Introduction

While one of the main goals of total knee arthroplasty (TKA) is to restore normal knee function, sagittal plane kinematics after TKA are known to differ from normal knee kinematics [1,2]. As a result, normative and TKA knee joint kinematics have been studied extensively using a wide range of instrumentation and methodologies [3]. One such technique is motion analysis, which provides numerous advantages to traditional imaging methods. Motion analysis allows researchers to noninvasively quantify knee kinematics in TKA and normative participants during functional, weightbearing tasks, including gait [4], sit-to-stand [5], and stair climbing [6]. In addition, motion analysis permits the measurement of the patellofemoral joint across a range of knee flexion angles. The major limitation of this technique is soft tissue movement, which may cause errors in the estimation of the position and orientation of the underlying bone [7]. As a result, few studies have attempted to track sagittal tendon geometry after TKA using this method.

Measurement of sagittal knee tendon angles is an established method for assessing normal and TKA joint kinematics [8-10]. These tendons play a critical role in determining the orientation of the patella and provide an indirect assessment of knee function. This paper uses a two-dimensional model of the lower extremity to describe movements of knee structures in the sagittal plane of individuals with TKA. Using motion capture techniques, the relative positions of the patellar tendon and quadriceps tendon, as a function of knee flexion-extension, were estimated during step down and bilateral deep knee bend tasks. Data were compared to an age-matched normative group. To assess the validity of this technique, motion capture data were compared to radiographic data for each TKA patient.

Methods

Participants

Instrumentation/Apparatus

Procedures

Data collection occurred during routine clinical evaluations at the Orthopaedics Clinic at the Dr. Everett Chalmers Hospital. A physical exam and radiographs are a routine part of this evaluation. Prior to the radiographs, the examining surgeon placed radio-opaque markers over specific anatomical landmarks on both lower extremities. These landmarks were the: 1) mid-femur, 2) lateral femoral epicondyle, 3) tibial condyle, 4) tibial tubercle, 5) distal tibia, 6) superior pole of patella, and 7) inferior pole of patella (Figure 1a). Each patient then had lateral weight-bearing radiographs taken with their knees slightly flexed.

Figure 1

Figure 1 – a) Sagittal view of participant with reflective markers performing the step down task, b) a lateral weight-bearing radiograph of a total knee arthroplasty with radio-opaque markers.

Immediately after completion of the radiographs, patients were asked to enter a separate room containing the Vicon 512 motion capture system. Reflective markers were applied directly overtop of the radio-opaque markers (Figure 1b). The motion capture system was used to track the locations of these markers during weight bearing bilateral knee bend and stair descent tasks. A 20cm high step with a support handle was positioned within the calibrated workspace for the stair descent task. The order of tasks was randomized across individuals. Both activities produced high loads at the knee joint and are typical activities of daily living. Patients were asked to perform 6 bilateral knee bends to the best of their ability. Patients were permitted to position their arms in front of their trunk to aid with balance. For the stair descent task, patients were asked to perform 3 trials with the left leg leading off the step, and 3 trials with the right left leading off the step. If necessary, patients were permitted to hold onto the support rail for stability.

Data Analysis

Results

TKA Versus Age-matched Normative Data

The results for the bilateral deep knee bend and the step down task were very similar. Therefore, only the deep knee bend data is provided here. As the knee flexed from 30º-70º during the bilateral deep knee bend, the orientation of the patellar tendon was calculated relative to the longitudinal axis of the patella (Figure 2) and relative to the longitudinal axis of the tibia (Figure 3). For both the TKA and normative group, the mean range of motion of the patellar tendon angle with respect to the patella was relatively small. This indicates that the tendon maintains a relatively fixed position with respect to the patella in the sagittal plane. Age-matched data was more variable than the TKA data and showed differences in amplitude for the patellar tendon angle (Figure 2). The mean relative angle between the tibia and the patellar tendon did not change as expected in the TKA group as a function of knee flexion (Figure 3). The patellar tendon angle was directed anteriorly throughout the task. At higher values of knee flexion, the tendon will be directed posteriorly, yet this transition is not apparent by 70º of knee flexion. The normative group demonstrated a more acceptable pattern of motion, showing a negative slope with increasing knee flexion.

Figure 2

Figure 2a – Mean ± one standard deviation for patellar tendon and patellar axis in the TKA group

Figure 3

Figure 2b – Mean ± one standard deviation for patellar tendon and patellar axis in the normative group

Figure 4

Figure 3a – Mean ± one standard deviation for patellar tendon and tibial axis for TKA group

Figure 5

Figure 3b – Mean ± one standard deviation for patellar tendon and tibial axis

In the TKA group, the relative angle between the longitudinal axes of the femur and patella (an approximation of the quadriceps tendon) increased linearly from 27º to 52º during flexion of the knee from 30º-70º (Figure 4a). Similar ranges and pattern of motion were noted in the age-matched normative data (Figure 4b). However, the normative group showed lower amplitudes and higher variability for this angle compared to the TKA group.

The midpoint of the patella translated an average of 1.5 cm posteriorly as the knee flexed in the TKA group. Similar results were noted for the normative group (average of 1.3 cm). Again, the TKA group showed more consistent results than the normative group (Figure 5).

