Total Knee Arthroplasty - part 4 pptx

III Kinematics 18 Understanding and Interpreting In Vivo Kinematic Studies – 115 S.A Banks 19 The Importance of the ACL for the Function of the Knee: Relevance to Future Developments in Total Knee Arthroplasty – 121 A M Chaudhari, C O Dyrby, T P Andriacchi 20 Kinematics of Mobile Bearing Total Knee Arthroplasty – 126 D A Dennis, R D Komistek 21 Cruciate Deficiency in the Replaced Knee – 141 J Victor 22 Kinematic Characteristics of the Unicompartmental Knee J.-N Argenson, R D Komistek, D A Dennis 23 In Vitro Kinematics of the Replaced Knee S Incavo, B Beynnon, K Coughlin 24 The Virtual Knee – 159 B W McKinnon, J K Otto, S McGuan – 152 – 148 115 Chapter 18 · Understanding and Interpreting In Vivo Kinematic Studies – S.A Banks 18 18 18 Understanding and Interpreting In Vivo Kinematic Studies S A Banks Summary Radiographic imaging and shape-matching techniques have been used since the late 1980s to quantify the motions of knee replacements in vivo These studies have shown how knee implants move in vivo, how implant design affects knee kinematics, and how different surgical and design factors influence knee mechanics and patient function In general, knee implants that definitively control the anteroposterior position of the femur with respect to the tibia achieve greater weight-bearing flexion and exhibit kinematics that are more likely to result in better patient function and implant longevity Three-dimensional Kinematics from Two-dimensional Images By the late 1980s, total knee arthroplasty (TKA) had become a fairly routine procedure for the treatment of severe knee arthritis.A wide variety of implant designs were being utilized with predictable success and reasonable durability The focus of designers was shifting from basic knee function and implant fixation to improving knee performance and implant longevity In part, what was needed to continue evolving knee replacements was more precise information on how knee replacements moved once implanted Unfortunately, the gait laboratories and CT scanners of the day were not able to provide accurate three-dimensional (3D) kinematic information about knee replacement motion during weight-bearing dynamic activities In 1988 I was given the mandate to develop a better method for measuring knee kinematics for my doctoral dissertation Having failed to use the gait laboratory motion capture system to accurately measure implant motion, W Andrew Hodge suggested we should directly image the joint with X-ray fluoroscopy and develop an image-based motion measurement technique This “shape-matching” approach proved to work well [1, 2], and this technique and its evolved forms have been used since to provide a better understanding of knee replacement function The details of shape-matching-based motion measurement are beyond the scope of this volume,but the process follows several logical steps: Radiographic images are produced when X-rays pass through space and are attenuated by the patient’s anatomy before striking a sensitive medium to cause a chemical or electrical reaction.The Xray beam emanates from a point source, with rays diverging in all directions, creating a central or perspective projection of the object - in essence a shadow (⊡ Fig 18-1) The location of the X-ray source with respect to the image plane can be measured so that the same optics can be reproduced on a computer Computer-aided design (CAD) information is available for the knee implant components, and bone surfaces can be reconstructed from CT or MR (⊡ Fig 18-2), making it a simple process to synthesize on the computer images of implants at any possi- Fluo ros c opi c Im age X-ray spot ⊡ Fig 18-1 Fluoroscopic and radiographic projections are created by a spot source of rays so that the image is a “perspective”projection, or shadow, that is a three-dimensional function of the projection geometry and the position and orientation of the bones This geometry allows three-dimensional kinematics to be derived from sequences of two-dimensional radiographic images 116 III Kinematics ⊡ Fig 18-2 Three-dimensional measurement of dynamic knee motion using fluoroscopy and shape-matching techniques has been performed for natural knees (left), knees with partial arthroplasty (middle), and knees with total arthroplasty (right) The bone surface models