Association of trunk sway and gait-cycle variability measured by triaxial accelerometry with heart rate-based walking efficiency in patients with mild hemiparesis.

Abstract

The walking efficiency of individuals who have experienced a stroke and have mild hemiparesis may be reduced even in the absence of visually apparent gait abnormalities. We investigated the association between several gait parameters assessed by a triaxial accelerometer and walking efficiency in this population. Patients with first-ever stroke and mild hemiparesis who could ambulate independently without assistive devices (n = 36) were included. Gait assessments were conducted when the patients were able to walk continuously for ≥14 m and for 3 min. We used the root mean square (RMS) of acceleration to quantify the patients' trunk sway during walking. Trunk asymmetry and gait-cycle variability were assessed with the Lissajous index and coefficient of variation (CV), respectively. Walking efficiency was evaluated with the Physiological Cost Index (PCI). Correlation and multiple regression analyses were performed, with age and lower-limb motor function assessed by the Stroke Impairment Assessment Set (SIAS) as covariates. The RMS results (composite: ρ = 0.65, mediolateral: ρ = 0.64, vertical: ρ = 0.65, anteroposterior: ρ = 0.51) and CV (ρ = 0.49) were significantly positively correlated with the PCI results (p < 0.01). The SIAS lower-limb score was significantly negatively correlated with the PCI (ρ = -0.47, p < 0.01). No significant associations were observed for the Lissajous index or age. The multiple regression analysis identified all RMS components and CV as independent predictors of PCI results. In patients with mild hemiparesis, trunk sway and gait-cycle variability were associated with decreased walking efficiency. The mediolateral and vertical RMS values and CV may serve as sensitive indicators of gait-related energy inefficiency.