Abstract 30: Associations of Accelerometer-Measured Machine-Learning Classified Sitting With All-Cause and Cardiovascular Disease Mortality Among Older Women: The Opach Study

• Published in: Circulation (New York, N.Y.) Vol. 147; no. Suppl_1; p. A30
• Main Authors: Nguyen, Steve, Bellettiere, John, Di, Chongzhi, Anuskiewicz, Blake, Natarajan, Loki, Lamonte, Michael J, LaCroix, Andrea Z
• Format: Journal Article
• Language: English
• Published: Lippincott Williams & Wilkins 28.02.2023
• ISSN: 0009-7322; 1524-4539
• DOI: 10.1161/circ.147.suppl_1.30

IntroductionSedentary behavior (SB) is a recognized mortality and CVD risk factor. Most studies with accelerometry classified SB using cut-points, which do not capture postural transitions as accurately as thigh-worn devices. The recently published convolutional neural network hip accelerometer posture (CHAP) algorithm more accurately classifies sitting than cut-point methods. HypothesisHigher amounts of CHAP-classified sitting time (ST) and mean sitting bout duration (MSBD) are associated with higher all-cause (ACM) and CVD mortality risk. MethodsOlder women (n=6,056; mean age=79±7; 34% Black, 17% Hispanic) in the Objective Physical Activity and Cardiovascular Health (OPACH) study without prior MI or stroke wore accelerometers for 7 days in May 2012-April 2014 and were followed through February 19, 2022 for mortality. The CHAP algorithm has been shown to have higher sensitivity (97.1% vs 88.2%) and specificity (88.6% vs 59.7%) for classifying sitting compared to the 100 counts/minute cut-point. Cox models estimated hazard ratios (HR) and 95% confidence intervals (CI) for ACM and CVD mortality adjusting for age, race/ethnicity, education, alcohol, smoking, multimorbidity, self-rated health, physical functioning, HDL, triglycerides, SBP, and log hs-CRP. ResultsThere were 1,808 deaths and 651 CVD deaths over a median follow-up of 8.4 years. The HR (95% CI; P-trend) comparing women in the highest ST quartile (>11.6 hr/day) to those in the lowest (<9.2 hr/day) was 1.43 (1.23-1.66; <0.001) for ACM and 1.64 (1.27-2.14; <0.001) for CVD mortality. The HR (95% CI; P-trend) comparing women in the highest MSBD quartile (>15 minutes) to those in the lowest (<9.3 minutes) was 1.33 (1.15-1.55; <0.001) for ACM and 1.50 (1.16-1.94; <0.001) for CVD mortality. The HR (95% CI) for women with the highest ST and MSBD was 1.63 (1.38-1.93) for ACM and 1.95 (1.47-2.60) for CVD mortality. ConclusionsST and MSBD were positively associated with ACM and CVD mortality risk, supporting interventions aimed at reducing both ST and MSBD.