Ground reaction force patterns in knees with and without radiographic osteoarthritis and pain: descriptive analyses of a large cohort (the Multicenter Osteoarthritis Study)

• Published in: Osteoarthritis and cartilage 2021-08, Vol.29 (8), p.1138-1146
• Main Authors: Costello, K.E., Felson, D.T., Neogi, T., Segal, N.A., Lewis, C.E., Gross, K.D., Nevitt, M.C., Lewis, C.L., Kumar, D.
• Format: Article
• Language: English
• Subjects: Aged; Arthralgia - physiopathology; Biomechanical Phenomena - physiology; Cohort Studies; Female; Gait Analysis; Ground reaction force; Humans; Knee; Knee Joint - diagnostic imaging; Knee Joint - physiopathology; Male; Middle Aged; Osteoarthritis, Knee - diagnostic imaging; Osteoarthritis, Knee - physiopathology; Pain; Principal Component Analysis; Radiographic osteoarthritis; Radiography
• ISSN: 1063-4584; 1522-9653
• DOI: 10.1016/j.joca.2021.03.009

To compare ground reaction force patterns (GRF) during walking among legs defined by presence or absence of knee pain and/or radiographic knee osteoarthritis (ROA). Principal component analysis extracted major modes of variation (PCs) in GRF data from the Multicenter Osteoarthritis Study during self-paced walking. Legs were categorized as pain + ROA (n = 168), ROA only (n = 303), pain only (n = 476), or control (n = 1877). Relationships between group and GRF PCs were examined using Generalized Estimating Equations, adjusted for age, sex, body mass index, race, and clinic site with and without additional adjustment for gait speed. With or without speed adjustment, pain + ROA had flatter vertical GRF waveforms than control (speed adjusted PC2 difference [95%CI]: −66 [-113,-20]), pain + ROA and ROA only had higher lateral GRF at impact and greater mid-stance medial GRF than control (speed adjusted PC3 difference: 9 [3,16] and 6 [2,10], respectively), and ROA only had higher early vs late medial GRF than control (speed adjusted PC2 difference: 7 [2,13]). Pain only had flatter vertical GRF waveforms and a smaller difference between anterior and posterior GRF than control only without speed adjustment. In this large sample, sustained mid-stance loading and higher impact loads were identified in legs with ROA or ROA and pain, even when adjusting for differences in gait speed and other confounders. While it remains to be seen whether these features precede or result from ROA and pain, the presence of these patterns in the speed-adjusted models could have implications on gait interventions aimed to change joint loading.