- Email: [email protected]
Determinants of patellar tracking in total knee arthroplasty
Available online at www.sciencedirect.com
Clinical Biomechanics 23 (2008) 900–910 www.elsevier.com/locate/clinbiomech
Determinants of patellar tracking in total knee arthroplasty C. Anglin a,b,c,*, J.M. Brimacombe a, A.J. Hodgson a, B.A. Masri d, N.V. Greidanus d, J. Tonetti e, D.R. Wilson d,f,g a
Department of Mechanical Engineering, University of British Columbia, Vancouver, Canada Centre for Bioengineering Research and Education, University of Calgary, Calgary, Canada c Department of Civil Engineering, University of Calgary, Calgary, Canada d Department of Orthopaedics, University of British Columbia, Vancouver, Canada e Orthopaedic and Trauma Department, Hoˆpital Michallon, Grenoble, France f Division of Orthopaedic Engineering Research, University of British Columbia, Vancouver, Canada g Vancouver Coastal Health Research Institute, Vancouver, Canada b
Received 16 August 2007; accepted 1 April 2008
Abstract Background. Optimizing patellar tracking in total knee arthroplasty is a surgical priority. Despite this, a comparison of the effects of different component placements on patellar tracking is not available; the biomechanical impact of the patellar resection angle has not been studied; and the similarity between intraoperative and postoperative effects, fundamental to improving patellar tracking, is unknown. Our objective was to compare the impact of the major controllable femoral, tibial and patellar component positions on patellar kinematics during both passive and loaded flexion. Methods. We tested eight cadaveric knee specimens in two rigs, simulating intraoperative and weightbearing flexion. Optoelectronic marker arrays were attached to the femur, tibia and patella to record kinematics throughout the range of motion. We modified posteriorstabilized fixed-bearing knee components to allow for five types of variations in component placement in addition to the neutral position: femoral component rotation, tibial component rotation, patellar resection angle, patellar component medialization and additional patellar thickness, for a total of 11 individual variations. Findings. The major determinants of patellar tilt and shift were patellar component medialization, patellar resection angle and femoral component rotation. The relative order of these variables depended on the structure (bone or component), kinematic parameter (tilt or shift) and flexion angle (early or late flexion). Effects of component changes were consistent between the intraoperative and weightbearing rigs. Interpretation. To improve patellar tracking, and thereby the clinical outcome, surgeons should focus on patellar component medialization, patellar resection angle and femoral component rotation. These have been linked with anterior knee pain as well. Neither tibial component rotation nor patellar thickness should be adjusted to improve patellar tracking. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Total knee replacement; Total knee arthroplasty; Patellar tracking; Patella; Femur; Tibia; Kinematics; Surgical technique; Component rotation; Patellar component medialization; Patellar resection; Patellar thickness; Computer-assisted surgery
1. Introduction A major goal during knee replacement surgery is to achieve normal patellar tracking within the femoral *
Corresponding author. Present Address: Department of Civil Engineering, 2500 University Drive NW, Calgary, AB, Canada T2N 1N4. E-mail address: [email protected] (C. Anglin). 0268-0033/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2008.04.001
groove. Abnormal tracking can lead to subluxation, higher contact forces, smaller contact areas or excessive soft tissue tensions (Barrack and Burak, 2001; Rand, 2004; Stiehl, 2005). These issues may result in postoperative complications such as anterior knee pain, component wear, component loosening and patellar fracture. A limitation of the current literature is that a comparison of the impact of the major controllable component
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
positioning factors on patellar kinematics is not available. In previous studies, researchers have examined the effects of one or two patellar, tibial or femoral component placement changes on patellar tracking to identify guidelines for those individual component placements that lead to improved clinical outcomes (Section 4 provides specific references); however, it is not possible to directly compare the results between studies due to differences in experimental protocol and due to the high natural variability among specimens (Katchburian et al., 2003). Therefore, the relative effects of the different component positioning factors are not known. A comprehensive comparison and approximate ordering of the effects of the major component positioning factors on patellar kinematics is essential for rational component and instrumentation design as well as for planning both traditional and computer-assisted surgery. The patellar resection angle has been relatively ignored in the literature, despite the fact that two research groups have now linked asymmetric resection to anterior knee pain, patellar complications and bony impingement (Baldini et al., 2006, 2007; Pagnano and Trousdale, 2000). The resection angle, furthermore, is directly under the surgeon’s control. To our knowledge, the biomechanical effects of the patellar resection angle have not been studied. Comparing the effects of the resection angle to those of other component variables should indicate whether or not the resection angle deserves greater focus. Patellar component medialization has been the subject of several investigations (Lee et al., 1999; Miller et al., 2001b; Yoshii et al., 1992) but each study tested only one level of medialization. It is therefore unclear whether the observed effects grow in proportion to the amount of medialization. A sudden change in patellar mechanics would indicate either that a larger proportional benefit could be achieved by greater medialization, or that a stricter constraint is needed on the amount of medialization to avoid the negative consequences. Surgeons aim to optimize knee component placement for postoperative performance but it is not clear how well intraoperative observations of patellar tracking during passive knee flexion relate to patellar tracking in activity. The similarity in geometry between the two conditions suggests that the results should be comparable, whereas differences in loading suggest that the effects of component changes may differ between the two conditions. Since all studies of component placement that we are aware of have tested specimens in a weightbearing setup, it is unknown whether or not changes made to component placement intraoperatively have the same effect under weightbearing. This is relevant to both traditional and computer-assisted surgery. The objectives of our study were to compare the impact of the major controllable femoral, tibial and patellar component positions on patellar kinematics (specifically patellar tilt and shift) including multiple levels of medialization and patellar resection angle, during both passive and loaded flexion. Improving patellar tracking should ultimately improve the clinical outcome of total knee arthroplasty.
901
2. Methods We measured femoral, tibial and patellar kinematics throughout flexion and extension using cadaveric specimens implanted with modified knee arthroplasty components that allowed 11 different component positions. All specimens were tested in two different rigs that simulated intraoperative and weightbearing flexion, respectively. The extensive methods, including the designs of the modified component, are described in detail elsewhere (Anglin et al., 2008; Brimacombe, 2006); the most relevant details are highlighted below. 2.1. Specimens and component variations An experienced surgeon implanted modified NexGen LPS posterior-stabilized, fixed-bearing components (Zimmer, Warsaw, USA) into eight cadaveric knee specimens (6 female/2 male, ages 51–80) using a parapatellar approach. The lateral retinaculum remained intact throughout testing. The medial incision was closed during each test using two towel clamps, the locations of which were marked to ensure repeated positioning. The all-polyethylene patellar component matched the femoral component geometry in mid- to deep flexion (Ma et al., 2007). Modified femoral, tibial and patellar components allowed the following clinically-representative variations: (1) 5° internal and 5° external femoral component rotation (FR); (2) 5° internal and 5° external tibial component rotation (TR); (3) 7.5° and 15° lateral and 7.5° medial patellar resection angle (PA); (4) 2.5 and 5.0 mm patellar component medialization (PM); and, (5) 3 mm increased patellar thickness (PT). All components also included a neutral, baseline position (0°, 0 mm), as described below. We selected these variables because they are directly under the surgeon’s control and therefore have the potential to significantly impact patellar kinematics both intraoperatively and postoperatively. Our institutional review board approved the study. 2.2. Design of modified components 2.2.1. Femoral component rotation A long threaded rod, attached permanently to the femoral component, was passed through the intramedullary canal of the specimen such that it was accessible above the resected femur. Sufficient bone was removed anteriorly and posteriorly to allow 5° internal and external rotation; the precise angle was established by comparing the component rotation to a mounted protractor. The neutral
902
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
location (described below) was marked on the femur while it was still intact.
all component variations in passive intraoperative flexion and then with all variations in weightbearing flexion.
2.2.2. Tibial component rotation We designed a custom 3 mm baseplate with two matching sets of three screw holes, corresponding to 0°, +5° and 5° rotation; the standard tibial plate fit onto the baseplate. An extra 3 mm was resected from the bone to accommodate the baseplate. A trial insert was used to permit repeated access to the baseplate screws to change the rotation angle.
2.5. Intraoperative rig
2.2.3. Patellar variations A baseplate was secured to the resected patellar bone surface to accommodate the patellar variations. Sufficient bone was removed such that the final thickness of the patella was within 1 mm of its original thickness. To simulate the different resection angles, we added a neutral disk or wedge, all having the same central thickness; ‘medial’ wedges corresponded to underresection of the medial patellar bone. This disk or wedge was attached to the baseplate via metal pegs that fit into machined holes in the baseplate. A magnet embedded in the baseplate resisted tensile forces that would cause separation of the components. Different patellar positions were achieved by inserting the pegs into different sets of holes on the baseplate. A cylindrical disk was used to create the additional thickness. The bottom of the polyethylene patellar component was modified to attach to the remaining construct, without changing the component diameter. Three different sizes (32, 35, 38 mm) were created for the patellar components. 2.3. Definitions of neutral placement Neutral femoral component rotation was defined as 3° externally rotated from the posterior condylar line. Neutral tibial component rotation was defined by aligning the central axis of the tibial plate with the line connecting the centre of the posterior notch with the junction between the mid and medial thirds of the tibial tubercle. Neutral patellar resection was defined as having equal medial and lateral patellar thicknesses at one-quarter distance across the mediolateral width. Neutral patellar component placement meant central mediolateral and superoinferior placement on the patellar bone, i.e. at the geometric centre of the cut surface of the patella. Neutral patellar thickness was within 1 mm of the original thickness.
This custom-designed apparatus simulated the range of motion test performed by a surgeon to judge patellar tracking intraoperatively (see Fig. 1; see also Anglin et al., 2008 for a photograph of the apparatus). The three degree-offreedom ‘hip’ joint was pushed manually (using a pushrod) along a track toward the three degree-of-freedom ‘ankle’ joint, up to 120° flexion. We maintained a constant speed by using a stopwatch, completing each half-cycle in 15 s. A spring connected between the quadriceps clamp and a point distal to the hip joint (parallel to the femur) replicated passive resistance of the quadriceps muscle and maintained the patella in contact with the anterior surface of the femur. 2.6. Weightbearing Rig This Oxford-type rig simulated a squat or stair-climbing (see Fig. 2; see also Anglin et al., 2008 for a photograph of the apparatus). Many researchers have used similar rigs (Ezzet et al., 2001; Miller et al., 2001a,b; Whiteside and Nakamura, 2003). The knee was mounted vertically between the mechanical ‘ankle’ and ‘hip’ joints, with the tray-weight producing a 54 N resultant load on the hip joint. As with the intraoperative rig, all six degrees of freedom of the knee were unconstrained. The quadriceps tendon clamp was attached via a cable and a pulley to a motor (parallel to the femur). Tension in the cable simulated the extensors and counteracted the flexion moment created by the weight at the hip joint. Shortening and lengthening the cable with the motor caused the knee to flex and extend. We flexed the knee from full extension to 90° flexion. Greater flexion would have stressed the joint and tendon clamp excessively. The quadriceps line of action represented the resultant force vector of all four muscle components. This relatively simple flexion model did not simulate the complex muscle patterns of gait, but was valuable for many reasons: it allowed us to study deeper flexion, which is usually where anterior knee pain occurs; it avoided the potential error
2.4. Testing order The order of testing for different types of component placement (e.g. femoral or tibial component rotation) was randomized. Baseline kinematic measurements, with all components in neutral positions, were taken between each type of component variation to ensure repeatability. At least two cycles were performed for all variations and the results were averaged. Each specimen was tested first with
Spring Hand Hand Force Force
Fig. 1. ‘Intraoperative’ passive-flexion rig.
