Upper and lower trunk accelerometer-based metrics during gait as predictors of physical function in older adults.

Abstract

Physical function and activities of daily living (ADLs) are key determinants of independence in older adults, with walking being an essential component of ADLs. Gait assessment using multiple sensors may enable early detection of functional decline and cause-specific intervention. We hypothesized that specific accelerometer-based gait metrics from both the upper and lower trunk sensors are associated with declining mobility in older adults. Gait analysis was conducted in community-dwelling older adults with declining mobility using accelerometers attached to the upper and lower trunk. Physical function was evaluated using the five-repetition sit-to-stand test, single-leg standing time, three-meter Timed Up and Go test, preferred gait speed, and the locomotive syndrome risk test. Associations between physical performance and accelerometer-based metrics were examined using correlation, regression, and principal component analyses. Gait speed correlated strongly with Root Mean Square (RMS)-vertical (VT) and with several performance metrics. Upper RMS-VT, upper improved Harmonic Ratio in the anteroposterior direction (iHR-AP), upper iHR-VT, lower iHR-AP, and the upper/lower trunk RMS ratio in the mediolateral direction ([U/L] RMS ratio-ML) associated with performance metrics. After adjustment for age, sex, and gait speed, upper and lower iHR-AP and upper iHR-VT remained associated (p < 0.05). Principal component analysis indicated that principal component 1, which was correlated with gait speed, included RMS-VT, whereas iHRs belonged to a different component. Accelerometer-based metrics obtained from both the upper and lower trunk reflect physical function and may serve as indicators of early gait impairment, as well as for future interventional studies.