Abstract
BACKGROUND: The effects of small, realistic changes in physical activity and sedentary behaviour on population-level mortality are unclear. We aimed to estimate the proportion of deaths preventable by 5-min and 10-min incremental increases in moderate-to-vigorous intensity physical activity (MVPA) and 30-min and 60-min reductions in daily sedentary time.
METHODS: We did an individual participant data meta-analysis of prospective cohort studies. We included studies with device-measured physical activity and sedentary time. We estimated the proportion of deaths prevented (potential impact fractions; PIFs) by changes in (1) the approximately 20% least active participants (high-risk approach) and (2) all participants except the approximately 20% most active (population-based approach). We calculated PIFs from adjusted hazard ratios estimated for 5-min and 10-min increases in MVPA and 30-min and 60-min reductions in sedentary time from observed levels across the activity distribution.
FINDINGS: We included seven cohorts from Norway, Sweden, and the USA (n=40 327; 4895 deaths). Data from the UK Biobank (n=94 719; 3487 deaths) were analysed separately. A 5-min/day increase in MVPA in the least active participants might prevent 6·0% (95% CI 4·3-7·4) of all deaths. A similar increase in MVPA in all participants except the most active might prevent 10·0% (6·3-13·4) of all deaths. Reducing sedentary time by 30 min/day might prevent 3·0% (2·0-4·1) of all deaths in the high-risk approach and 7·3% (4·8-9·6) in the population-based approach. Results from the UK Biobank were of a smaller magnitude but still substantial-eg, reducing sedentary time by 30 min/day in all except the most active participants was associated with preventing 4·5% (2·8-6·1) of total deaths.
INTERPRETATION: Small and realistic increases in MVPA of 5 min/day might prevent up to 6% of all deaths in a high-risk approach and 10% of all deaths in population-based approach. Reducing sedentary time by 30 min/day might prevent a smaller, but still meaningful, proportion of deaths in the two risk scenarios.
FUNDING: None.