Free-Living Physical Activity Energy Expenditure Based on Accelerometry Versus Heart Rate in Community-Dwelling Older Adults

Defining the intensity and volume of physical activity (PA) for older people using device-based monitoring still lacks universal recommendations as most studies have focused on younger populations. We compared PA estimates from wearable accelerometers and heart rate monitors to investigate their correspondence in overall PA volume and across PA intensity categories: light, moderate, and vigorous. Participants were community-dwelling older adults ( N = 253), aged (mean ± SD ) 78.4 ± 3.4 years, who wore a heart rate monitor and a thigh accelerometer for at least 3 days in their daily lives. Two novel calibration methods were utilized to estimate physical activity energy expenditure (PAEE) from heart rate and accelerometry separately. Heart rate-based estimates indicated higher PAEE compared with accelerometry, especially at low intensities (13.2 ± 9.5 vs. 7.1 ± 2.3 kJ·kg −1 ·day −1 , p < .001) and vigorous intensities (3.3 ± 6.9 vs. 0.4 ± 1.2 kJ·kg −1 ·day −1 , p < .001). However, estimates for moderate PA were consistent between the two methods (15.7 ± 14.7 vs. 14.3 ± 6.2 kJ·kg −1 ·day −1 , p = .096). Male sex and beta blocker use were associated with greater differences in total PAEE between the methods (β [95% confidence interval] 6.1 [0.16, 12.0] and 7.4 [1.50, 13.4] kJ·kg −1 ·day −1 , respectively). Bland–Altman analyses indicated that heart rate-based PAEE was disproportionately higher than accelerometry-based PAEE in individuals at the higher end of the PAEE distribution, whereas the opposite was true for individuals at the lower end. These findings suggest that accelerometry and heart rate provide differing estimates of PAEE for the most and the least active older adults.