Dual-joint modeling for estimation of total knee replacement contact forces during locomotion

• Published in: Journal of biomechanical engineering Vol. 135; no. 2; p. 021013
• Main Authors: Hast, Michael W, Piazza, Stephen J
• Format: Journal Article
• Language: English
• Published: United States 01.02.2013
• Subjects: Aged, 80 and over; Algorithms; Arthroplasty, Replacement, Knee; Biomechanical Phenomena; Humans; Knee Joint - physiology; Knee Joint - surgery; Leg Bones - physiology; Locomotion; Male; Mechanical Phenomena; Models, Biological; Muscles - physiology
• ISSN: 1528-8951
• DOI: 10.1115/1.4023320

Model-based estimation of in vivo contact forces arising between components of a total knee replacement is challenging because such forces depend upon accurate modeling of muscles, tendons, ligaments, contact, and multibody dynamics. Here we describe an approach to solving this problem with results that are tested by comparison to knee loads measured in vivo for a single subject and made available through the Grand Challenge Competition to Predict in vivo Tibiofemoral Loads. The approach makes use of a "dual-joint" paradigm in which the knee joint is alternately represented by (1) a ball-joint knee for inverse dynamic computation of required muscle controls and (2) a 12 degree-of-freedom (DOF) knee with elastic foundation contact at the tibiofemoral and patellofemoral articulations for forward dynamic integration. Measured external forces and kinematics were applied as a feedback controller and static optimization attempted to track measured knee flexion angles and electromyographic (EMG) activity. The resulting simulations showed excellent tracking of knee flexion (average RMS error of 2.53 deg) and EMG (muscle activations within ±10% envelopes of normalized measured EMG signals). Simulated tibiofemoral contact forces agreed qualitatively with measured contact forces, but their RMS errors were approximately 25% of the peak measured values. These results demonstrate the potential of a dual-joint modeling approach to predict joint contact forces from kinesiological data measured in the motion laboratory. It is anticipated that errors in the estimation of contact force will be reduced as more accurate subject-specific models of muscles and other soft tissues are developed.