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910 54 N
M Motor
903
part of the analysis. Geometric predictions represent what the relative tilt and shift of the patellar bone would be if the position and orientation of the patellar component were to remain fixed within the femoral groove; the difference therefore represents component movement with respect to the groove. For example, 5 mm medialization of the patellar component with respect to the patellar bone should geometrically result in a 5 mm lateral shift of the patellar bone (resulting in reduced tension in the lateral retinaculum). A measurement of 3 mm lateral bone shift implies that the patellar component itself, together with the bone remnant, must have shifted 2 mm medially. Geometric predictions for each component variable were overlaid on the measured results; differences, where relevant, are described in Section 4. 2.8. Statistical analysis
Fig. 2. ‘Postoperative’ weightbearing rig.
of an incorrect muscle activation pattern, since no one pattern has widespread acceptance nor can one pattern capture all activities; and importantly, a gait pattern would not have been meaningful in the passive intraoperative rig to which we wished to compare the results.
2.7. Kinematic analysis Femoral, tibial and patellar motions were recorded throughout flexion and extension using an Optotrak optoelectronic system (Northern Digital, Waterloo, Canada). Marker arrays, each with four infrared-emitting diodes (IREDs), were attached firmly to the femur, tibia and patella using K-wires. Superior, inferior, medial and lateral landmarks were then digitized on each bone using the Optotrak pointer in order to relate the marker array coordinate system to an anatomical coordinate system. Patellar tilt and shift were selected as the most clinically relevant kinematic parameters. Patellar tilt was defined as rotation about the superoinferior axis of the patella, while patellar shift was defined as translation parallel to the epicondylar (flexion) axis of the femur (Cole et al., 1993; Katchburian et al., 2003). Results are reported for the extension cycle only, for the sake of graphical clarity. Our main analysis reports the kinematics of the patellar bone remnant relative to the femoral bone, because this is where the marker arrays were attached. Since the kinematics of the patellar component relative to the femoral component are also of interest, we determined these changes by examining the difference between the measured effects and the geometrically-predicted effects. This was an important
We evaluated the hypothesis that there were no differences in patellar bone shift and tilt between the baseline measurements and the component variations using a three-way repeated-measures ANOVA (2 rigs 11 variations 6 flexion angles) with a = 0.05. Differences for individual variables relative to baseline were tested using post-hoc Bonferroni paired t-tests relative to a single control (baseline), i.e. with a = 0.005 to account for the 10 comparisons to baseline; we did not evaluate differences between component variables. We analyzed the data at 15° increments, but focused on two flexion angles in particular: 15°, where loads are low, the femoral groove is shallow, and subluxation is a greater concern; and 90° where loads are high and the patella is normally well-seated in the femoral groove. Results for the full range of flexion, including each variable and both rigs, are detailed elsewhere (Brimacombe, 2006). 2.9. Ordering of component variable effects To synthesize the results of the individual component changes, we ordered the component changes based on their effects on tilt and shift for both the bone remnant and the patellar component in early and late flexion. The largest numerical result (based on the mean effect) received the highest ranking; the intention was not to create a strict ranking, but to imply that factors higher in the list had a greater impact than those lower in the list. Comparing the results as we did assumes that the component-change magnitudes represented equally probable clinical variations; this is appropriate since we selected the test magnitudes based on typical clinical values reported in the literature. We consider this preferable to normalizing to the change per degree or millimeter of component positioning since different levels of variability occur for each component and we wished to reflect standard clinical practice. However, we excluded the 15° patellar resection angle from the ordering because, although it can occur, it is a more unusual case. We also averaged the results at the two levels
904
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
of medialization (2.5 mm and 5.0 mm) to represent a more typical level of medialization. 3. Results Component changes clearly had different levels of influence on both tilt and shift (Figs. 3 and 4). Tilt of the patellar bone remnant was significantly affected by the patellar resection angle, resulting in relative lateral tilt for a medial resection angle and relative medial tilt for a lateral resection angle (P < 0.001). External femoral component rotation caused significant lateral tilt relative to the bone in early flexion (P < 0.005) but had a diminished effect in later flexion. Both tibial component rotation and increased patellar thickness had small and/or inconsistent influences on tilt.
Shift of the patellar bone remnant was significantly affected by 5 mm medialization of the patellar component: medialization of the component caused lateral shift of the patellar bone remnant (P < 0.001). Shift was also more lateral following external femoral component rotation (P < 0.001, for early weightbearing flexion). Shift was affected by all patellar resection angles in late flexion (P < 0.001), whereas in early flexion, changes in shift were smaller and less consistent. The effect of tibial component rotation was minor in the weightbearing rig and highly inconsistent in the intraoperative rig. The effect of increasing the patellar thickness was negligible. Doubling the amount of medialization roughly doubled the amount of shift: the mean change in shift for 2.5 mm component medialization was 1.9 mm lateral compared to a mean change of 3.9 mm lateral shift for 5.0 mm med-
Fig. 3. Change in tilt of the patellar bone with respect to the femoral bone for 10 variations in component placement, relative to the baseline tilt at 15° and 90° flexion. The 10 variations were: 5° internal and external femoral component rotation, 5° internal and external tibial component rotation, 7.5° medial, 7.5° lateral and 15° lateral patellar resection angle, 2.5 and 5.0 mm patellar component medialization and 3 mm additional patellar thickness. Significant differences from baseline (P < 0.005) are marked with an asterisk. The horizontal lines (—) refer to the geometrically-predicted tilt of the patellar bone if the tilt of the patellar component were to remain constant with respect to the femoral groove. The difference between this line and the measured value represents tilt of the patellar component relative to the femoral component. Lateral tilt is positive.
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
905
Fig. 4. Change in shift of the patellar bone with respect to the femoral bone for 10 variations in component placement, relative to the baseline shift at 15° and 90° flexion. The 10 variations were: 5° internal and external femoral component rotation, 5° internal and external tibial component rotation, 7.5° medial, 7.5° lateral and 15° lateral patellar resection angle, 2.5 and 5.0 mm patellar component medialization and 3 mm additional patellar thickness. Significant differences from baseline (P < 0.005) are marked with an asterisk. The horizontal lines (—) refer to the geometrically-predicted shift of the patellar bone if the shift of the patellar component were to remain constant with respect to the femoral groove. The difference between this line and the measured value represents shift of the patellar component relative to the femoral component. Lateral shift is positive.
ialization, with an overall ratio of 2.0 (SD, 0.3). Doubling the amount of medialization also roughly doubled the amount of tilt: the mean change in tilt for 2.5 mm component medialization was 1.7° lateral compared to a mean change of 3.2° lateral tilt for 5.0 mm medialization, with an overall ratio of 1.9 (SD, 0.2). For the resection angle, changes in tilt were proportional to both the direction and size of the resection angle: the medial:lateral ratio was 1.0 (SD, 0.2), while the 15°:7.5° ratio was 2.0 (SD, 0.3), for flexion angles greater than 15°. By contrast, the effect on shift was disproportionately large as the size of the resection angle increased: the change in shift due to a 15° resection angle was 2.6 (SD, 0.5) times greater than for the 7.5° resection angle (for 30° to 90° flexion). While there were some variations in the patterns between the intraoperative and postoperative rigs, these differences were not statistically significant (P > 0.05).
Correlations between the intraoperative and weightbearing kinematics are reported elsewhere (Anglin et al., 2008). Ordering the component variable effects (Table 1) revealed that the variables that produced the largest changes in patellar tracking were: patellar component medialization (PM), patellar resection angle (PA) and femoral component rotation (FR). Their relative order depended on the structure of interest (bone or component), kinematic parameter (tilt or shift) and flexion angle (early or late flexion). Tibial component rotation (TR) and patellar thickness (PT) consistently ranked last. 4. Discussion This study was unique because it made direct comparisons regarding the impact of the most important surgeon-controlled component variables on patellar
906
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
Table 1 Approximate ordering of the effects of the component variables on tilt and shift of the patellar bone and patellar component in early and late flexion Effects on: Patellar Bone Relative to Femoral Bone
TILT Early Late Flexion Flexion PA FR (PM) (TR) (PT)
PA FR PM (TR) (PT)
SHIFT Early Late Flexion Flexion FR PM (PA) (TR) (PT)
PM PA (FR) (TR) (PT)
R A N K I N G
1 2 3 4 5
Effects on: Patellar Component Relative to Femoral Component
TILT Early Late Flexion Flexion PM PA (FR) (TR) (PT)
FR PM PA (TR) (PT)
SHIFT Early Late Flexion Flexion PM (FR) (PA) (TR) (PT)
PM FR PA (TR) (PT)
Variables higher in the list had a greater or more consistent impact on the results than those lower in the list. Variables in brackets had a minimal or variable effect. Abbreviations are: PA = patellar resection angle, PM = patellar component medialization, FR = femoral component rotation, TR = tibial component rotation, PT = additional patellar thickness. (The order presented is for the weightbearing rig. The only differences in the intraoperative rig were that: the effect of PM for late-flexion bone remnant tilt exceeded FR; the effects of PA on component tilt and TR on component shift were greater; and the effects of femoral component rotation were reduced.)