Validation of Motion Capture Data

Comparisons of radiographic and motion capture results for patellar tendon angle (PA/PT) and quadriceps angle (PA/FEM) showed good agreement for both step down and knee bend tests. Average absolute differences in angles between the two methods for the step down task were 2.9º (± 0.1º) for PA/PT and 5.6º (± 0.2º) for PA/FEM. For the bilateral deep knee bend task, the average absolute differences in angles were 2.6º (±0.1º) for PA/PT and 5.9º (±0.2º) for PA/FEM. Radiographic and motion capture data for a typical patient is shown in Figure 6.

Figure 6

Figure 4a – Mean ± one standard deviation for quadriceps tendon and patellar axis for TKA group

Figure 7

Figure 4b – Mean ± one standard deviation for quadriceps tendon and patellar axis for Normative group

Figure 8

Figure 5a – Mean ± one standard deviation for anteroposterior patellar displacement for TKA group

Figure 9

Figure 5b – Mean ± one standard deviation for anteroposterior patellar displacement for Normative group

Figure 10

Figure 6. Comparisons of radiographic (*) and motion capture tendon angle data (+) versus knee flexion for one participant.

Discussion

Tracking the patellar and quadriceps tendon angles, as they vary with knee flexion, provides a simple assessment of knee function. This study examined sagittal plane knee kinematics during functional, weight-bearing tasks using motion capture techniques.

For both groups, the patellar tendon angles were similar to those reported by Van Eijden et al. (1986), demonstrating a relatively fixed angle with respect to the longitudinal axis of the patella throughout knee flexion. Previous research has shown that, with respect to the tibia, the patellar tendon is typically directed anteriorly during the first 70º of knee flexion and posteriorly for the remainder of flexion [8,10]. While the normative group shows an anteriorly directed patellar tendon that is gradually shifting towards neutral as knee flexion increases, the TKA group did not show this pattern. Instead, the patellar tendon remained in a fairly consistent anterior position throughout knee flexion.

For this study, it was assumed that the longitudinal axis of the femur approximated the orientation of the quadriceps tendon during knee flexion. Previous research has shown that during 30º-70º of knee flexion, the quadriceps tendon maintains a relatively fixed angle with respect to the longitudinal axis of the femur [8]. In both groups, the mean relative angle between the longitudinal axes of the patella and femur increased linearly with flexion of the knee (Figure 4). For the normative group, results were similar to Van Eijden et al. (1986), despite differences in knee load between the studies. The TKA group showed a similar pattern and range of motion to the normative data; however, the amplitude of the signal was higher suggesting a different tendon orientation.

Linear displacements of the patella demonstrated a similar pattern of motion across the two groups (Figure 5). Higher variability is noticeable as the knee approached extension for both TKA and normative data. Results for this parameter were similar to those provided by Van Eijden et al. (1986).

The second objective of this study was to validate the use of a Vicon motion capture system to measure sagittal knee tendon angles. The motion capture system used in this study was a passive optoelectronic device that used infrared cameras to track skin-mounted reflective markers as the person moved. Average absolute differences between the two methods revealed that the motion capture system was able to approximate the patella tendon angles more accurately than the quadriceps tendon angle for both tasks. Average absolute differences were 2.9º (±0.1º) for PA/PT and 5.6º (±0.2º) for PA/FEM for the step down task. For the bilateral deep knee bend task, the average absolute differences in angles were 2.6º (±0.1º) for PA/PT and 5.9º (±0.2º) for PA/FEM.

The larger error for PA/FEM for both motor tasks is likely due to several reasons. First, the motion capture method used the locations of markers on the mid-femur marker and lateral femoral condyle to estimate the quadriceps tendon angle. However, the mid-femur marker was not always visible on the radiographs of each patient. In these cases, a more proximal femur point of reference was used to measure the PA/FEM angle from the radiographs. As a result, small errors in the angle measures may have been generated. Another factor that may have affected the error results was that the motion capture method used the three-dimensional coordinates of the markers to compute angles. For the radiographic images, if the patient was not aligned properly during the procedure, the measured projection angles would be smaller than the true angle. While the radiographs were subjectively examined to ensure appropriate alignment of the patient, small errors could result from minor malalignment of the patient. Finally, errors may have also occurred through movement of the mid-femur marker relative to the underlying bone during the performance of the motor tasks.

The kinematic model used in this study was limited to the sagittal plane and does not address the other planes of motion for the knee joint or the Q-angle. However, the robust sagittal measures provide a reliable indicator of knee function. Due to subject differences in range of motion, data was cropped to common knee flexion-extension ranges. While the data could have been modeled to provide estimates of tendon angles at certain knee joint angles, we opted for a conservative approach that used only the available raw data.

Conclusions

This technique provides a method to measure patellar movements after TKA with results comparable to radiographs. Motion capture techniques permit data capture during functional, weight-bearing activities, thereby providing information on knee motion across a range of flexion angles. The results of this study suggest that motion capture is a valid and reliable technique for assessing sagittal plane knee tendon geometry. Further research aims to develop three-dimensional mechanical models of the knee to assess movements in individuals with TKA.

Acknowledgments