JMIR Publications

ABSTRACT

Background:

Age-related cognitive decline and gait slowing are common in older adults. While an association between cognitive function and gait speed has been demonstrated, its temporal sequence remains unclear, especially in Chinese older people.

Objective:

To examine the longitudinal reciprocal association of gait speed with cognition and the possible temporal sequence of changes in cognition and gait speed.

Methods:

Data were derived from the China Health and Retirement Longitudinal Study. Usual gait speed was measured over two 2.5-meter walks. Global cognition was assessed based on mental intactness and episodic memory. Cross-lagged panel models and linear mixed-effects models were performed to examine the hypothesized association between cognition and gait speed.

Results:

Totally 3,009 participants aged over 60 years (mean age 66.4 ± 5.4 years; 47.3% female) were included. After accounting for baseline assessments and potential confounders, faster gait speed at baseline was associated with better global cognition at follow-up (β = 0.056, P < 0.01), and simultaneously, initial cognition was also positively associated with subsequent gait speed (β = 0.117, P < 0.001). The longitudinal reciprocal associations were also significant and similar for specific domains of mental intactness and episodic memory (all P < 0.05). Notably, the effect from baseline global cognition to follow-up gait speed was significantly greater than the reverse pathway (2 = 6.50, P for difference = 0.011). Totally 3,009 participants aged over 60 years (mean age 66.4 ± 5.4 years; 47.3% female) were included. After accounting for baseline assessments and potential confounders, faster gait speed at baseline was associated with better global cognition at follow-up (β = 0.056, P < 0.01), and simultaneously, initial cognition was also positively associated with subsequent gait speed (β = 0.117, P < 0.001). The longitudinal reciprocal associations were also significant and similar for specific domains of mental intactness and episodic memory (all P < 0.05). Notably, the effect from baseline global cognition to follow-up gait speed was significantly greater than the reverse pathway (P for difference = 0.011). Trajectories from linear mixed-effects analyses further supported these dynamic relationships, showing that slower baseline gait speed predicted more declines in cognitive scores and lower baseline cognitive scores predicted steeper decreases in gait speed over time.

Conclusions:

There is a bidirectional association over time between usual gait speed and cognitive function among Chinese older adults, which might show ideas for identifying the causal direction of changes in gait speed and cognitive function. To prevent age-related gait slowing, early measures to maintain normal cognitive function are particularly recommended.