kinematics, using the same specimens and protocol. It was also unique because: (1) we assessed this impact using simulations of both intraoperative and weightbearing mechanics; (2) we investigated the effect of the patellar resection angle; (3) we studied two different levels of medialization and resection angle; and (4) we determined an approximate ordering of the effects of variations in component positioning that might occur in clinical practice, which has never been possible previously. Changes to component placement largely had the predicted geometric effect on the patellar bone remnant (represented by the horizontal lines in Figs. 3 and 4). This indicates that, on average, the tilt and shift of the patellar component remained constant within the femoral groove to within a few degrees and a few millimeters, especially in the weightbearing rig; individual specimens showed greater differences. Discrepancies between the geometrically-predicted tilt or shift and the measured values suggest the influence of the soft tissues on patellar kinematics. 4.1. Neutral alignment kinematics (0°, 0 mm) The baseline kinematics in this study were similar in magnitude to those reported previously with arthroplasty components (Armstrong et al., 2003; Chew et al., 1997; Hsu et al., 1996; Jenny et al., 2002; Lee et al., 1997; Miller et al., 2001a; Rhoads et al., 1990). Kinematic patterns vary considerably among studies, however, in part due to the natural variability among individuals and in part due to differences in experimental protocol (Katchburian et al., 2003). In the present study, the patella generally shifted and tilted laterally in early flexion, approaching neutral shift and tilt in later flexion. Since this is the first study sim-
ulating intraoperative flexion, there are no studies available for comparison. Absolute patellar kinematics may be affected by a variety of factors. We minimized the impact of these factors by focusing our results and conclusions on the effects of changes in component position. 4.2. Femoral component rotation (FR) External femoral component rotation had less effect on patellar tilt and shift at 90° as compared to 15°, which is consistent with the geometry: although the femoral groove is deeper at 90°, femoral component rotation at this flexion angle occurs almost perpendicular to the surface of the patella. In fact, in deeper flexion (as seen in the intraoperative rig), the effects on shift of internal and external femoral rotation crossed over and were opposite to those in early flexion because the trochlea itself rotates in the opposite direction relative to the central femoral axis (Brimacombe, 2006). In contrast, internal femoral component rotation produced similar or greater tilt at 90° compared to 15° in the weightbearing rig, suggesting that the soft tissues contribute more to patellar kinematics at 90° than joint geometry. The direction and magnitude of the changes in tilt and shift that we measured were similar to those found in previous cadaveric studies (Anouchi et al., 1993; Armstrong et al., 2003; Miller et al., 2001a,b; Rhoads et al., 1990), although it is difficult to compare results directly due to differing experimental protocols. A CT-based clinical study (Matsuda et al., 2001) found a correlation between femoral component rotation and patellar tilt at 30° knee flexion, consistent with our early-flexion results. An intraoperative study (Akagi et al., 1999) found that externally rotating the
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
femoral component by 3–5° improved patellar tracking and reduced the need for lateral retinacular release; this is consistent with our finding of a lateral shift of the patellar bone remnant in early flexion, which would reduce the Q-angle and reduce the tension in the lateral retinaculum. While this study shows that femoral component rotation could be used to improve patellar tracking, the additional effects of femoral component rotation on tibiofemoral mechanics limit the degree to which a surgeon can exploit this feature. 4.3. Tibial component rotation (TR) Tibial component rotation does not appear to be a useful technique for improving patellar tracking. The only other cadaveric study of tibial component rotation (Nagamine et al., 1994) found no significant effect on patellar tilt or shift except with 15° of external component rotation. In an in vivo imaging study (Matsuda et al., 2001), a significant correlation was reported between tibial component rotation (averaging 17° internal) and patellar tilt (averaging 3° lateral); however this amount of tibial component rotation is very large, the degree of tilt was small and the correlation was weaker than for femoral rotation. They found no relationship between tibial component rotation and patellar shift. Internal rotation of the tibial component and combined internal rotation of the tibial and femoral components have been linked to anterior knee pain (Barrack et al., 2001). There are two possible explanations of the discrepancy between this link and the low ranking in our study: (1) internal component rotation may occur in concert with other surgical errors, or (2) tibial component rotation may have a large impact on some knees but not on others, as evidenced by the considerable variability that we found in intraoperative shift. Putting these results in the context of our study, which investigates the factors under the surgeon’s control to affect patellar tracking, the effects of tibial component rotation cannot be used predictably during surgery, and the impact appears to be minimal during weightbearing.
907
(bone plus component) back in the lateral direction. This effect may be larger in a patient with a tight lateral retinaculum or a subluxing patella. The higher effect intraoperatively suggests that the retinacula indeed play an important role, although their role appears to be reduced under weightbearing. Our findings are consistent with the trends reported in clinical studies (Kawano et al., 2002; Chan and Gill, 1999; Gomes et al., 1988), i.e. that medial underresection causes lateral tilt of the bone and medial tilt of the component. Some authors (Chan and Gill, 1999; Kawano et al., 2002) have suggested that limited underresection of the medial side may be beneficial, whereas others (Ledger et al., 2005) have concluded that cut angle had little impact on tilt; the difference may be partially a judgment on how much is an important amount. One group (Chan and Gill, 1999) has argued that a thicker medial side mimics the original shape more closely; another (Kawano et al., 2002) has argued that since the patella tilts laterally in the natural knee, a 5° oblique cut would therefore result in neutral tilt of the patellar component. None of the clinical studies observed a noticeable effect on shift relative to the femoral component (Chan and Gill, 1999; Gomes et al., 1988; Kawano et al., 2002). We likewise found a minimal effect for the 7.5° resection angles; however component shift for the 15° resection angle was large and significant. Anterior knee pain has been reported to be more prevalent in patients with asymmetric patellar resurfacing compared to those with symmetric resurfacing (Baldini et al., 2006, 2007; Pagnano and Trousdale, 2000). There are many possible sources for the pain, including bony impingement on the femoral component from excessive tilt, especially under load (Baldini et al., 2006, 2007), poor implant design, elevated soft tissue strains or other errors in surgical technique. Underresection is also generally associated with a thicker patellar construct (Ledger et al., 2005), which may lead to overstuffing of the joint. A further disadvantage that we found is that the effect on shift was disproportionate to the resection angle, indicating that the consequences of a poor resection angle were increasingly severe as the resection angle increased. These results suggest that excessive obliquity should be avoided.
4.4. Patellar resection angle (PA) 4.5. Patellar medialization (PM) Achieving a desired patellar cut during total knee arthroplasty is difficult due to the small, hard bone and limited visibility. Cut asymmetry is common (Baldini et al., 2006, 2007; Chan and Gill, 1999; Gomes et al., 1988; Kawano et al., 2002; Ledger et al., 2005; Pagnano and Trousdale, 2000) even though most surgeons strive for equal medial and lateral thicknesses. The large, significant impact that the patellar resection angle had on patellar bone tilt was geometrically expected. The result, e.g. medial tilt of the patellar bone for a lateral wedge, would tend to increase tension in the opposing (in this case, lateral) soft tissue restraints, which may explain the small but consistent rotation of the whole construct
Lateral shift of the patellar bone due to medialization of the patellar component likely reduces tension in the lateral retinaculum, which may reduce the need for lateral release (Hofmann et al., 1997; Lewonowski et al., 1997). Lateral tilt and shift of the patellar bone due to component medialization is consistent with previous biomechanical studies (Lee et al., 1999; Miller et al., 2001b; Yoshii et al., 1992) and clinical studies (Chan and Gill, 1999; Lewonowski et al., 1997; Nelissen et al., 1995). The lateral tilt, which has not previously been explained, is likely induced by the lateral displacement of the quadriceps attachment relative to the patellar component, which thereby creates a
908
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
moment about the patellar component (Brimacombe, 2006). Slight medialization, on the order of 2.5 mm, may offer a balance between the desirable reduction in lateral subluxation and the undesirable increase in lateral tilt, which could result in bony impingement. The proportionality of changes in tilt and shift with medialization means that there was no sudden change in the consequences of medialization; it also means that the kinematic effects are reasonably predictable. The roughly linear relationship that we found suggests that the benefits and drawbacks can be traded off equally. 4.6. Patellar thickness (PT) Similar to our results, a clinical study recently reported no gross, visible effect on patellar tracking in vivo intraoperatively with additional patellar thicknesses up to 8 mm (Bengs and Scott, 2006). In contrast, the one previous biomechanical study determined that increasing patellar thickness increased lateral tilt and medial shift (Hsu et al., 1996). Radiographically, one study observed changes in tilt (Laughlin et al., 1996) whereas another did not (Kawano et al., 2002). The hypothesis that tracking is not affected by changes in patellar thickness, as we found, is supported by the argument that the Q-angle and patellofemoral joint geometry have not been altered in the coronal plane. However, these results may be different in the presence of abnormally tight soft tissues, e.g. due to a tight lateral retinaculum in the valgus knee, or due to overstuffing of the joint by an oversized femoral component. The main disadvantage appears to be a reduced range of motion (RoM): the intraoperative clinical study described above revealed a 3° reduction in RoM for every 2 mm of additional patellar thickness (Bengs and Scott, 2006). 4.7. Approximate ordering of effects of component placement variations Three component variables (patellar component medialization, patellar resection angle and femoral component rotation) consistently produced the largest effects on patellar kinematics, although the ordering among these factors depended on whether the structure of interest was the patellar bone or component, whether the kinematic parameter was tilt or shift and whether it was early or late flexion (see Table 1). If the greatest clinical concern is subluxation of the patellar component, then the surgeon should focus on patellar component medialization. If, on the other hand, tilt of the patellar bone remnant, which can affect retinacular tensions or cause impingement, is the greater clinical concern, then the surgeon should focus on the patellar resection angle. 4.8. Limitations The primary limitation of this study is that of any cadaveric study, i.e. test rigs do not completely replicate
