A Cross-Sectional and Longitudinal Study of the Relationship Between Walking Speed and Cognitive Function in Community-Dwelling Elderly People

Previous reports have shown links between cognitive function and physical performance in the elderly people, but it is unclear whether some specific cognitive domains are more strongly associated with measures of physical function such as walking speed. We investigated cross-sectional and longitudinal relationships between performance in five cognitive tests and walking speed among community-dwelling elderly people in the Dijon center (France) of the Three-City Study. At baseline, 3,769 participants aged 65-85 years had measurements of 6-m walking speed, global cognition (Mini-Mental State Examination), verbal fluency (Isaacs Set Test [IST]), psychomotor speed (Trail Making Test part A [TMT-A]), executive function (TMT part B), and memory (Benton Visual Retention Test). After a mean follow-up of 7 years, walking speed was again measured in 1,732 of these participants. In cross-sectional analyses, slower maximum walking speed (MWS) at baseline was significantly associated with poorer performance in each cognitive test. The association was stronger with TMT-A (beta [SE] = -.127 [0.014], p < .0001) and IST (beta [SE] = .120 [0.014], p < .0001) than with the other tests. Only TMT-A (beta [SE] = -.053 [0.021], p = .01) and IST (beta [SE] = .063 [0.022], p = .004) were associated with the degree of MWS decline over time. This study shows both cross-sectional and longitudinal associations between cognition and walking speed among community-dwelling elderly people. Poorer verbal fluency and slower psychomotor speed were more specifically associated with slower baseline MWS and with a stronger decline in MWS over time.