Abstract
Introduction: Gait speed is a functional vital sign used to categorize community ambulation after stroke. The relationship between gait speed and quantity (i.e., step counts) of community ambulation post-stroke has been established. Yet, quantity alone does not depict a complete picture of community function. Gait quality metrics, such as symmetry, play a role in falls risk, and speak to functional motor impairments post-stroke. However, the relationship between gait speed and community-measured gait quality remains unknown. The purpose of this study was to investigate the relationship between gait speed and gait quality in a community setting after stroke.
Methods: A 10-meter walk test was used to evaluate each preferred gait speed (PGS) and fastest gait speed (FGS). The mean of three trials for each PGS/FGS was used. Participants then wore inertial measurement units (IMUs) over a one-week period in their home/community setting. IMUs recorded both step counts (quantity) and time in stance phase of each the paretic and non-paretic lower limb during the gait cycle. Symmetry (quality) was calculated by dividing the stance time of the paretic by the non-paretic limb, yielding a symmetry index (SI). To assess the relationship between gait speed and gait quality, we first standardized the magnitude of asymmetry by subtracting SI from one. We used a linear regression model to determine whether gait speed predicted gait asymmetry, controlling for NIHSS.
Results: Of 32 participants, the mean±SD age was 61±9 years, and time since stroke was 66±104 months. The sample included 15 (47%) females and 13 (41%) left CVAs. In our cohort, 22 (69%) had mild strokes (NIHSS 0-5) and 10 (31%) had moderate strokes (NIHSS 6-14). After controlling for NIHSS, each PGS (β = -0.237, SE = 0.085, t = -2.796, p = 0.009, CI [-0.411, -0.063]) and FGS (β = -0.141, SE = 0.050, t = -2.842, p = 0.008, CI [-0.243, -0.039]) significantly predicted gait quality. PGS model explained 17%, while FGS model explained 18% of the variance in gait quality. Figure 1 shows the linear relationship between each PGS (A) and FGS (B) to the amount of gait asymmetry. Regarding gait quantity, no significant relationship was found between PGS and FGS to gait quantity within our sample.
Conclusions: Both PGS and FGS are significant predictors of gait quality, but not quantity, in a community setting after stroke. Our results are limited by the small sample size. Future work with a larger sample is warranted.