physiological loading in vivo. Nevertheless the consistency in results relative to baseline between the intraoperative and weightbearing simulations and the clear distinctions between component effects support the general conclusions. Furthermore, a parametric study such as we have performed (i.e. comparing the effects of different component positions on the same specimens) is only possible ex vivo. Unlike clinical studies, a biomechanical study has the advantage of being able to isolate the effects of particular component variables. We were also able to investigate kinematics throughout the range of motion, in contrast to most clinical studies, which only report results at a single static flexion angle. The effects of any discrepancies in absolute kinematics between the ex vivo model and in vivo kinematics were minimized by our focus on relative effects, i.e. changes in kinematics due to changes in component position. A second limitation, common to almost all cadaveric studies, is that we did not measure tilt and shift of the patellar component relative to the femoral component directly; most studies do not mention it at all. We calculated these indirectly by subtracting the measured bone kinematics from the geometrically-predicted values. Both the boneto-bone and the component-to-component values are important. Bone-to-bone values are of interest because tilt and shift can affect the medial and lateral soft tissue tensions and because these values correspond to what the surgeon can observe most easily during the intraoperative assessment. Component-to-component values are important because they relate to component longevity and can be compared to postoperative radiographic analyses. A third limitation is that we only studied a single component design. Due to the specific bone cuts made to accommodate the components, the need to machine modified components, and the practical limitations on the number of trials per specimen, it would not have been possible to repeat these measurements with a second design that had different geometry or tibiofemoral rotational freedom. It is possible that the relatively deep trochlear groove of the NexGen design captured the patella more than some other designs; however recognition of the importance of a deeper groove for patellar tracking (Chew et al., 1997; Yoshii et al., 1992) means that this is present in most current designs. The angled orientation of the NexGen femoral groove, and the marginally higher lateral flange, could result in different absolute kinematics than for a symmetric component, but these differences are unlikely to affect the ordering of the effects of changes in component placement, especially given that the proximal groove is shallow. A biomechanical study of three different component designs, including the NexGen, did not reveal significant differences among the designs (Chew et al., 1997). Likewise, a fluoroscopic study (Stiehl et al., 2001) did not detect significant differences between posterior-stabilized and cruciate-retaining designs in sagittal plane patellar kinematics. In our opinion, therefore, although the absolute kinematics and the precise relative results (e.g. Figs. 3 and 4) may differ
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
subtly among designs, we expect that the overall findings regarding which component changes have the greatest impact on patellar kinematics will be valid across designs. A fourth limitation is that cadaveric studies only represent the immediate postoperative period. Two studies (Laughlin et al., 1996; Shih et al., 2004) have recorded increasingly lateral tracking over time; this future change should be considered when improving tracking intraoperatively. 5. Conclusions To improve patellar tracking, the surgeon should focus on patellar medialization (aiming for slight medialization), patellar resection angle (aiming for symmetry) and femoral component rotation (aiming for anatomic reconstruction, favouring external rotation). The choice of which component position to change for the greatest impact depends on whether the change should primarily affect tilt or shift of the component or the bone remnant, in late flexion or early flexion. The ordering table is unique in the literature and was only possible by testing numerous component variables on each specimen. The patellar resection angle, which has not previously been studied, ranked highly against the other component variables, suggesting that it deserves more attention than it has previously received, especially given its link to anterior knee pain. The consistency in component placement effects between the weightbearing rig and the simulated intraoperative assessment, which has not previously been reported, supports the use of intraoperative assessment of component changes to improve postoperative mechanics. Reducing patellar maltracking could lead to an improved clinical outcome by reducing anterior knee pain, patellar fracture, component wear and component loosening. Acknowledgements We wish to thank the Natural Sciences and Engineering Research Council of Canada (NSERC), Praxim (Grenoble, France), the Canadian Arthritis Network, and the Michael Smith Foundation for Health Research for their generous financial support, and Zimmer Canada for the components and use of the instruments. We would also like to thank Karen Ho for her assistance in creating the experimental apparatus schematics. References Akagi, M., Matsusue, Y., Mata, T., Asada, Y., Horiguchi, M., Iida, H., Nakamura, T., 1999. Effect of rotational alignment on patellar tracking in total knee arthroplasty. Clin. Orthop. Relat. Res. 366, 155–163. Anglin, C., Brimacombe, J.M., Wilson, D.R., Masri, B.A., Greidanus, N.V., Tonetti, J., Hodgson, A.J., 2008. Intraoperative vs. weightbearing patellar kinematics in total knee arthroplasty: a cadaveric study. Clin. Biomech. 23, 60–70. Anouchi, Y.S., Whiteside, L.A., Kaiser, A.D., Milliano, M.T., 1993. The effects of axial rotational alignment of the femoral component on knee
909
stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens. Clin. Orthop. Relat. Res. 287, 170–177. Armstrong, A.D., Brien, H.J.C., Dunning, C.E., King, G.J.W., Johnson, J.A., Chess, D.G., 2003. Patellar position after total knee arthroplasty. J. Arthroplasty 18 (4), 458–465. Baldini, A., Anderson, J.A., Cerulli-Mariani, P., Kalyvas, J., Pavlov, H., Sculco, T.P., 2007. Patellofemoral evaluation after total knee arthroplasty: validation of a new weight-bearing axial radiographic view. J. Bone Joint Surg. Am. 89 (8), 1810–1817. Baldini, A., Anderson, J.A., Zampetti, P., Pavlov, H., Sculco, T.P., 2006. A new patellofemoral scoring system for total knee arthroplasty. Clin. Orthop. Relat. Res. 452, 150–154. Barrack, R.L., Burak, C., 2001. Patella in total knee arthroplasty. Clin. Orthop. Relat. Res. 389, 62–73. Barrack, R.L., Schrader, T., Bertot, A.J., Wolfe, M.W., Myers, L., 2001. Component rotation and anterior knee pain after total knee arthroplasty. Clin. Orthop. Relat. Res. 392, 46–55. Bengs, B.C., Scott, R.D., 2006. The effect of patellar thickness on intraoperative knee flexion and patellar tracking in total knee arthroplasty. J. Arthroplasty 21 (5), 650–655. Brimacombe, J.M., 2006. Effects of component placement on patellar kinematics and loading in intraoperative and postoperative loading configurations. M.A.Sc. thesis, University of British Columbia, Vancouver. Chan, K.C., Gill, G.S., 1999. Postoperative patellar tilt in total knee arthroplasty. J. Arthroplasty 14 (3), 300–304. Chew, J.T.H., Stewart, N.J., Hanssen, A.D., Zong-Ping, L., Rand, J.A., An, K-N., 1997. Differences in patellar tracking and knee kinematics among three different total knee designs. Clin. Orthop. Relat. Res. 345, 87–98. Cole, G.K., Nigg, B.M., Ronsky, J.L., Yeadon, M.R., 1993. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization proposal. J. Biomech. Eng. 115, 344–349. Ezzet, K.A., Hershey, A.L., D’Lima, D.D., Irby, S.E., Kaufman, K.R., Colwell, C.W., 2001. Patellar tracking in total knee arthroplasty – inset versus onset design. J. Arthroplasty 16 (7), 838–843. Gomes, L.S.M., Bechtold, J.E., Gustilo, R.B., 1988. Patellar prosthesis positioning in total knee arthroplasty. Clin. Orthop. Relat. Res. 236, 72–81. Hofmann, A.A., Tkach, T.K., Evanich, C.J., Camargo, M.P., Zhang, Y., 1997. Patellar component medialization in total knee arthroplasty. J. Arthroplasty 12 (2), 155–160. Hsu, H.-C., Zong-Ping, L., Rand, J.A., An, K.-N., 1996. Influence of patellar thickness on patellar tracking and patellofemoral contact characteristics after total knee arthroplasty. J. Arthroplasty 11 (1), 69– 80. Jenny, J.Y., Lefebvre, Y., Vernizeau, M., Lavaste, F., Skalli, W., 2002. In vitro analysis of the continuous active patellofemoral kinematics of the normal and prosthetic knee. Rev. Chir. Orthop. Re´par. l’App. Mot. 88 (8), 797–802. Katchburian, M.V., Bull, A.M.J., Shih, Y.-F., Heatley, F.W., Amis, A.A., 2003. Measurement of patellar tracking: assessment and analysis of the literature. Clin. Orthop. Relat. Res. 412, 241–259. Kawano, T., Miura, H., Nagamine, R., Urabe, K., Matsuda, S., Takaaki, M.-O., Iwamoto, Y., 2002. Factors affecting patellar tracking after total knee arthroplasty. J. Arthroplasty 17 (7), 942–947. Laughlin, R.T., Werries, B.A., Verhulst, S.J., Hayes, J.M., 1996. Patellar tilt in total knee arthroplasty. Am. J. Orthop. 23, 300–304. Ledger, M., Shakespeare, D., Scaddan, M., 2005. Accuracy of patellar resection in total knee replacement: a study using the medial pivot knee. Knee 12, 13–19. Lee, T.Q., Budoff, J.E., Glaser, F.E., 1999. Patellar component positioning in total knee arthroplasty. Clin. Orthop. Relat. Res. 366, 274– 281. Lee, T.Q., Gerken, A.P., Glaser, F.E., Kim, W.C., Anzel, S.H., 1997. Patellofemoral joint kinematics and contact pressures in total knee arthroplasty. Clin. Orthop. Relat. Res. 340, 257–266.
910
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
Lewonowski, K., Dorr, L.D., McPherson, E.J., Huber, G., Wan, Z., 1997. Medialization of the patella in total knee arthroplasty. J. Arthroplasty 12 (2), 161–167. Ma, H.M., Lu, Y.C., Kwok, T.G., Ho, F.Y., Huang, C.Y., Huang, C.H., 2007. The effect of the design of the femoral component on the conformity of the patellofemoral joint in total knee replacement. J. Bone Joint Surg. Br. 89, 408–412. Matsuda, S., Miura, H., Nagamine, R., Urabe, K., Hirata, G., Iwamoto, Y., 2001. Effect of femoral and tibial component position on patellar tracking following total knee arthroplasty: 10-year follow-up of MillerGalante I knees. Am. J. Knee Surg. 14, 152–156. Miller, M.C., Berger, R.A., Petrella, A.J., Karmas, A., Rubash, H.E., 2001a. Optimizing femoral component rotation in total knee arthroplasty. Clin. Orthop. Relat. Res. 392, 38–45. Miller, M.C., Zhang, A.X., Petrella, A.J., Berger, R.A., Rubash, H.E., 2001b. The effect of component placement on knee kinetics after arthroplasty with an unconstrained prosthesis. J. Orthop. Res. 19 (4), 614–620. Nagamine, R., Whiteside, L.A., White, S.E., McCarthy, D.S., 1994. Patellar tracking after total knee arthroplasty: the effect of tibial tray malrotation and articular surface configuration. Clin. Orthop. Relat. Res. 304, 263–271. Nelissen, R.G.H.H., Weidenheim, L., Mikhail, W.E.M., 1995. The influence of the position of the patellar component on tracking in total knee arthroplasty. Int. Orthop. 19, 224–228.
Pagnano, M.W., Trousdale, R.T., 2000. Asymmetric patella resurfacing in total knee arthroplasty. Am. J. Knee Surg. 13, 228–233. Rand, J.A., 2004. Extensor mechanism complications following total knee arthroplasty. J. Bone Joint Surg. Am. 86A (9), 2062– 2072. Rhoads, D.D., Noble, P.C., Reuben, J.D., Mahoney, O.M., Tullos, H.S., 1990. The effect of femoral component position on patellar tracking after total knee arthroplasty. Clin. Orthop. Relat. Res. 260, 43–51. Shih, H.-N., Shih, L.-Y., Wong, Y.-C., Hsu, R.W.-W., 2004. Long-term changes of the nonresurfaced patella after total knee arthroplasty. J. Bone Joint Surg. Am. 86, 935–939. Stiehl, J.B., 2005. A clinical overview: patellofemoral joint and application to total knee arthroplasty. J. Biomech. 38, 209–214. Stiehl, J.B., Komistek, R.D., Dennis, D.A., Keblish, P.A., 2001. Kinematics of the patellofemoral joint in total knee arthroplasty. J. Arthroplasty 16 (6), 706–714. Whiteside, L.A., Nakamura, T., 2003. Effect of femoral component design on unresurfaced patellas in knee arthroplasty. Clin. Orthop. Relat. Res. 410, 189–198. Yoshii, I., Whiteside, L.A., Anouchi, Y.S., 1992. The effect of patellar button placement and femoral component design on patellar tracking in total knee arthroplasty. Clin. Orthop. Relat. Res. 275, 211–219.
