Relationship Between Knee Biomechanics and Pain in People With Knee Osteoarthritis: A Systematic Review and Meta‐Analysis

OBJECTIVE: Our primary aim was to determine the cross-sectional relationship between knee biomechanics during gait and pain in people with medial knee osteoarthritis. Our secondary aim was to evaluate differences in knee biomechanics between symptomatic and asymptomatic participants with medial knee osteoarthritis. METHODS: Four online databases were searched from inception to July 2021. Eligible studies included people with medial/nonspecific knee osteoarthritis and a reported relationship between knee biomechanics during gait and pain or biomechanics of symptomatic and asymptomatic participants. Two reviewers independently extracted data and evaluated risk of bias. Random-effects meta-analyses were performed when three or more studies reported the same biomechanical variable for pooling (knee adduction moment KAM, KAM impulse, varus thrust, and peak knee flexion moment KFM). RESULTS: Forty studies were included. Methodological quality ranged from 4 to 9/10. Forty-seven unique biomechanical variables were reported. For the KAM, there was no correlation with pain for peak values pooled (early stance and overall) (r = 0.00, 95% confidence interval 95% CI: -0.12, 0.11, k = 16), a small negative correlation for early stance peak alone (r = -0.09, 95% CI -0.18, -0.002, k = 12), and a medium positive correlation for the overall peak during stance (r = 0.30, 95% CI 0.17, 0.42, k = 4). Metaregression identified that body mass index moderated the peak KAM-pain relationship (P < 0.001). KAM impulse had a small positive correlation with pain (r = 0.23, 95% CI 0.04, 0.40, k = 5), and people with varus thrust had 3.84 greater odds of reporting pain compared with people without (95% CI 1.72, 8.53, k = 3). Meta-analyses for the peak KFM and pain correlation and secondary aim were nonsignificant. CONCLUSION: Some knee gait biomechanics were associated with pain in this cohort. Longitudinal studies are required to determine causality.