Distribution of device-measured 24-hour movement behaviours across the cognitive spectrum in older adults: the HUNT study

Abstract

Abstract Background Key daily physical activities (e.g., standing and walking), postures (sitting and lying), and sleep constitute the 24-h day and have distinct health effects and functional implications. Traditionally studied separately, the distribution of these movement behaviours and sleep across the cognitive spectrum remains insufficiently characterised. We aimed to describe and compare the distribution of 24-h movement behaviours and sleep across the cognitive spectrum. Methods We analysed cross-sectional, population-based data from 4,699 community-dwelling older adults (≥ 70 years) participating in the HUNT4 70 + study. Cognitive status was classified as mild cognitive impairment (MCI, n = 1,690) or dementia ( n = 360) according to DSM-5 criteria. Participants not classified with MCI or dementia were categorised as cognitively unimpaired. Accelerometers worn on the thigh and lower back (≤ 7 days) were used to assess standing, walking, sitting, lying awake, and sleep using validated machine-learning models. Inverse probability weighting and multiple imputation by chained equations were used to address selection bias. Group differences were examined using survey-weighted regression models adjusted for potential confounders including age, sex, education, marital status, self-rated health, obesity, depression, alcohol consumption, and smoking. Results In fully adjusted models, older adults with MCI spent less time sitting (ß -12.4 [95% CI -20.0, -4.4] min/ 24 h) and more time lying awake (ß 10.3 [95% CI 2.7, 17.9]) and sleeping (ß 3.9 [95% CI 0.8, 7.0] min/24 h) than their cognitively unimpaired peers (all p < 0.05), while time spent standing and walking was similar between the two groups. In fully adjusted models, older adults with dementia spent less time standing (ß -29.0 [95% CI -41.3, -16.5] and ß -30.3 [95% CI -42.5, -18.2] min/24 h) and walking (ß -16.1 [95% CI -19.9, -12.4] and ß -17.2 [95% CI -20.9, -13.4] min/24 h), and more time lying awake (ß 36.3 [95% CI 17.1, 55.4] and ß 46.5 [95% CI 28.7, 64.5] min/24 h), than those with MCI and cognitively unimpaired peers (all p ≤ 0.001), and more time sleeping (ß 7.5 [95% CI 0.9, 14.9] min/24 h) than their cognitively unimpaired peers ( p = 0.04). Conclusion Across the cognitive spectrum, older adults exhibit distinct distributions of 24-h movement behaviours and sleep. The clinical importance of the observed differences remains unknown, and future research on the independent and compositional effects of these behaviours is warranted.