Associations between carbon monoxide exposure, physical activity, and outdoor play preferences among UK children: evidence from the millennium cohort study

BACKGROUND: Physical inactivity and air pollution are significant global health concerns, yet limited evidence examines how residential air pollution is associated with physical activity behaviours among children. Existing research has mainly focused on adults, older adults, or self-reported activity outcomes. This study examined the association between residential carbon monoxide (CO) exposure, objectively measured physical activity, and children's reported preference for outdoor sports and games using data from the Millennium Cohort Study. METHODS: This study used a cross-sectional secondary analysis of Sweep 4 Millennium Cohort Study data. Physical activity was measured using accelerometer-derived total steps over one week. Residential CO exposure was derived from linked MEDIx air pollution deciles and categorised into relatively lower and higher exposure groups. Linear regression was used to examine the association between CO exposure and total steps. Logistic regression was used to examine the association between CO exposure and preference for outdoor sports and games. Models were adjusted for sex, ethnicity, and socio-economic position, with survey and non-response weights applied. RESULTS: The analytic sample included 6,281 children. In the adjusted linear regression model, children living in areas with higher CO exposure recorded, on average, 356.26 more steps than those living in areas with lower CO exposure. In the adjusted logistic regression model, higher CO exposure was associated with slightly higher odds of reporting a preference for outdoor sports and games. However, the adjusted step-count model explained only a small proportion of variation in total steps. CONCLUSION: Higher residential CO exposure was modestly associated with higher total steps and greater reported preference for outdoor sports and games. These findings should be interpreted cautiously because the study was cross-sectional, explanatory power was low, and residual confounding by neighbourhood, household, school travel, and built-environment factors is likely. Future studies should examine multiple pollutants alongside more detailed contextual and environmental measures.