Relationships between physical activity, dementia mortality, and cognition in Finnish twins

Physical activity (PA) has been associated with better cognition or decreased dementia incidence in many earlier studies. Since both PA and cognition are, to a large extent, heritable traits, the association found in earlier studies may be confounded by genetic selection. The aim of this thesis study was to ascertain whether midlife PA is associated with late-life dementia mortality and cognition and if objectively measured late-life PA is associated with late-life cognition and whether the possible associations are contributable to genetic factors and childhood shared environment. The thesis study was implemented in the older Finnish Twin Cohort. Vigorous PA and the volume of PA have been reported in questionnaires in 1975 and 1981. Dementia mortality was followed from the time of the questionnaire in 1981 to the end of the year 2011 (n=21,524). All participants over 65 years of age were offered the possibility to participate in a telephone cognition interview. The cognition interview participants born in 1940–1944 were also offered the possibility to participate in accelerometer monitorings. The participation rates for questionnaires in 1975 and 1981, cognition interviews and accelerometer monitorings were high considering the length of the follow-up (89%, 84%, 78% and 54% among twins alive and with an address in Finland). In the accelerometer study, there was a trend toward a slightly selected population regarding better education and healthier lifestyle. Among the 21,524 twins followed for dementia mortality, there were 353 dementia deaths. Long-term vigorous PA was associated with decreased dementia mortality (hazard ratio (HR) in the fully adjusted model 0.60, 95% confidence intervals (CI) 0.39 – 0.93). The volume of PA and long-term quantitative physical inactivity were not significantly associated with dementia mortality. Among 3050 twins aged 65 years and older and who had participated in cognition interviews, 204 had cognitive impairment. Long-term vigorous PA was significantly associated with better cognition (beta estimate 0.91, 95% CI 0.47 – 1.35), belonging to the most inactive quintile in both questionnaire years 1975 and 1981 was inversely associated with cognition, and volume of PA was not significantly associated with cognition. Late-life objectively measured light PA was positively and sedentary behavior inversely associated with late-life cognition, but the effect sizes were small. In analyses comparing twins with their co-twins, the point estimates were quite similar to those found at the individual level but were statistically non-significant. The number of twin pairs discordant for PA and dementia mortality or cognition was very small. The results indicate that midlife vigorous activity is significantly associated with decreased dementia mortality and better late-life cognition, but the associations are likely to be explained in part by genetic factors and childhood shared environment. The association of objectively measured PA and cognition in late-life seems weak and, also, explained in part by genetic factors and childhood shared environment.