can be created from CT and MR scans and the implant models are obtained from the manufacturer or 3D laser scans ble position These synthetic views can be iteratively modified until they match the views obtained from patients Once matched, the positions and orientations of the models represent the physical position and orientation of the patient’s implants that created the radiographic projection Many groups the world over have used shape-matching techniques for determining implant motion from single-plane radiographic views, studying a range of activities including gait [3], stair-climbing [4], and deep knee bends [5].Although the details of the methods vary, measurement precision for each moving segment is typically 0.5-1.0 mm for implant motions parallel to the image plane and 0.5°-1.0° for rotations Importantly, this is monocular vision,not stereo or binocular,and all of these techniques have much reduced accuracy for translations perpendicular to the image plane, where precisions are typically 3.0-6.0 mm.If these measurement errors are extended to the articular surfaces, one can typically expect measurement uncertainties of greater than 1.2 mm for single observations of condylar contact or separation tramedullary rods or extramedullary techniques seek alignment orthogonal to the distal femur The anterior bow of the femoral shaft results in the femoral implant component being flexed forward in the sagittal plane by 5°-7° Similarly, tibial implant techniques range from alignment perpendicular to the long axis of the tibia to an alignment matching the normal posterior slope of the tibial plateau The net result of typical surgical placement is that the implants are in 5°-12° of relative hyperextension.Simultaneous measures of skeletal flexion,using goniometry or motion capture (⊡ Fig 18-3), and of implant flexion using fluoroscopy have shown an average 9.5° of implant hyperextension compared with the skeletal flexion angle [6] There are at least three important ramifications of this simple and intuitive observation First, implants that have hyperextension stops will likely experience much greater contact and possible wear than the designers anticipated [6, 7] Posterior stabilized designs with tibial Positional Findings 18 Findings from image-based TKA studies can be organized into positional and dynamic observations The positional observations relate closely to how implant design and surgical alignment influence articular contact and knee function at the extreme ranges of motion Knee implants typically are designed to maximize tibiofemoral contact area with the knee in extension and to accommodate 10°-15° of hyperextension Implant wear testing is performed such that the implants reach 0° relative flexion at simulated early stance Yet neither context takes account of the fact that surgical alignment may place the implants in positions which differ from 0° relative flexion Femoral components implanted using in- ⊡ Fig 18-3 Knees with well-aligned implants commonly show implant hyperextension Anterior bow of the femur and posterior slope of the tibial plateau bias implant alignment by an average of 10° hyperextension Thus, when the knee is fully extended at toe-off during gait (left), the implants are in hyperextension (right) 117 Chapter 18 · Understanding and Interpreting In Vivo Kinematic Studies – S.A Banks posts and some PCL-retaining designs accommodate limited hyperextension, often 5°-15° With the implants routinely placed in almost 10° of hyperextension at 0° of knee flexion, many of these designs will experience anterior impingement during routine activity Second, standard evaluations of TKA designs, whether by computer or by machine,do not account for implant alignment.The evaluations assume that the straight leg corresponds to 0° of implant flexion Given that many designs have surfaces with changing curvatures in early flexion, it is possible that these tests will predict performance differing from the clinical experience Third, implant features designed to guide implant motions at particular flexion angles will engage later in the flexion arc Post and cam mechanisms in posterior stabilized knees will engage at approximately 10° greater anatomical flexion than anticipated by the design In very deep flexion, there is some concern that the proximal “edge” of the