Although an overall public health target of 10,000 steps per day has been advocated, the dose–response relationship for each health benefit of physical activity may differ.
A representative community sample of 2458 Australian residents aged 55–85 wore a pedometer for a week in 2005–2007 and completed a health assessment. Age-standardized steps per day were compared with multiple markers of health using locally weighted regression to produce smoothed dose–response curves and then to select the steps per day matching 60% or 80% of the range in each health marker.
There is a linear relationship between activity level and markers of inflammation throughout the range of steps per day; this is also true for BMI in women and high density lipoprotein in men. For other markers, including waist:hip ratio, fasting glucose, depression, and SF-36 scores, the benefit of physical activity is mostly in the lower half of the distribution.
Older adults have no plateau in the curve for some health outcomes, even beyond 12,000 steps per day. For other markers, however, there is a threshold effect, indicating that most of the benefit is achieved by 8000 steps per day, supporting this as a suitable public health target for older adults.
Ewald, Attia, and McElduff are with the Centre for Clinical Epidemiology & Biostatistics, University of Newcastle, Callaghan, New South Wales, Australia.