Differences in mean levels of physical capability (95% confidence intervals) per one standard deviation difference in sedentary time, moderate-to-vigorous physical activity and physical activity energy expenditure at age 60–64 years using multivariable linear regression models.

Model 1: adjusted for sex.
Model 2: adjusted for sex, height and weight.
Model 3: adjusted for sex, height, weight, education level, occupational class, smoking status and long-term limiting illness or disability.
n = 1,646 for grip strength; n = 1,710 for chair rise speed; n = 1,713 for standing balance time and n = 1,609 for TUG speed.
Associations highlighted in bold are statistically significant at p<0.05
* Each one unit (standard deviation) change equates to: 2.1 hours/day difference in time spent sedentary; a 60 min/day difference in moderate-to-vigorous physical activity and a 14.7 kJ/kg/day difference in physical activity energy expenditure.
Effect estimates are from analyses using the multiple imputation by chained equations method run across 10 imputed datasets and using Rubin’s combination rules to combine datasets.
Definitions: Sedentary time was defined as a MET value of <1.5 in accordance with current convention [30] and MVPA as ≥3.0 METs using an individualised estimate of RMR to define one MET [31].
Differences in mean levels of physical capability (95% confidence intervals) per one standard deviation difference in sedentary time, moderate-to-vigorous physical activity and physical activity energy expenditure at age 60–64 years using multivariable linear regression models.