has been postulated for epidural anesthesia; however, in a study comparing epidural and general anesthesia there was no difference in intraoperatively obtained blood markers for fibrinolysis or thrombogenesis.

Pain management strategies after partial and total knee arthroplasty

Many different pain management modalities have been used to alleviate pain after total or partial knee arthroplasty. Most surgeons advocate a multimodal approach that includes a preoperative dose of a COX-2 antiinflammatory and gabapentin, which has been shown to be beneficial in patients with chronic pain who have TKA. The use of femoral nerve catheters is now under scrutiny because they can inhibit postoperative mobilization, especially on and after the day of surgery. Intraarticular injections that infiltrate the surrounding soft tissues with either a bupivacaine or ropivacaine (lower cardiotoxicity) product or that are placed in the intraarticular space have been studied, as has the use of liposomal-encased bupivacaine. Use of one nonnarcotic oral preoperative medication with either a femoral nerve block or intraarticular injection can give excellent pain relief. These modalities have been compared in prospective studies, with several studies finding no significant difference in visual analog scale (VAS) scores or narcotic usage after surgery. The use of liposomal-encased bupivacaine has come under some scrutiny because of its expense and lack of improvement in some reports over plain bupivacaine injections combined with epinephrine, with or without addition of ketorolac and dexamethasone. The PILLAR study, however, did show an advantage in postoperative pain control after primary TKA using 20 mL of liposomal bupivacaine with 20 mL of 0.5% plain bupivacaine and 80 mL of normal saline in a targeted infiltration technique.

A recent publication has reported that up to 40% of patients remain chronic users of narcotics after primary TKA; therefore, surgeons should make all efforts to wean patients off narcotics after surgery. Cryoneurolysis of the anterior femoral cutaneous and infrapatellar branch of the saphenous nerve before TKA surgery has been shown to decrease the amount of daily morphine equivalents needed after TKA, and this may be a beneficial treatment in high-risk patients or those who are under chronic pain management.

Blood preservation management in partial and total knee arthroplasty

The use of tranexamic acid either intravenously, topically before closure, or orally has been shown to significantly decrease postoperative hemoglobin drop and the need for postoperative transfusions after primary TKA. All of the delivery modes have been shown to be safe with no increased risk of thromboembolic events after surgery in the proper patient population. Advocates for intravenous administration argue that there is no wait to close the operative approach, whereas advocates of topical administration believe it may be a safer route of administration. For intravenous administration, the dose should be 10 to 15 mg/kg or 1 g, with consideration of a preoperative dose given 20 minutes before tourniquet inflation and a repeat dose given about 15 minutes before tourniquet deflation. For topically administered dosing, 1.5 to 3 gdiluted in 100 mL normal saline should be placed in the wound and intracapsular space for 5 minutes before tourniquet deflation. Some surgeons who still use drains administer the dose through the drain and then activate the closed suction after 5 minutes. Contraindications to intravenous tranexamic acid use include a history of a clotting disorder, bleeding disorder, subarachnoid hemorrhage, pulmonary embolus, DVT, cardiovascular accident, or myocardial infarction, as well as the presence of a coronary stent. Oral tranexamic acid and has been shown to be effective. Doses of 2000 mg given 2 hours before surgery and then two doses after surgery (4 to 6 hours and 12 hours postoperatively) have been reported to be effective and equivalent to intravenous or topical dosing. It is also more cost effective.

Intramedullary and extramedullary alignment instrumentation

Intramedullary alignment instrumentation is crucial on the femoral side of a TKA because femoral landmarks are not easily palpable. The entry portal for the femoral alignment rod typically is placed a few millimeters medial to the midline, at a point anterior to the origin of the PCL. Preoperative radiographs should be scrutinized for a wide canal or excessive femoral bowing because these conditions may result in alignment errors. Cadaver studies have shown that positioning of the entry point of the femoral intramedullary alignment rod significantly affects the resulting alignment of the distal femoral cut by as much as 5 degrees in the sagittal plane.

Extramedullary femoral alignment is useful only in limbs with severe lateral femoral bowing, femoral malunion, or stenosis from a previous fracture, or when an ipsilateral total hip arthroplasty or other hardware fills the isthmus of the intramedullary canal. A palpable marker can be placed over the center of the femoral head based on preoperative hip radiographs or by fluoroscopic imaging with the patient on the operating table. The anterior superior iliac spine has been shown to be unreliable for determining the hip center and should not be used as the primary landmark when extramedullary femoral alignment is chosen. Currently, when preexisting deformity or hardware is present, more modern techniques such as computer navigation or custom cutting blocks are used (see Computer-Assisted Alignment Technique).

The use of tibial intramedullary alignment guides is slightly more controversial. One concern about the use of these guides is the risk of fat embolism. A greater elevation of pulmonary arterial pressures and slightly diminished cardiac indices were found in patients undergoing bilateral TKAs using intramedullary tibial alignment guides compared with extramedullary tibial alignment guides and venting of the femoral intramedullary canals; however, these slight changes were not believed to constitute any contraindication to the use of intramedullary alignment devices. Because of a 12% prevalence of postoperative neurologic changes believed to be consistent with fat embolism after bilateral TKA, the use of pulmonary arterial pressure monitoring has been recommended by some surgeons. Drilling an oversized 12.7-mm hole in the distal femur and using an 8-mm fluted rod were shown to eliminate the negative cardiopulmonary effects of intramedullary femoral alignment rods.

The relative accuracy of intramedullary and extramedullary tibial alignment also has been debated. In one study, 94% of tibial components were within 2 degrees of 90 degrees with intramedullary alignment compared with 85% with extramedullary alignment. Another study found extramedullary alignment to be more accurate, with 88% of tibial components within 2 degrees of the 90-degree goal, whereas only 72% of the components placed using intramedullary alignment met this criterion. Neutral tibial component alignment was reported to be obtained with intramedullary devices in 83% of varus knees, but in only 37% of valgus knees; tibial bowing was more common in valgus knees ( Fig. 7.40A ), and the use of long leg films for templating and intraoperatively double checking the alignment of the tibial cut with an extramedullary device were recommended ( Fig. 7.40B ). In a cadaver study comparing intramedullary and extramedullary devices to computer navigation for tibial alignment, intramedullary alignment was not as accurate as extramedullary alignment in determining posterior slope. The anterior tibial crest also has been used by some surgeons during extramedullary alignment of the tibia. This technique has been shown in a cadaver study to be variable, resulting in a range from 3.2 degrees varus to 2.1 degrees valgus with a consistent anterior slope in the sagittal plane. Currently, most surgeons at our institution use intramedullary femoral alignment with extramedullary tibial alignment; one uses computer navigation techniques as well. When determining where the center of the ankle lies between the malleoli, some now advocate a slightly medial-based position from the center of the malleolar axis in the coronal plane. This has been reported to be approximately 2 mm medial to this center point, so many surgeons aim the tip of the tibial extramedullary alignment rod slightly medial to the center of the ankle, and the amount of tibial torsion and subcutaneous tissue present can affect where the radiographic center of the ankle lies. Assessment of a long-standing radiograph can help determine where the radiographic center of the tibia lies on an individual basis and should be reviewed before every case.