EventsThe 2nd International Conference on Bioengineering
Published
This submission belongs to the session S5. Biomechanics and Sports Medicine of the event The 2nd International Conference on Bioengineering
Published date
16 Nov, 2026
Contribution&Funding
Conceptualization,
[Alison Chang]
Methodology,
[Kaitlyn Wei] ; [Jun Yao] ; [Alison Chang] ; [Alison Li]
Software,
[Kaitlyn Wei] ; [Alison Li]
Validation,
[Alison Li] ; [Jun Yao] ; [Alison Chang] ; [Kaitlyn Wei]
Formal analysis,
[Alison Li] ; [Kaitlyn Wei]
Investigation,
[Jun Yao] ; [Alison Li] ; [Kaitlyn Wei]
Resources,
[Alison Chang]
Data curation,
[Alison Chang]
Writing - original draft,
[Kaitlyn Wei] ; [Alison Li]
Writing - review,
[editing] ; [Alison Chang] ; [Kaitlyn Wei]
Supervision,
[Alison Chang] ; [Jun Yao]
Project administration,
[Alison Chang]
Funding acquisition,
[Alison Chang]
Funding: --
Citation
Kaitlyn Wei, Alison Li, Alison Chang, Jun Yao, Contributions of Gait Waveforms and Patient-Reported Knee Pain to Radiographic Knee Osteoarthritis Severity, in Proceedings of The 2nd International Conference on Bioengineering, Barcelona, 10 November–13 November 2026, MDPI: Basel, Switzerland
Share
Email
Facebook
Twitter
LinkedIn

Contributions of Gait Waveforms and Patient-Reported Knee Pain to Radiographic Knee Osteoarthritis Severity

Alison Li 1
Jun Yao 1
1. Physical Therapy and Human Movement Sciences, Northwestern, Chicago, 60611, USA
Abstract

Introduction. Most gait studies in knee osteoarthritis (KOA) quantify biomechanics using discrete variables, discarding much of the information contained within continuous gait waveforms. Which gait waveform features are associated with radiographic disease severity and whether they provide information complementary to self-reported knee pain remain unclear.  

Methods. Functional principal component analysis (fPCA) extracted waveform-derived gait features, elastic-net variable selection identified robust gait features, and linear mixed-effects model (LMM) with dominance analysis (DA) quantified the independent contributions of gait waveform features and knee pain to radiographic KOA severity.  

Results. Among 246 participants with KOA (476 knees), two consensus gait features–knee flexion/extension fPC3 and knee adduction/abduction fPC1–were selected in over 90% of elastic-net folds. Knee flexion fPC3 reflects the knee flexion excursion from late stance (39%) to initial swing (67%) of the gait cycle, with higher scores indicating a healthier, spring-like pattern and lower scores a stiffer, more restricted pattern. Knee adduction fPC1 reflects a frontal-plane varus excursion at midstance (19% of the gait cycle), consistent with established varus-thrust findings. An LMM with Kellgren–Lawrence (KL) grade as the outcome showed that greater self-reported knee pain (mild vs. minimal: β = 0.463; moderate-to-severe vs. minimal: β = 0.662), lower knee flexion/extension fPC3 (β = −0.212), and greater knee adduction/abduction fPC1 scores (β = 0.169) were independently associated with greater radiographic severity after adjusting for age, sex, BMI, and gait speed (marginal R² = 0.231, 95% CI: 0.173–0.329; all p < .001). Dominance analysis demonstrated that self-reported knee pain, knee flexion/extension fPC3, and knee adduction/abduction fPC1 accounted for 25.6%, 20.3%, and 12.3% of the model's explained variance, respectively.  

Discussion/Conclusion. Patient-reported knee pain and waveform-derived gait features provide complementary information about radiographic KOA severity. These findings support integrating symptom assessment with gait mechanics to improve individualized KOA evaluation and rehabilitation strategies. 

Keywords
knee osteoarthritis
Kellgren-Lawrence grade
gait analysis
functional principal component analysis
linear mixed-effects model
dominance analysis
patient-reported pain
From Bioinnovation to Adoption: A Practical Translational Pathway
Advanced Control of Miniaturized Robots in Real-World Environments for Biomedical Applications