femoral condyles (where the articular and bone-cut surfaces meet) will dig into the tibial articular surface Normal implant alignment means this phenomenon will occur 10° later in the flexion arc, if at all Fluoroscopic evaluations have elucidated the mechanics of total knee arthroplasties in deeply flexed postures It has long been assumed that greater posterior femoral translation on the tibia permits greater knee flexion [8] In a study of 16 different TKA designs in patients with excellent clinical outcomes, a significant linear relationship was seen between the amount of posterior femoral translation and maximum weight-bearing flexion [9] This relationship, 1.4° greater flexion for each additional millimeter of posterior femoral translation, held true for all types of TKA design (⊡ Fig 18-4) Implant designs that definitively controlled tibiofemoral position in flexion achieved greater femoral “roll-back”, and demon- ⊡ Fig 18-5 Correlation of restoration of posterior condylar offset (postoperative minus preoperative) with postoperative flexion gain(+)/loss(-) for 150 consecutive knees Overlapping points are not shown (Reprinted from [7]) strated greater weight-bearing flexion than designs that required the soft tissue and muscles to control tibiofemoral position These findings suggest that the flexion space,particularly in PCL-retaining TKA,ought not to be made too loose, as additional laxity may allow unwanted anterior translation of the femur and a concomitant decrease in maximum weight-bearing flexion Similar analyses have shown the importance of posterior condylar geometry on knee flexion range Bellemans et al [10] showed a significant linear relationship between changes in the posterior condylar offset, the maximum AP distance from the femoral shaft to the most posterior point on the condyles, and changes in the passive ROM They found that reducing the posterior condylar offset by mm from its anatomical value decreased the passive ROM by 6° (⊡ Fig 18-5) This finding is particularly relevant for surgeons using anterior referencing instrumentation: When a knee measures in-between component sizes, common practice argues for selecting the smaller component This will typically reduce the anatomical posterior condylar offset by several millimeters,potentially reducing the flexion range by 10° or more! Using the larger femoral component, when possible, or adjusting the position of the smaller femoral component can reduce the effect on the posterior condylar offset and provide the patient with the best possible range of motion Dynamic Characteristics ⊡ Fig 18-4 Maximum weight-bearing knee flexion as a function of femoral AP position for 121 knees Femoral posterior positions are negative, anterior is positive, and zero represents the AP midpoint of the tibial component Circles represent posterior-stabilized knees, asterisks represent posterior cruciate-retaining fixed bearing knees, and triangles represent the mobile-bearing knees The solid line shows the linear regression with a slope of 1.4° more flexion per millimeter femoral posterior translation (R=0.64, p90), without substantial ligamentous laxity or pain, and the ACL was considered to be functionally present in all cases Under fluoroscopic surveillance, each subject was asked to perform successive weight-bearing deep knee bend maneuvers up to maximum flexion, while the knee kinematic patterns were assessed at full extension and at 15°,30°,45°,60°,75°,and 90° of knee flexion.Likewise,each subject was asked to perform the normal stance phase of gait, while the kinematics were analyzed at heel-strike, at 33% and 66% of stance phase, and at toe-off The contact position between the medial femoral condyle (medial UKA) or the lateral femoral condyle (lateral UKA) and the tibia was determined using a threedimensional (3D) model-fitting technique [10] The 3D computer-aided design (CAD) solid models of the femoral and tibial components were overlaid and fit onto the two-dimensional (2D) fluoroscopic perspective view images (⊡ Fig 22-1) 22 149 Chapter 22 · Kinematic Characteristics of the Unicompartmental Knee – J N Argenson et al ⊡ Fig 22-1 Example of the 3D model-fitting process showing a fluoroscopic image (far left), 3D overlay (center left), pure sagittal view (center right), and top view (far right) AP Position (mm) [- Post., + Ant.] 