Clinical Biomechanics 23 (2008) 900–910 www.elsevier.com/locate/clinbiomech
Determinants of patellar tracking in total knee arthroplasty C. Anglin a,b,c,*, J.M. Brimacombe a, A.J. Hodgson a, B.A. Masri d, N.V. Greidanus d, J. Tonetti e, D.R. Wilson d,f,g a
Department of Mechanical Engineering, University of British Columbia, Vancouver, Canada Centre for Bioengineering Research and Education, University of Calgary, Calgary, Canada c Department of Civil Engineering, University of Calgary, Calgary, Canada d Department of Orthopaedics, University of British Columbia, Vancouver, Canada e Orthopaedic and Trauma Department, Hoˆpital Michallon, Grenoble, France f Division of Orthopaedic Engineering Research, University of British Columbia, Vancouver, Canada g Vancouver Coastal Health Research Institute, Vancouver, Canada b
Received 16 August 2007; accepted 1 April 2008
Abstract Background. Optimizing patellar tracking in total knee arthroplasty is a surgical priority. Despite this, a comparison of the effects of different component placements on patellar tracking is not available; the biomechanical impact of the patellar resection angle has not been studied; and the similarity between intraoperative and postoperative effects, fundamental to improving patellar tracking, is unknown. Our objective was to compare the impact of the major controllable femoral, tibial and patellar component positions on patellar kinematics during both passive and loaded flexion. Methods. We tested eight cadaveric knee specimens in two rigs, simulating intraoperative and weightbearing flexion. Optoelectronic marker arrays were attached to the femur, tibia and patella to record kinematics throughout the range of motion. We modified posteriorstabilized fixed-bearing knee components to allow for five types of variations in component placement in addition to the neutral position: femoral component rotation, tibial component rotation, patellar resection angle, patellar component medialization and additional patellar thickness, for a total of 11 individual variations. Findings. The major determinants of patellar tilt and shift were patellar component medialization, patellar resection angle and femoral component rotation. The relative order of these variables depended on the structure (bone or component), kinematic parameter (tilt or shift) and flexion angle (early or late flexion). Effects of component changes were consistent between the intraoperative and weightbearing rigs. Interpretation. To improve patellar tracking, and thereby the clinical outcome, surgeons should focus on patellar component medialization, patellar resection angle and femoral component rotation. These have been linked with anterior knee pain as well. Neither tibial component rotation nor patellar thickness should be adjusted to improve patellar tracking. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Total knee replacement; Total knee arthroplasty; Patellar tracking; Patella; Femur; Tibia; Kinematics; Surgical technique; Component rotation; Patellar component medialization; Patellar resection; Patellar thickness; Computer-assisted surgery
1. Introduction A major goal during knee replacement surgery is to achieve normal patellar tracking within the femoral *
Corresponding author. Present Address: Department of Civil Engineering, 2500 University Drive NW, Calgary, AB, Canada T2N 1N4. E-mail address: [email protected] (C. Anglin). 0268-0033/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2008.04.001
groove. Abnormal tracking can lead to subluxation, higher contact forces, smaller contact areas or excessive soft tissue tensions (Barrack and Burak, 2001; Rand, 2004; Stiehl, 2005). These issues may result in postoperative complications such as anterior knee pain, component wear, component loosening and patellar fracture. A limitation of the current literature is that a comparison of the impact of the major controllable component
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
positioning factors on patellar kinematics is not available. In previous studies, researchers have examined the effects of one or two patellar, tibial or femoral component placement changes on patellar tracking to identify guidelines for those individual component placements that lead to improved clinical outcomes (Section 4 provides specific references); however, it is not possible to directly compare the results between studies due to differences in experimental protocol and due to the high natural variability among specimens (Katchburian et al., 2003). Therefore, the relative effects of the different component positioning factors are not known. A comprehensive comparison and approximate ordering of the effects of the major component positioning factors on patellar kinematics is essential for rational component and instrumentation design as well as for planning both traditional and computer-assisted surgery. The patellar resection angle has been relatively ignored in the literature, despite the fact that two research groups have now linked asymmetric resection to anterior knee pain, patellar complications and bony impingement (Baldini et al., 2006, 2007; Pagnano and Trousdale, 2000). The resection angle, furthermore, is directly under the surgeon’s control. To our knowledge, the biomechanical effects of the patellar resection angle have not been studied. Comparing the effects of the resection angle to those of other component variables should indicate whether or not the resection angle deserves greater focus. Patellar component medialization has been the subject of several investigations (Lee et al., 1999; Miller et al., 2001b; Yoshii et al., 1992) but each study tested only one level of medialization. It is therefore unclear whether the observed effects grow in proportion to the amount of medialization. A sudden change in patellar mechanics would indicate either that a larger proportional benefit could be achieved by greater medialization, or that a stricter constraint is needed on the amount of medialization to avoid the negative consequences. Surgeons aim to optimize knee component placement for postoperative performance but it is not clear how well intraoperative observations of patellar tracking during passive knee flexion relate to patellar tracking in activity. The similarity in geometry between the two conditions suggests that the results should be comparable, whereas differences in loading suggest that the effects of component changes may differ between the two conditions. Since all studies of component placement that we are aware of have tested specimens in a weightbearing setup, it is unknown whether or not changes made to component placement intraoperatively have the same effect under weightbearing. This is relevant to both traditional and computer-assisted surgery. The objectives of our study were to compare the impact of the major controllable femoral, tibial and patellar component positions on patellar kinematics (specifically patellar tilt and shift) including multiple levels of medialization and patellar resection angle, during both passive and loaded flexion. Improving patellar tracking should ultimately improve the clinical outcome of total knee arthroplasty.
901
2. Methods We measured femoral, tibial and patellar kinematics throughout flexion and extension using cadaveric specimens implanted with modified knee arthroplasty components that allowed 11 different component positions. All specimens were tested in two different rigs that simulated intraoperative and weightbearing flexion, respectively. The extensive methods, including the designs of the modified component, are described in detail elsewhere (Anglin et al., 2008; Brimacombe, 2006); the most relevant details are highlighted below. 2.1. Specimens and component variations An experienced surgeon implanted modified NexGen LPS posterior-stabilized, fixed-bearing components (Zimmer, Warsaw, USA) into eight cadaveric knee specimens (6 female/2 male, ages 51–80) using a parapatellar approach. The lateral retinaculum remained intact throughout testing. The medial incision was closed during each test using two towel clamps, the locations of which were marked to ensure repeated positioning. The all-polyethylene patellar component matched the femoral component geometry in mid- to deep flexion (Ma et al., 2007). Modified femoral, tibial and patellar components allowed the following clinically-representative variations: (1) 5° internal and 5° external femoral component rotation (FR); (2) 5° internal and 5° external tibial component rotation (TR); (3) 7.5° and 15° lateral and 7.5° medial patellar resection angle (PA); (4) 2.5 and 5.0 mm patellar component medialization (PM); and, (5) 3 mm increased patellar thickness (PT). All components also included a neutral, baseline position (0°, 0 mm), as described below. We selected these variables because they are directly under the surgeon’s control and therefore have the potential to significantly impact patellar kinematics both intraoperatively and postoperatively. Our institutional review board approved the study. 2.2. Design of modified components 2.2.1. Femoral component rotation A long threaded rod, attached permanently to the femoral component, was passed through the intramedullary canal of the specimen such that it was accessible above the resected femur. Sufficient bone was removed anteriorly and posteriorly to allow 5° internal and external rotation; the precise angle was established by comparing the component rotation to a mounted protractor. The neutral
902
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
location (described below) was marked on the femur while it was still intact.
all component variations in passive intraoperative flexion and then with all variations in weightbearing flexion.
2.2.2. Tibial component rotation We designed a custom 3 mm baseplate with two matching sets of three screw holes, corresponding to 0°, +5° and 5° rotation; the standard tibial plate fit onto the baseplate. An extra 3 mm was resected from the bone to accommodate the baseplate. A trial insert was used to permit repeated access to the baseplate screws to change the rotation angle.
2.5. Intraoperative rig
2.2.3. Patellar variations A baseplate was secured to the resected patellar bone surface to accommodate the patellar variations. Sufficient bone was removed such that the final thickness of the patella was within 1 mm of its original thickness. To simulate the different resection angles, we added a neutral disk or wedge, all having the same central thickness; ‘medial’ wedges corresponded to underresection of the medial patellar bone. This disk or wedge was attached to the baseplate via metal pegs that fit into machined holes in the baseplate. A magnet embedded in the baseplate resisted tensile forces that would cause separation of the components. Different patellar positions were achieved by inserting the pegs into different sets of holes on the baseplate. A cylindrical disk was used to create the additional thickness. The bottom of the polyethylene patellar component was modified to attach to the remaining construct, without changing the component diameter. Three different sizes (32, 35, 38 mm) were created for the patellar components. 2.3. Definitions of neutral placement Neutral femoral component rotation was defined as 3° externally rotated from the posterior condylar line. Neutral tibial component rotation was defined by aligning the central axis of the tibial plate with the line connecting the centre of the posterior notch with the junction between the mid and medial thirds of the tibial tubercle. Neutral patellar resection was defined as having equal medial and lateral patellar thicknesses at one-quarter distance across the mediolateral width. Neutral patellar component placement meant central mediolateral and superoinferior placement on the patellar bone, i.e. at the geometric centre of the cut surface of the patella. Neutral patellar thickness was within 1 mm of the original thickness.
This custom-designed apparatus simulated the range of motion test performed by a surgeon to judge patellar tracking intraoperatively (see Fig. 1; see also Anglin et al., 2008 for a photograph of the apparatus). The three degree-offreedom ‘hip’ joint was pushed manually (using a pushrod) along a track toward the three degree-of-freedom ‘ankle’ joint, up to 120° flexion. We maintained a constant speed by using a stopwatch, completing each half-cycle in 15 s. A spring connected between the quadriceps clamp and a point distal to the hip joint (parallel to the femur) replicated passive resistance of the quadriceps muscle and maintained the patella in contact with the anterior surface of the femur. 2.6. Weightbearing Rig This Oxford-type rig simulated a squat or stair-climbing (see Fig. 2; see also Anglin et al., 2008 for a photograph of the apparatus). Many researchers have used similar rigs (Ezzet et al., 2001; Miller et al., 2001a,b; Whiteside and Nakamura, 2003). The knee was mounted vertically between the mechanical ‘ankle’ and ‘hip’ joints, with the tray-weight producing a 54 N resultant load on the hip joint. As with the intraoperative rig, all six degrees of freedom of the knee were unconstrained. The quadriceps tendon clamp was attached via a cable and a pulley to a motor (parallel to the femur). Tension in the cable simulated the extensors and counteracted the flexion moment created by the weight at the hip joint. Shortening and lengthening the cable with the motor caused the knee to flex and extend. We flexed the knee from full extension to 90° flexion. Greater flexion would have stressed the joint and tendon clamp excessively. The quadriceps line of action represented the resultant force vector of all four muscle components. This relatively simple flexion model did not simulate the complex muscle patterns of gait, but was valuable for many reasons: it allowed us to study deeper flexion, which is usually where anterior knee pain occurs; it avoided the potential error
2.4. Testing order The order of testing for different types of component placement (e.g. femoral or tibial component rotation) was randomized. Baseline kinematic measurements, with all components in neutral positions, were taken between each type of component variation to ensure repeatability. At least two cycles were performed for all variations and the results were averaged. Each specimen was tested first with
Spring Hand Hand Force Force
Fig. 1. ‘Intraoperative’ passive-flexion rig.
