Associations between appetite, physical activity and sedentary behaviour from hip- and wrist-worn accelerometers in older adults

Aim: The objective of the current study was to examine whether physical activity and sedentary behavior were associated with appetite among community-dwelling older adults. Methods: Cross-sectional analysis was performed on three cohort studies: the Longitudinal Aging Study Amsterdam (LASA); the Health, Aging and Body Composition Study (HABC Study) and the I'm Still Standing Study (ISS Study); (n = 1173, n = 162, n = 125; age range: 57-99, 85-95, 80-100 years; women: 51%, 56%, 61%, respectively). Physical activity and sedentary behavior were measured using hip-worn (LASA and HABC) and wrist-worn (ISS) accelerometers. Appetite was self-reported. Logistic regression models were fitted by accelerometer placement to explore the association between good appetite and various physical activity metrics (total activity, sedentary behavior, and time spent in different intensities of physical activity). Results: Among cohorts using hip-worn accelerometers, those having total activity within the highest tertile had more than double the odds of having good appetite compared with those within the lowest tertile (odds ratio [OR] 2.16 (1.15-4.06)). Each additional percent of daily sedentary behavior decreased the odds for having good appetite by 3% (OR 0.97 (0.95-0.996)), while each additional percent of daily light-intensity physical activity increased the odds for having good appetite by 4% (OR 1.02 (1.01-1.06)). No association was found between either physical activity or sedentary behavior and appetite for measurements with the wrist-worn accelerometers. Conclusions: Among community-dwelling older adults, the associations between appetite, accelerometer-assessed physical activity and sedentary behavior differ by accelerometer placement location. This study highlights the importance of careful interpretation of accelerometer data from different body locations and concurrent health outcomes.