10 -5 -10 -15 A contact position anterior to the midline of the tibia was denoted as positive, and a position posterior was denoted as negative For rotation, the angle between the longitudinal axis of the femoral component on the coronal view and the fixed axis passing through the tibial component was measured either medially or laterally (⊡ Fig 22-2) Error analyses for this 3D model-fitting technique have been conducted previously and demonstrated translational errors less of than 0.5 mm, and rotational errors of less than 0.5° [10] Results 15 30 45 60 Flexion Angle (Degrees) 75 90 ⊡ Fig 22-3 Average anteroposterior contact position, evaluated on the sagittal view, for subjects with a medial UKA, during deep knee bend 10.00 AP Position (mm) [- Post., + Ant.] ⊡ Fig 22-2 Technique used to determine the axial rotation of the UKA 5.00 0.00 -5.00 -10.00 -15.00 HS 33 66 Flexion Angle (Degrees) TO ⊡ Fig 22-4 Average anteroposterior contact position, evaluated on the sagittal view, for subjects with a medial UKA, during stance phase of gait, from heel-strike (HS) to toe-off (TO) Anteroposterior Translation Medial UKA On average, during DKB the femoral component moved 3.1 mm posteriorly from 0° to 45° and then 2.3 mm anteriorly from 45° to 90° (⊡ Fig 22-3) Thus from 0° to 90° the femoral components moved an average of 0.8 mm posteriorly Eight knees (47 %) displayed anterior tibiofemoral contact at 0° In the remaining nine knees the inital contact was more posterior At heel-strike, the average contact position for subjects with a medial UA was -0.2 mm (6.1 to -7.2), moving an average of 0.3 mm in the anterior direction to an average contact position of 0.3 mm (6.6 to -7.2) at 33% of gait stance phase (⊡ Fig 22-4).The subjects with a medial UKA remained in a similar position at 66% of stance phase with a contact position of 0.4 mm (7.7 to -6.0) From 66% of stance phase to toe-off, these subjects experienced an average anterior motion, having a contact position of 0.6 mm (7.2 to -8.0) at toe-off Eleven of the 15 subjects experienced less than 2.0 mm of medial UKA motion (anterior or posterior),which is similar to the medial condyle for the normal knee during gait Lateral UKA During DKB, two knees exhibited minimal anteroposterior motion of the femoral component from 0° to 90°,and one exhibited posterior motion of 7.6 mm Between 0° 22 AP Position (mm) [- Post., + Ant.] 10.00 5.00 0.00 -5.00 -10.00 -15.00 HS 33 66 Flexion Angle (Degrees) TO ⊡ Fig 22-5 Average anteroposterior contact position, evaluated on the sagittal view, for subjects with a lateral UKA, during stance phase of gait, from heel-strike (HS) to toe-off (TO) and 90°, all three knees exhibited both anterior and posterior motion at different points in the arc On average, during stance phase of gait, subjects with a lateral UA experienced -0.4 mm of posterior motion from heel-strike to toe-off (⊡ Fig 22-5) At heel-strike,the average contact position for subjects with a lateral UA was -5.7 mm (-3.9 to -83.9),at 33% of stance phase the average was -6.4 mm (-5.7 to -7.6), at 66% of stance phase the average was -7.3 mm (-2.4 to -9.9), and at toe-off the average contact position was -6.1 mm (-4.3 to -8.0) On average, the greatest amount of posterior motion occurred from heel-strike to 66% of stance phase (-1.6 mm),while an anterior slide of 1.2 mm occurred from 66% of stance phase to toe-off Overall, all four subjects experienced less than 2.1 mm of motion, whether the motion occurred in the anterior or the posterior direction Axial Tibiofemoral Rotation Medial UKA On average, during DKB the 17 knees displayed 3.3° of internal tibial rotation between the two components of the Axial Rotation Angle (Deg) III Kinematics 2.5 1.5 0.5 -0.5 -1 -1.5 -2 -2.5 15 30 45 60 75 Knees Flexion Angle (Degrees) 90 ⊡ Fig 22-6 Average axial rotation pattern for subjects having a medial UKA, during deep knee bend Rotation Angle (Degrees) 150 30.00 20.00 10.00 0.00 HS -10.00 33 66 TO -20.00 -30.00 Flexion Angle (Degrees) ⊡ Fig 22-7 Average axial rotation pattern for subjects with a medial UKA, during stance phase of gait, from heel-strike (HS) to toe-off (TO) prosthesis between 0° and 90° flexion (⊡ Fig 22-6) Three knees displayed external tibial rotation, two displayed negligible (5.0° (%) Normal rot.j 17 18 12 21 25 15 13 60 40 90 >10.0°