Muscle co-activation patterns during walking in those with severe knee osteoarthritis

Background Sensory and motor impairments have been found for those with knee osteoarthritis; however, how these impairments are manifested during functional movements such as walking is not well established. A few studies suggest an increase in co-activity among lower limb muscles. The objective of this study was to characterize the neuromuscular patterns of knee joint muscles during walking for those with severe knee osteoarthritis using pattern recognition techniques on the entire waveform. Methods Fifty-one subjects received a gait assessment within one-week prior to total knee replacement surgery. Subjects walked along a 6-m walkway at their preferred walking speed while surface electromyograms from seven muscles were recorded. The electromyographic data were entered into a pattern recognition procedure that captured both the amplitude and shape characteristics of electromyographic waveforms. ANOVA models tested whether differences existed both among and within muscle groups for these waveform characteristics. Findings Four principal patterns explained 97% of the variance in the waveform data, with principal pattern one explaining 86% of the total variance. There were statistically significant differences (P < 0.05) among muscle sites for all principal pattern scores. The analyses supported the hypothesis that similarities existed in patterns among muscles from different groups indicating (i) a general co-activity pattern and (ii) differential recruitment between muscles within a muscle group. Interpretation In addition to the roles during impact loading and propulsion, the muscle responses were consistent with attempts to (i) decrease medial knee joint loading, (ii) decrease peak knee joint loading during push off and (iii) increase stiffness during stance phase to improve joint stability. The technique employed provides a novel approach to quantify synergistic co-activity.