Abstract P149: Sedentary Behavior In Mid-life And Risk Of Change In Global Cognitive Function And Incident Dementia: The Atherosclerosis Risk In Communities Neurocognitive Study (ARIC-NCS)

Abstract

Introduction: Sedentary behavior (SB) has emerged as a risk factor that may exert health effects independent from moderate-to-vigorous intensity physical activity (PA). However, the role of SB on cognitive health and its interaction with PA has not been widely examined prospectively. We tested the hypotheses that higher levels of, and persistence of mid-life television viewing, an established proxy measure of leisure-time SB, are associated with faster rates of cognitive decline and greater incidence of dementia, and that these associations are attenuated by meeting/not meeting the 2018 PA guidelines. Methods: ARIC participants (n=10,700, mean age: 59 years, 44% male, 19% Black) with self-reported assessments of television viewing at visits 1 (1987-1989) and 3 (1993-1995) were included. Participants were categorized as low [“never”/“seldom”], medium [“sometimes”], or high [“often”/“very often”] television viewing. Persistence in television viewing was quantified as the same reported television viewing at visits 1 and 3 (n=6,462). Cognitive tests of working memory, language, and executive function were administered at visits 4 (1996-1998) and 5 (2011-2013), and a factor analysis was used to derive a global cognition factor score. Dementia diagnoses were based on a battery of cognitive tests and expert adjudication. Additional cases were identified through hospital discharge codes and diagnostic codes from death certificates. Linear mixed models estimated the associations of SB categories with change in global cognition. Time to dementia was estimated using Cox proportional hazards regression models. Models were adjusted for age, education, race-ARIC field center, and APOE-ε4. Interactions with PA, categorized as meeting/not meeting PA guidelines, were tested. Results: Over a median follow-up of 17.4 years, 1,063 dementia cases were observed. In unadjusted models, high vs. low television viewing was associated with higher incident dementia (hazard ratio [95% CI], 1.42 [1.18, 1.71]); however, this association was no longer significant in multivariable adjusted models (1.09 [0.90, 1.32]). Significantly faster rates of decline in global cognition were observed for participants with high (-0.04 standard deviation (SD) difference [-0.08, -0.002]) or medium (-0.04 SD difference [-0.08, -0.004]) vs. low television viewing in multivariable adjusted models. These associations were stronger when measured as persistent mid-life television viewing (stable high vs. stable low: -0.06 SD difference [-0.11, -0.01]). The interaction of television viewing with PA guidelines on change in global cognition was statistically null. Conclusions: TV viewing, a proxy SB measure, is associated with cognitive decline, but not incident dementia. Longitudinal studies with device-based measures are needed to test the associations of accumulated daily SB on cognition and dementia outcomes.