RESUMO
BACKGROUND: Gait slowing and cognitive decline are both common in older people. Although cross-sectionally related, the longitudinal associations between specific cognitive functions and gait speed are less well understood. We aimed to determine whether decline in specific cognitive domains are associated with change in gait speed. METHODS: Participants aged 60-85, randomly selected from the electoral roll, were assessed twice over 3 years. Gait speed was obtained using the GAITRite walkway. Raw scores from a cognitive battery were subjected to principal component analyses deriving summary domains of executive function, processing speed, memory, and visuospatial ability. Multivariable linear regression was used to examine the associations between change in each cognitive domain and change in gait speed, adjusting for covariates and stratifying for the presence of baseline cognitive impairment. RESULTS: Mean age at baseline was 71.1 years (SD = 6.7) and 56% (159/284) were men. Mean follow-up was 2.55 (0.47) years. Decline in executive function, but not other cognitive domains (p > .05), was associated with decline in gait speed, cm/s (ß = -3.55, 95% CI = -5.49, -1.61; p < .001), both in the presence and absence of baseline cognitive impairment. Stronger associations were seen for those with baseline multiple domain cognitive impairment (ß = -6.38, 95% CI = -12.49, -0.27) and nonamnestic single-domain cognitive impairment (ß = -7.74, 95% CI = -14.76, -0.72). CONCLUSION: Decline in nonamnestic function (specifically executive function) was associated with decline in gait speed irrespective of the presence of baseline cognitive impairment. Strategies to improve or maintain executive function may prevent gait slowing.