Robot-aided in vitro measurement of patellar stability with consideration to the influence of muscle loading

Abstract

Background: Anterior knee pain is often associated with patellar maltracking and instability. However, objective measurement of patellar stability under clinical and experimental conditions is difficult, and muscular activity influences the results. In the present study, a new experimental setting for in vitro measurement of patellar stability was developed and the mediolateral force-displacement behavior of the native knee analyzed with special emphasis on patellar tilt and muscle loading.

Methods: In the new experimental setup, two established testing methods were combined: an upright knee simulator for positioning and loading of the knee specimens, and an industry robot for mediolateral patellar displacement. A minimally invasive coupling and force control mechanism enabled unconstrained motion of the patella as well as measurement of patellar motion in all six degrees of freedom via an external ultrasonic motion-tracking system. Lateral and medial patellar displacement were measured on seven fresh-frozen human knee specimens in six flexion angles with varying muscle force levels, muscle force distributions, and displacement forces.

Results: Substantial repeatability was achieved for patellar shift (ICC(3,1) = 0.67) and tilt (ICC(3,1) = 0.75). Patellar lateral and medial shift decreased slightly with increasing flexion angle. Additional measurement of patellar tilt provided interesting insights into the different displacement mechanisms in lateral and medial directions. For lateral displacement, the patella tilted in the same (lateral) direction, and tilted in the opposite direction (again laterally) for medial displacement. With regard to asymmetric muscle loading, a significant influence (p < 0.03, up to 5 mm shift and 8° tilt) was found for lateral displacement and a reasonable relationship between muscle and patellar force, whereas no effect was visible in the medial direction.

Conclusion: The developed experimental setup delivered reproducible results and was found to be an excellent testing method for the in vitro analysis of patellar stability and future investigation of surgical techniques for patellar stabilization and total knee arthroplasty. We demonstrated a significant influence of asymmetric quadriceps loading on patellar stability. In particular, increased force application on the vastus lateralis muscle led to a clear increase of lateral patellar displacement.

Figure 1

Overview of the experimental setup. Arrangement of the knee simulator, KUKA robot, and ultrasonic motion tracking system (ZEBRIS) during execution of the experiments. The coordinate systems of the femur (x_F, y_F, z_F) and patella (x_P, y_P, z_P) are illustrated. Patellar shift is defined as translation of the patella along the mediolateral axis of the femur (z_F) and patellar tilt is defined as rotation of the patella around its proximodistal axis (x_P).

Figure 2

Positioning of patellar screws. We ensured that screws were positioned in one line, with their tips pointing together and to the center of the patella. The CT scan shows a transverse section of a typical patella at the screw level.

Figure 3

Details of the experimental setup. a Mechanism for fixation of the femur to the testing rig in an arbitrary position. A threaded bar was screwed through the femoral shaft and secured to a multiple adjustable frame (yellow double-arrows indicate possibilities for adjustment) using two ball joints. The tibia was not constrained further. b Robotic patellar attachment via a tip into a torx screw. The two tips are for medial and lateral displacement, respectively, but only one tip is touching the corresponding patellar screw at a time, thereby allowing unconstrained patellar rotation.

Figure 4

Influence of the quadriceps force level on patellar displacement. Patellar mediolateral shift (top) and tilt (bottom) during lateral (a) and medial (b) force application are illustrated, comparing the four measured levels of quadriceps force (30, 150, 300, 600 N). Mean values and standard deviations are shown for 50 N of mediolateral patellar displacement force and central quadriceps force distribution.

Figure 5

Influence of asymmetric quadriceps loading on patellar displacement for moderate muscle force. Patellar mediolateral shift (top) and tilt (bottom) during lateral (a) and medial (b) force application are illustrated, comparing the influence of three quadriceps force distributions (central, lateral, medial). Mean values and standard deviations are shown for 50 N mediolateral displacement force and a total of 150 N quadriceps force. Significant differences are identified by an asterisk.

Figure 6

Influence of asymmetric quadriceps loading on patellar displacement for high muscle force. Patellar mediolateral shift (top) and tilt (bottom) during lateral (a) and medial (b) force application are illustrated, comparing the influence of three quadriceps force distributions (central, lateral, medial). Mean values and standard deviations are shown for 50 N mediolateral displacement force and a total of 150 N quadriceps force. Significant differences are identified by an asterisk.