Exploration of physical activity, sedentary behavior and insulin level among short sleepers

Sufficient physical activity and sleep duration are essential for overall health. While one-third of the US population reports short sleep (<7 h), which is proven to link with negative health status. Current evidence on the relationship between physical activity, sedentary behavior, and serum insulin level in short sleep groups is limited.

The National Health and Nutrition Examination Survey (NHANES) was used to conduct this cross-sectional study of 8,494 adults (NHANES) 2007-2018. Serum insulin was quantitatively tested by human insulin immunoassay. Short sleep conditions were defined as ≤7 h per night. Physical activity conditions, including work activity, recreational activity, and sedentary behavior, were self-reported in NHANES by the Physical Activity Questionnaire using a 7-day recall method. The main analyses utilized weighted linear regression models due to the complex multistage sampling design of NHANES. Subgroup analysis and the influence of different lipid indices were explored in this study. In addition, a sensitivity analysis of participants without diabetes was conducted.

In fully adjusted models, increased levels of work and recreational activity significantly reduced insulin levels, with β values 95% CI = -0.002 (-0.003, 0.001) and β values 95% CI = -0.008 (-0.012, -0.003), respectively. However, sedentary behavior was positively associated with insulin levels, with a β value 95% CI =0.022 (0.009, 0.034). The sensitivity analysis further confirmed the benefits of recreational activity in controlling insulin levels. Through sex stratification analysis, it seemed that physical activity was more obviously impacted in the male than female groups.

Overall, our analysis demonstrates that in short sleepers, an increased level of work and recreational activity is beneficial to control the insulin level, and more sedentary time is harmful. However, this association might be discrepant in different sexes and different levels of lipid indices.