Walking speed and propulsion asymmetry as distinct diagnostic biomarkers of post-stroke gait impairment: Towards precision rehabilitation after stroke
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Stroke-induced hemiparesis often results in slow walking and an asymmetrical gait; however, these impairments vary widely across patients. Indeed, some individuals walk fast despite severe gait asymmetry, whereas others walk slowly despite mild asymmetry. We posit that speed and symmetry reflect distinct dimensions of walking and may serve as complementary biomarkers of stroke recovery. In this cross-sectional study, 57 individuals with chronic hemiparesis were classified as “slow” or “fast” and as having “severe” or “mild” asymmetry, using pre-defined thresholds (i.e., speed = 0.8 m/s; symmetry = 31.5%). Kruskal-Wallis and regression analyses were used to evaluate differences in gait biomechanics (i.e., limb and joint propulsive power and the metabolic cost of walking) and ambulatory function (i.e., speed, distance, and functional balance) associated with these impairment classifications. When examined independently, slower speeds and greater asymmetry were similarly associated with a higher metabolic cost of walking (Δ range: 77% to 83% higher, p < 0.05) and reduced balance and walking function (Δ range: 26% to 82% lower, p < 0.05). However, the two classification approaches revealed different underlying biomechanical mechanisms: whereas slower walking corresponded to reduced paretic propulsive power (Δ: 71% lower, p < 0.05), greater asymmetry indicated a distal-to-proximal shift in power generation (Δ: 200% shift from the ankle to the hip, p < 0.05). Furthermore, multiple regression analyses revealed that each classifier was an independent predictor of metabolic cost (R²=0.70, p < 0.001) and ambulatory function (R² range: 0.34 to 0.55, p < 0.001). These findings demonstrate the complementary diagnostic value of measuring both walking speed and propulsion asymmetry, supporting the development of multi-dimensional phenotypes to enable precision neurorehabilitation.