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910 54 N
M Motor
903
part of the analysis. Geometric predictions represent what the relative tilt and shift of the patellar bone would be if the position and orientation of the patellar component were to remain fixed within the femoral groove; the difference therefore represents component movement with respect to the groove. For example, 5 mm medialization of the patellar component with respect to the patellar bone should geometrically result in a 5 mm lateral shift of the patellar bone (resulting in reduced tension in the lateral retinaculum). A measurement of 3 mm lateral bone shift implies that the patellar component itself, together with the bone remnant, must have shifted 2 mm medially. Geometric predictions for each component variable were overlaid on the measured results; differences, where relevant, are described in Section 4. 2.8. Statistical analysis
Fig. 2. ‘Postoperative’ weightbearing rig.
of an incorrect muscle activation pattern, since no one pattern has widespread acceptance nor can one pattern capture all activities; and importantly, a gait pattern would not have been meaningful in the passive intraoperative rig to which we wished to compare the results.
2.7. Kinematic analysis Femoral, tibial and patellar motions were recorded throughout flexion and extension using an Optotrak optoelectronic system (Northern Digital, Waterloo, Canada). Marker arrays, each with four infrared-emitting diodes (IREDs), were attached firmly to the femur, tibia and patella using K-wires. Superior, inferior, medial and lateral landmarks were then digitized on each bone using the Optotrak pointer in order to relate the marker array coordinate system to an anatomical coordinate system. Patellar tilt and shift were selected as the most clinically relevant kinematic parameters. Patellar tilt was defined as rotation about the superoinferior axis of the patella, while patellar shift was defined as translation parallel to the epicondylar (flexion) axis of the femur (Cole et al., 1993; Katchburian et al., 2003). Results are reported for the extension cycle only, for the sake of graphical clarity. Our main analysis reports the kinematics of the patellar bone remnant relative to the femoral bone, because this is where the marker arrays were attached. Since the kinematics of the patellar component relative to the femoral component are also of interest, we determined these changes by examining the difference between the measured effects and the geometrically-predicted effects. This was an important
We evaluated the hypothesis that there were no differences in patellar bone shift and tilt between the baseline measurements and the component variations using a three-way repeated-measures ANOVA (2 rigs 11 variations 6 flexion angles) with a = 0.05. Differences for individual variables relative to baseline were tested using post-hoc Bonferroni paired t-tests relative to a single control (baseline), i.e. with a = 0.005 to account for the 10 comparisons to baseline; we did not evaluate differences between component variables. We analyzed the data at 15° increments, but focused on two flexion angles in particular: 15°, where loads are low, the femoral groove is shallow, and subluxation is a greater concern; and 90° where loads are high and the patella is normally well-seated in the femoral groove. Results for the full range of flexion, including each variable and both rigs, are detailed elsewhere (Brimacombe, 2006). 2.9. Ordering of component variable effects To synthesize the results of the individual component changes, we ordered the component changes based on their effects on tilt and shift for both the bone remnant and the patellar component in early and late flexion. The largest numerical result (based on the mean effect) received the highest ranking; the intention was not to create a strict ranking, but to imply that factors higher in the list had a greater impact than those lower in the list. Comparing the results as we did assumes that the component-change magnitudes represented equally probable clinical variations; this is appropriate since we selected the test magnitudes based on typical clinical values reported in the literature. We consider this preferable to normalizing to the change per degree or millimeter of component positioning since different levels of variability occur for each component and we wished to reflect standard clinical practice. However, we excluded the 15° patellar resection angle from the ordering because, although it can occur, it is a more unusual case. We also averaged the results at the two levels
904
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
of medialization (2.5 mm and 5.0 mm) to represent a more typical level of medialization. 3. Results Component changes clearly had different levels of influence on both tilt and shift (Figs. 3 and 4). Tilt of the patellar bone remnant was significantly affected by the patellar resection angle, resulting in relative lateral tilt for a medial resection angle and relative medial tilt for a lateral resection angle (P < 0.001). External femoral component rotation caused significant lateral tilt relative to the bone in early flexion (P < 0.005) but had a diminished effect in later flexion. Both tibial component rotation and increased patellar thickness had small and/or inconsistent influences on tilt.
Shift of the patellar bone remnant was significantly affected by 5 mm medialization of the patellar component: medialization of the component caused lateral shift of the patellar bone remnant (P < 0.001). Shift was also more lateral following external femoral component rotation (P < 0.001, for early weightbearing flexion). Shift was affected by all patellar resection angles in late flexion (P < 0.001), whereas in early flexion, changes in shift were smaller and less consistent. The effect of tibial component rotation was minor in the weightbearing rig and highly inconsistent in the intraoperative rig. The effect of increasing the patellar thickness was negligible. Doubling the amount of medialization roughly doubled the amount of shift: the mean change in shift for 2.5 mm component medialization was 1.9 mm lateral compared to a mean change of 3.9 mm lateral shift for 5.0 mm med-
Fig. 3. Change in tilt of the patellar bone with respect to the femoral bone for 10 variations in component placement, relative to the baseline tilt at 15° and 90° flexion. The 10 variations were: 5° internal and external femoral component rotation, 5° internal and external tibial component rotation, 7.5° medial, 7.5° lateral and 15° lateral patellar resection angle, 2.5 and 5.0 mm patellar component medialization and 3 mm additional patellar thickness. Significant differences from baseline (P < 0.005) are marked with an asterisk. The horizontal lines (—) refer to the geometrically-predicted tilt of the patellar bone if the tilt of the patellar component were to remain constant with respect to the femoral groove. The difference between this line and the measured value represents tilt of the patellar component relative to the femoral component. Lateral tilt is positive.
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
905
Fig. 4. Change in shift of the patellar bone with respect to the femoral bone for 10 variations in component placement, relative to the baseline shift at 15° and 90° flexion. The 10 variations were: 5° internal and external femoral component rotation, 5° internal and external tibial component rotation, 7.5° medial, 7.5° lateral and 15° lateral patellar resection angle, 2.5 and 5.0 mm patellar component medialization and 3 mm additional patellar thickness. Significant differences from baseline (P < 0.005) are marked with an asterisk. The horizontal lines (—) refer to the geometrically-predicted shift of the patellar bone if the shift of the patellar component were to remain constant with respect to the femoral groove. The difference between this line and the measured value represents shift of the patellar component relative to the femoral component. Lateral shift is positive.
ialization, with an overall ratio of 2.0 (SD, 0.3). Doubling the amount of medialization also roughly doubled the amount of tilt: the mean change in tilt for 2.5 mm component medialization was 1.7° lateral compared to a mean change of 3.2° lateral tilt for 5.0 mm medialization, with an overall ratio of 1.9 (SD, 0.2). For the resection angle, changes in tilt were proportional to both the direction and size of the resection angle: the medial:lateral ratio was 1.0 (SD, 0.2), while the 15°:7.5° ratio was 2.0 (SD, 0.3), for flexion angles greater than 15°. By contrast, the effect on shift was disproportionately large as the size of the resection angle increased: the change in shift due to a 15° resection angle was 2.6 (SD, 0.5) times greater than for the 7.5° resection angle (for 30° to 90° flexion). While there were some variations in the patterns between the intraoperative and postoperative rigs, these differences were not statistically significant (P > 0.05).
Correlations between the intraoperative and weightbearing kinematics are reported elsewhere (Anglin et al., 2008). Ordering the component variable effects (Table 1) revealed that the variables that produced the largest changes in patellar tracking were: patellar component medialization (PM), patellar resection angle (PA) and femoral component rotation (FR). Their relative order depended on the structure of interest (bone or component), kinematic parameter (tilt or shift) and flexion angle (early or late flexion). Tibial component rotation (TR) and patellar thickness (PT) consistently ranked last. 4. Discussion This study was unique because it made direct comparisons regarding the impact of the most important surgeon-controlled component variables on patellar
906
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
Table 1 Approximate ordering of the effects of the component variables on tilt and shift of the patellar bone and patellar component in early and late flexion Effects on: Patellar Bone Relative to Femoral Bone
TILT Early Late Flexion Flexion PA FR (PM) (TR) (PT)
PA FR PM (TR) (PT)
SHIFT Early Late Flexion Flexion FR PM (PA) (TR) (PT)
PM PA (FR) (TR) (PT)
R A N K I N G
1 2 3 4 5
Effects on: Patellar Component Relative to Femoral Component
TILT Early Late Flexion Flexion PM PA (FR) (TR) (PT)
FR PM PA (TR) (PT)
SHIFT Early Late Flexion Flexion PM (FR) (PA) (TR) (PT)
PM FR PA (TR) (PT)
Variables higher in the list had a greater or more consistent impact on the results than those lower in the list. Variables in brackets had a minimal or variable effect. Abbreviations are: PA = patellar resection angle, PM = patellar component medialization, FR = femoral component rotation, TR = tibial component rotation, PT = additional patellar thickness. (The order presented is for the weightbearing rig. The only differences in the intraoperative rig were that: the effect of PM for late-flexion bone remnant tilt exceeded FR; the effects of PA on component tilt and TR on component shift were greater; and the effects of femoral component rotation were reduced.)
