Following the footsteps: wearable sensors quantify principal component informed kinetic gait outcomes before and after knee arthroplasty.

Abstract

A common feature of knee osteoarthritis (OA) gait is a stiff-knee pattern characterized by minimal range of motion and more constant loading through the stance phase. Previous work has shown that knee joint kinematics and kinetics following arthroplasty surgery often do not return to age-matched control levels for most patients. Quantification of these gait patterns has traditionally relied on laboratory-based optoelectronic motion capture and synchronized force plates, which are costly, spatially constrained, and limit the clinical scalability of relevant gait metrics. The purpose of this study was to establish an inertial measurement unit (IMU) protocol capable of representing clinically relevant features of the knee adduction and flexion moments during gait in patients with knee OA, with the goal of supporting out of laboratory OA kinetic gait assessment. Multivariate relationships between the knee flexion and adduction moment patterns representing 'stiff-knee' kinetic gait patterns and lower extremity accelerations and angular velocities from lower-extremity worn inertial sensors were examined in patients with end-stage knee OA. Shank-mounted IMU models explained significant variance in the pattern of the knee flexion and adduction moment waveforms (R² = 0.55-0.64, p < 0.01), and inclusion of foot-mounted sensors further improved model performance (R² = 0.67-0.69, p < 0.01). Agreement analyses demonstrated minimal bias with moderate limits of agreement between IMU-predicted and motion capture-derived PC2 scores. These findings demonstrated that OA-specific, clinically relevant temporal features of knee moments can be reasonably estimated using IMU data, supporting the potential for real-world OA gait assessment.