Timing of physical activity across adulthood on later-life cognition: 30 years follow-up in the 1946 British birth cohort

Abstract

Background: To assess how timing, frequency and maintenance of being physically active, spanning over 30 years in adulthood, is associated with later-life cognitive function.

Methods: Participants (n=1417, 53% female) were from the prospective longitudinal cohort study, 1946 British birth cohort. Participation in leisure time physical activity was reported five times between ages 36 and 69, categorised into: not active (no participation in physical activity/month); moderately active (participated 1-4 times/month); most active (participated 5 or more times/month). Cognition at age 69 was assessed by tests of cognitive state (Addenbrooke's Cognitive Examination-III), verbal memory (word learning test) and processing speed (visual search speed).

Results: Being physically active, at all assessments in adulthood, was associated with higher cognition at age 69. For cognitive state and verbal memory, the effect sizes were similar across all adult ages, and between those who were moderately and most physically active. The strongest association was between sustained cumulative physical activity and later-life cognitive state, in a dose-response manner. Adjusting for childhood cognition, childhood socioeconomic position and education largely attenuated these associations but results mainly remained significant at the 5% level.

Conclusions: Being physically active at any time in adulthood, and to any extent, is linked with higher later-life cognitive state, but lifelong maintenance of physical activity was most optimal. These relationships were partly explained by childhood cognition and education, but independent of cardiovascular and mental health and APOE-E4, suggestive of the importance of education on the lifelong impacts of physical activity.

Keywords: alzheimer's disease; cognition; dementia; neuroepidemiology.

Figure 1

Flow chart of the analytical sample. ACE-III, Addenbrooke’s Cognitive Examination-III; NSHD, National Survey of Health and Development; SEP, socioeconomic position; VSS, Visual Search Speed; WLT, Word Learning test.

Figure 2

Associations between physical activity at assessed time periods (ages 36, 43, 53, 60 and 69), and cumulative physical activity across these time periods, with standardised cognitive function tests at age 69. Standardised coefficients (difference in SD of the cognitive test) and 95% CIs are presented from multiple linear regression sex-adjusted models, which compare cognitive test scores between: (A) those who were not active in a time period with those who were moderately active (participated in activity 1–4 times per month); and those who were most active (participated ≥5 times per month); (B) those who were never physically active in adulthood with those who were active (participated in activity ≥1 time per month) at varying durations across adulthood. ACE-III, Addenbrooke’s Cognitive Examination-III total test scores.

Figure 3

Associations between accumulative adulthood physical activity with Adenbrooke’s Cognitive Examination-III total score at age 69. Coefficients and 95% CIs are presented from multivariable linear regression models comparing those who were never physically active in adulthood with those who were active (participated in activity ≥1 times per month) at varying frequencies across adulthood. Models were adjusted for sex (blue); childhood cognition, childhood socioeconomic position (SEP) and education (red); cardiovascular (CV) health at age 69 (green); emotional mental health (MH) at age 69 (orange) and APOE-E4 status (grey).

Figure 4

Associations between life course patterns of adulthood physical activity, compared with those never active, on the Adenbrooke’s Cognitive Examination (ACE-III) total score at age 69. coefficients (difference in the cognitive test) and 95% CIs are presented from multiple linear regression sex-adjusted models, which compare cognitive test scores between those who were never physically active in adulthood with those who were active (participated in activity ≥1 times per month) at varying periods across adulthood.