kinematics, using the same specimens and protocol. It was also unique because: (1) we assessed this impact using simulations of both intraoperative and weightbearing mechanics; (2) we investigated the effect of the patellar resection angle; (3) we studied two different levels of medialization and resection angle; and (4) we determined an approximate ordering of the effects of variations in component positioning that might occur in clinical practice, which has never been possible previously. Changes to component placement largely had the predicted geometric effect on the patellar bone remnant (represented by the horizontal lines in Figs. 3 and 4). This indicates that, on average, the tilt and shift of the patellar component remained constant within the femoral groove to within a few degrees and a few millimeters, especially in the weightbearing rig; individual specimens showed greater differences. Discrepancies between the geometrically-predicted tilt or shift and the measured values suggest the influence of the soft tissues on patellar kinematics. 4.1. Neutral alignment kinematics (0°, 0 mm) The baseline kinematics in this study were similar in magnitude to those reported previously with arthroplasty components (Armstrong et al., 2003; Chew et al., 1997; Hsu et al., 1996; Jenny et al., 2002; Lee et al., 1997; Miller et al., 2001a; Rhoads et al., 1990). Kinematic patterns vary considerably among studies, however, in part due to the natural variability among individuals and in part due to differences in experimental protocol (Katchburian et al., 2003). In the present study, the patella generally shifted and tilted laterally in early flexion, approaching neutral shift and tilt in later flexion. Since this is the first study sim-
ulating intraoperative flexion, there are no studies available for comparison. Absolute patellar kinematics may be affected by a variety of factors. We minimized the impact of these factors by focusing our results and conclusions on the effects of changes in component position. 4.2. Femoral component rotation (FR) External femoral component rotation had less effect on patellar tilt and shift at 90° as compared to 15°, which is consistent with the geometry: although the femoral groove is deeper at 90°, femoral component rotation at this flexion angle occurs almost perpendicular to the surface of the patella. In fact, in deeper flexion (as seen in the intraoperative rig), the effects on shift of internal and external femoral rotation crossed over and were opposite to those in early flexion because the trochlea itself rotates in the opposite direction relative to the central femoral axis (Brimacombe, 2006). In contrast, internal femoral component rotation produced similar or greater tilt at 90° compared to 15° in the weightbearing rig, suggesting that the soft tissues contribute more to patellar kinematics at 90° than joint geometry. The direction and magnitude of the changes in tilt and shift that we measured were similar to those found in previous cadaveric studies (Anouchi et al., 1993; Armstrong et al., 2003; Miller et al., 2001a,b; Rhoads et al., 1990), although it is difficult to compare results directly due to differing experimental protocols. A CT-based clinical study (Matsuda et al., 2001) found a correlation between femoral component rotation and patellar tilt at 30° knee flexion, consistent with our early-flexion results. An intraoperative study (Akagi et al., 1999) found that externally rotating the
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
femoral component by 3–5° improved patellar tracking and reduced the need for lateral retinacular release; this is consistent with our finding of a lateral shift of the patellar bone remnant in early flexion, which would reduce the Q-angle and reduce the tension in the lateral retinaculum. While this study shows that femoral component rotation could be used to improve patellar tracking, the additional effects of femoral component rotation on tibiofemoral mechanics limit the degree to which a surgeon can exploit this feature. 4.3. Tibial component rotation (TR) Tibial component rotation does not appear to be a useful technique for improving patellar tracking. The only other cadaveric study of tibial component rotation (Nagamine et al., 1994) found no significant effect on patellar tilt or shift except with 15° of external component rotation. In an in vivo imaging study (Matsuda et al., 2001), a significant correlation was reported between tibial component rotation (averaging 17° internal) and patellar tilt (averaging 3° lateral); however this amount of tibial component rotation is very large, the degree of tilt was small and the correlation was weaker than for femoral rotation. They found no relationship between tibial component rotation and patellar shift. Internal rotation of the tibial component and combined internal rotation of the tibial and femoral components have been linked to anterior knee pain (Barrack et al., 2001). There are two possible explanations of the discrepancy between this link and the low ranking in our study: (1) internal component rotation may occur in concert with other surgical errors, or (2) tibial component rotation may have a large impact on some knees but not on others, as evidenced by the considerable variability that we found in intraoperative shift. Putting these results in the context of our study, which investigates the factors under the surgeon’s control to affect patellar tracking, the effects of tibial component rotation cannot be used predictably during surgery, and the impact appears to be minimal during weightbearing.
907
(bone plus component) back in the lateral direction. This effect may be larger in a patient with a tight lateral retinaculum or a subluxing patella. The higher effect intraoperatively suggests that the retinacula indeed play an important role, although their role appears to be reduced under weightbearing. Our findings are consistent with the trends reported in clinical studies (Kawano et al., 2002; Chan and Gill, 1999; Gomes et al., 1988), i.e. that medial underresection causes lateral tilt of the bone and medial tilt of the component. Some authors (Chan and Gill, 1999; Kawano et al., 2002) have suggested that limited underresection of the medial side may be beneficial, whereas others (Ledger et al., 2005) have concluded that cut angle had little impact on tilt; the difference may be partially a judgment on how much is an important amount. One group (Chan and Gill, 1999) has argued that a thicker medial side mimics the original shape more closely; another (Kawano et al., 2002) has argued that since the patella tilts laterally in the natural knee, a 5° oblique cut would therefore result in neutral tilt of the patellar component. None of the clinical studies observed a noticeable effect on shift relative to the femoral component (Chan and Gill, 1999; Gomes et al., 1988; Kawano et al., 2002). We likewise found a minimal effect for the 7.5° resection angles; however component shift for the 15° resection angle was large and significant. Anterior knee pain has been reported to be more prevalent in patients with asymmetric patellar resurfacing compared to those with symmetric resurfacing (Baldini et al., 2006, 2007; Pagnano and Trousdale, 2000). There are many possible sources for the pain, including bony impingement on the femoral component from excessive tilt, especially under load (Baldini et al., 2006, 2007), poor implant design, elevated soft tissue strains or other errors in surgical technique. Underresection is also generally associated with a thicker patellar construct (Ledger et al., 2005), which may lead to overstuffing of the joint. A further disadvantage that we found is that the effect on shift was disproportionate to the resection angle, indicating that the consequences of a poor resection angle were increasingly severe as the resection angle increased. These results suggest that excessive obliquity should be avoided.
4.4. Patellar resection angle (PA) 4.5. Patellar medialization (PM) Achieving a desired patellar cut during total knee arthroplasty is difficult due to the small, hard bone and limited visibility. Cut asymmetry is common (Baldini et al., 2006, 2007; Chan and Gill, 1999; Gomes et al., 1988; Kawano et al., 2002; Ledger et al., 2005; Pagnano and Trousdale, 2000) even though most surgeons strive for equal medial and lateral thicknesses. The large, significant impact that the patellar resection angle had on patellar bone tilt was geometrically expected. The result, e.g. medial tilt of the patellar bone for a lateral wedge, would tend to increase tension in the opposing (in this case, lateral) soft tissue restraints, which may explain the small but consistent rotation of the whole construct
Lateral shift of the patellar bone due to medialization of the patellar component likely reduces tension in the lateral retinaculum, which may reduce the need for lateral release (Hofmann et al., 1997; Lewonowski et al., 1997). Lateral tilt and shift of the patellar bone due to component medialization is consistent with previous biomechanical studies (Lee et al., 1999; Miller et al., 2001b; Yoshii et al., 1992) and clinical studies (Chan and Gill, 1999; Lewonowski et al., 1997; Nelissen et al., 1995). The lateral tilt, which has not previously been explained, is likely induced by the lateral displacement of the quadriceps attachment relative to the patellar component, which thereby creates a
908
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
moment about the patellar component (Brimacombe, 2006). Slight medialization, on the order of 2.5 mm, may offer a balance between the desirable reduction in lateral subluxation and the undesirable increase in lateral tilt, which could result in bony impingement. The proportionality of changes in tilt and shift with medialization means that there was no sudden change in the consequences of medialization; it also means that the kinematic effects are reasonably predictable. The roughly linear relationship that we found suggests that the benefits and drawbacks can be traded off equally. 4.6. Patellar thickness (PT) Similar to our results, a clinical study recently reported no gross, visible effect on patellar tracking in vivo intraoperatively with additional patellar thicknesses up to 8 mm (Bengs and Scott, 2006). In contrast, the one previous biomechanical study determined that increasing patellar thickness increased lateral tilt and medial shift (Hsu et al., 1996). Radiographically, one study observed changes in tilt (Laughlin et al., 1996) whereas another did not (Kawano et al., 2002). The hypothesis that tracking is not affected by changes in patellar thickness, as we found, is supported by the argument that the Q-angle and patellofemoral joint geometry have not been altered in the coronal plane. However, these results may be different in the presence of abnormally tight soft tissues, e.g. due to a tight lateral retinaculum in the valgus knee, or due to overstuffing of the joint by an oversized femoral component. The main disadvantage appears to be a reduced range of motion (RoM): the intraoperative clinical study described above revealed a 3° reduction in RoM for every 2 mm of additional patellar thickness (Bengs and Scott, 2006). 4.7. Approximate ordering of effects of component placement variations Three component variables (patellar component medialization, patellar resection angle and femoral component rotation) consistently produced the largest effects on patellar kinematics, although the ordering among these factors depended on whether the structure of interest was the patellar bone or component, whether the kinematic parameter was tilt or shift and whether it was early or late flexion (see Table 1). If the greatest clinical concern is subluxation of the patellar component, then the surgeon should focus on patellar component medialization. If, on the other hand, tilt of the patellar bone remnant, which can affect retinacular tensions or cause impingement, is the greater clinical concern, then the surgeon should focus on the patellar resection angle. 4.8. Limitations The primary limitation of this study is that of any cadaveric study, i.e. test rigs do not completely replicate
