Gait variability is more strongly associated with walking speed than with age

BACKGROUND: Based on a complex systems perspective, quantifying the structure of gait variability using nonlinear analysis techniques have been posited to be sensitive to age-related declines in the biological subsystems. This study tested this claim by investigating relationships between age and various measures of gait across the adult lifespan. METHODS: Thirty-four adults ranging from 23 to 71 years old completed a 6-minute walk at their preferred speed. Seventeen measures of gait were quantified, including mean spatiotemporal measures of gait and linear and nonlinear measures of gait variability. RESULTS: Contrary to expectation, older age was positively correlated (p values <.05) only with coefficient of variation of stride length and sample entropy of stride time (R²s = .12 and.15, respectively). Age was instead most strongly associated (p < .05) with slower preferred gait speed (R² =.23). Compared to age, gait variability measures were more strongly associated with preferred gait speed. Slower walking was associated (p values <.05) with increased coefficient of variation of stride length (R² =.34), sample entropy of stride length (R² =.14), the correlation dimension of knee motion and vertical pelvis motion (R² =.34 and R² =.26, respectively), and decreased sample entropy of stride time (R² =.51) and dynamic stability of mediolateral pelvis motion (R² =.11). CONCLUSIONS: These findings emphasize that measures of gait variability may not be particularly sensitive to the age-related declines in the biological subsystems but rather, are more predictive of preferred walking speed.