physiological loading in vivo. Nevertheless the consistency in results relative to baseline between the intraoperative and weightbearing simulations and the clear distinctions between component effects support the general conclusions. Furthermore, a parametric study such as we have performed (i.e. comparing the effects of different component positions on the same specimens) is only possible ex vivo. Unlike clinical studies, a biomechanical study has the advantage of being able to isolate the effects of particular component variables. We were also able to investigate kinematics throughout the range of motion, in contrast to most clinical studies, which only report results at a single static flexion angle. The effects of any discrepancies in absolute kinematics between the ex vivo model and in vivo kinematics were minimized by our focus on relative effects, i.e. changes in kinematics due to changes in component position. A second limitation, common to almost all cadaveric studies, is that we did not measure tilt and shift of the patellar component relative to the femoral component directly; most studies do not mention it at all. We calculated these indirectly by subtracting the measured bone kinematics from the geometrically-predicted values. Both the boneto-bone and the component-to-component values are important. Bone-to-bone values are of interest because tilt and shift can affect the medial and lateral soft tissue tensions and because these values correspond to what the surgeon can observe most easily during the intraoperative assessment. Component-to-component values are important because they relate to component longevity and can be compared to postoperative radiographic analyses. A third limitation is that we only studied a single component design. Due to the specific bone cuts made to accommodate the components, the need to machine modified components, and the practical limitations on the number of trials per specimen, it would not have been possible to repeat these measurements with a second design that had different geometry or tibiofemoral rotational freedom. It is possible that the relatively deep trochlear groove of the NexGen design captured the patella more than some other designs; however recognition of the importance of a deeper groove for patellar tracking (Chew et al., 1997; Yoshii et al., 1992) means that this is present in most current designs. The angled orientation of the NexGen femoral groove, and the marginally higher lateral flange, could result in different absolute kinematics than for a symmetric component, but these differences are unlikely to affect the ordering of the effects of changes in component placement, especially given that the proximal groove is shallow. A biomechanical study of three different component designs, including the NexGen, did not reveal significant differences among the designs (Chew et al., 1997). Likewise, a fluoroscopic study (Stiehl et al., 2001) did not detect significant differences between posterior-stabilized and cruciate-retaining designs in sagittal plane patellar kinematics. In our opinion, therefore, although the absolute kinematics and the precise relative results (e.g. Figs. 3 and 4) may differ
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
subtly among designs, we expect that the overall findings regarding which component changes have the greatest impact on patellar kinematics will be valid across designs. A fourth limitation is that cadaveric studies only represent the immediate postoperative period. Two studies (Laughlin et al., 1996; Shih et al., 2004) have recorded increasingly lateral tracking over time; this future change should be considered when improving tracking intraoperatively. 5. Conclusions To improve patellar tracking, the surgeon should focus on patellar medialization (aiming for slight medialization), patellar resection angle (aiming for symmetry) and femoral component rotation (aiming for anatomic reconstruction, favouring external rotation). The choice of which component position to change for the greatest impact depends on whether the change should primarily affect tilt or shift of the component or the bone remnant, in late flexion or early flexion. The ordering table is unique in the literature and was only possible by testing numerous component variables on each specimen. The patellar resection angle, which has not previously been studied, ranked highly against the other component variables, suggesting that it deserves more attention than it has previously received, especially given its link to anterior knee pain. The consistency in component placement effects between the weightbearing rig and the simulated intraoperative assessment, which has not previously been reported, supports the use of intraoperative assessment of component changes to improve postoperative mechanics. Reducing patellar maltracking could lead to an improved clinical outcome by reducing anterior knee pain, patellar fracture, component wear and component loosening. Acknowledgements We wish to thank the Natural Sciences and Engineering Research Council of Canada (NSERC), Praxim (Grenoble, France), the Canadian Arthritis Network, and the Michael Smith Foundation for Health Research for their generous financial support, and Zimmer Canada for the components and use of the instruments. We would also like to thank Karen Ho for her assistance in creating the experimental apparatus schematics. References Akagi, M., Matsusue, Y., Mata, T., Asada, Y., Horiguchi, M., Iida, H., Nakamura, T., 1999. Effect of rotational alignment on patellar tracking in total knee arthroplasty. Clin. Orthop. Relat. Res. 366, 155–163. Anglin, C., Brimacombe, J.M., Wilson, D.R., Masri, B.A., Greidanus, N.V., Tonetti, J., Hodgson, A.J., 2008. Intraoperative vs. weightbearing patellar kinematics in total knee arthroplasty: a cadaveric study. Clin. Biomech. 23, 60–70. Anouchi, Y.S., Whiteside, L.A., Kaiser, A.D., Milliano, M.T., 1993. The effects of axial rotational alignment of the femoral component on knee
909
stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens. Clin. Orthop. Relat. Res. 287, 170–177. Armstrong, A.D., Brien, H.J.C., Dunning, C.E., King, G.J.W., Johnson, J.A., Chess, D.G., 2003. Patellar position after total knee arthroplasty. J. Arthroplasty 18 (4), 458–465. Baldini, A., Anderson, J.A., Cerulli-Mariani, P., Kalyvas, J., Pavlov, H., Sculco, T.P., 2007. Patellofemoral evaluation after total knee arthroplasty: validation of a new weight-bearing axial radiographic view. J. Bone Joint Surg. Am. 89 (8), 1810–1817. Baldini, A., Anderson, J.A., Zampetti, P., Pavlov, H., Sculco, T.P., 2006. A new patellofemoral scoring system for total knee arthroplasty. Clin. Orthop. Relat. Res. 452, 150–154. Barrack, R.L., Burak, C., 2001. Patella in total knee arthroplasty. Clin. Orthop. Relat. Res. 389, 62–73. Barrack, R.L., Schrader, T., Bertot, A.J., Wolfe, M.W., Myers, L., 2001. Component rotation and anterior knee pain after total knee arthroplasty. Clin. Orthop. Relat. Res. 392, 46–55. Bengs, B.C., Scott, R.D., 2006. The effect of patellar thickness on intraoperative knee flexion and patellar tracking in total knee arthroplasty. J. Arthroplasty 21 (5), 650–655. Brimacombe, J.M., 2006. Effects of component placement on patellar kinematics and loading in intraoperative and postoperative loading configurations. M.A.Sc. thesis, University of British Columbia, Vancouver. Chan, K.C., Gill, G.S., 1999. Postoperative patellar tilt in total knee arthroplasty. J. Arthroplasty 14 (3), 300–304. Chew, J.T.H., Stewart, N.J., Hanssen, A.D., Zong-Ping, L., Rand, J.A., An, K-N., 1997. Differences in patellar tracking and knee kinematics among three different total knee designs. Clin. Orthop. Relat. Res. 345, 87–98. Cole, G.K., Nigg, B.M., Ronsky, J.L., Yeadon, M.R., 1993. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization proposal. J. Biomech. Eng. 115, 344–349. Ezzet, K.A., Hershey, A.L., D’Lima, D.D., Irby, S.E., Kaufman, K.R., Colwell, C.W., 2001. Patellar tracking in total knee arthroplasty – inset versus onset design. J. Arthroplasty 16 (7), 838–843. Gomes, L.S.M., Bechtold, J.E., Gustilo, R.B., 1988. Patellar prosthesis positioning in total knee arthroplasty. Clin. Orthop. Relat. Res. 236, 72–81. Hofmann, A.A., Tkach, T.K., Evanich, C.J., Camargo, M.P., Zhang, Y., 1997. Patellar component medialization in total knee arthroplasty. J. Arthroplasty 12 (2), 155–160. Hsu, H.-C., Zong-Ping, L., Rand, J.A., An, K.-N., 1996. Influence of patellar thickness on patellar tracking and patellofemoral contact characteristics after total knee arthroplasty. J. Arthroplasty 11 (1), 69– 80. Jenny, J.Y., Lefebvre, Y., Vernizeau, M., Lavaste, F., Skalli, W., 2002. In vitro analysis of the continuous active patellofemoral kinematics of the normal and prosthetic knee. Rev. Chir. Orthop. Re´par. l’App. Mot. 88 (8), 797–802. Katchburian, M.V., Bull, A.M.J., Shih, Y.-F., Heatley, F.W., Amis, A.A., 2003. Measurement of patellar tracking: assessment and analysis of the literature. Clin. Orthop. Relat. Res. 412, 241–259. Kawano, T., Miura, H., Nagamine, R., Urabe, K., Matsuda, S., Takaaki, M.-O., Iwamoto, Y., 2002. Factors affecting patellar tracking after total knee arthroplasty. J. Arthroplasty 17 (7), 942–947. Laughlin, R.T., Werries, B.A., Verhulst, S.J., Hayes, J.M., 1996. Patellar tilt in total knee arthroplasty. Am. J. Orthop. 23, 300–304. Ledger, M., Shakespeare, D., Scaddan, M., 2005. Accuracy of patellar resection in total knee replacement: a study using the medial pivot knee. Knee 12, 13–19. Lee, T.Q., Budoff, J.E., Glaser, F.E., 1999. Patellar component positioning in total knee arthroplasty. Clin. Orthop. Relat. Res. 366, 274– 281. Lee, T.Q., Gerken, A.P., Glaser, F.E., Kim, W.C., Anzel, S.H., 1997. Patellofemoral joint kinematics and contact pressures in total knee arthroplasty. Clin. Orthop. Relat. Res. 340, 257–266.
910
C. Anglin et al. / Clinical Biomechanics 23 (2008) 900–910
Lewonowski, K., Dorr, L.D., McPherson, E.J., Huber, G., Wan, Z., 1997. Medialization of the patella in total knee arthroplasty. J. Arthroplasty 12 (2), 161–167. Ma, H.M., Lu, Y.C., Kwok, T.G., Ho, F.Y., Huang, C.Y., Huang, C.H., 2007. The effect of the design of the femoral component on the conformity of the patellofemoral joint in total knee replacement. J. Bone Joint Surg. Br. 89, 408–412. Matsuda, S., Miura, H., Nagamine, R., Urabe, K., Hirata, G., Iwamoto, Y., 2001. Effect of femoral and tibial component position on patellar tracking following total knee arthroplasty: 10-year follow-up of MillerGalante I knees. Am. J. Knee Surg. 14, 152–156. Miller, M.C., Berger, R.A., Petrella, A.J., Karmas, A., Rubash, H.E., 2001a. Optimizing femoral component rotation in total knee arthroplasty. Clin. Orthop. Relat. Res. 392, 38–45. Miller, M.C., Zhang, A.X., Petrella, A.J., Berger, R.A., Rubash, H.E., 2001b. The effect of component placement on knee kinetics after arthroplasty with an unconstrained prosthesis. J. Orthop. Res. 19 (4), 614–620. Nagamine, R., Whiteside, L.A., White, S.E., McCarthy, D.S., 1994. Patellar tracking after total knee arthroplasty: the effect of tibial tray malrotation and articular surface configuration. Clin. Orthop. Relat. Res. 304, 263–271. Nelissen, R.G.H.H., Weidenheim, L., Mikhail, W.E.M., 1995. The influence of the position of the patellar component on tracking in total knee arthroplasty. Int. Orthop. 19, 224–228.
Pagnano, M.W., Trousdale, R.T., 2000. Asymmetric patella resurfacing in total knee arthroplasty. Am. J. Knee Surg. 13, 228–233. Rand, J.A., 2004. Extensor mechanism complications following total knee arthroplasty. J. Bone Joint Surg. Am. 86A (9), 2062– 2072. Rhoads, D.D., Noble, P.C., Reuben, J.D., Mahoney, O.M., Tullos, H.S., 1990. The effect of femoral component position on patellar tracking after total knee arthroplasty. Clin. Orthop. Relat. Res. 260, 43–51. Shih, H.-N., Shih, L.-Y., Wong, Y.-C., Hsu, R.W.-W., 2004. Long-term changes of the nonresurfaced patella after total knee arthroplasty. J. Bone Joint Surg. Am. 86, 935–939. Stiehl, J.B., 2005. A clinical overview: patellofemoral joint and application to total knee arthroplasty. J. Biomech. 38, 209–214. Stiehl, J.B., Komistek, R.D., Dennis, D.A., Keblish, P.A., 2001. Kinematics of the patellofemoral joint in total knee arthroplasty. J. Arthroplasty 16 (6), 706–714. Whiteside, L.A., Nakamura, T., 2003. Effect of femoral component design on unresurfaced patellas in knee arthroplasty. Clin. Orthop. Relat. Res. 410, 189–198. Yoshii, I., Whiteside, L.A., Anouchi, Y.S., 1992. The effect of patellar button placement and femoral component design on patellar tracking in total knee arthroplasty. Clin. Orthop. Relat. Res. 275, 211–219.