The long-term impact of restricting cycling and walking during high air pollution days on all-cause mortality: Health impact Assessment study

Regular active commuting, such as cycling and walking to and from the workplace, is associated with lower all-cause mortality through increased physical activity (PA). However, active commuting may increase intake of fine particles (PM2.5), causing negative health effects. The purpose of this study is to estimate the combined risk of PA and air pollution for all-cause mortality among active commuters who, on days with high PM2.5 levels, switch to commuting by public transportation or work from home. Towards this purpose, we developed a Health Impact Assessment model for six cities (Helsinki, London, Sao Paulo, Warsaw, Beijing, New Delhi) using daily, city-specific PM2.5 concentrations. For each city we estimated combined Relative Risk (RR) due to all-cause mortality for the PA benefits and PM2.5 risks with different thresholds concentrations. Everyday cycling to work resulted in annual all-cause mortality risk reductions ranging from 28 averted deaths per 1000 cyclists (95% confidence interval (CI): 20-38) in Sao Paolo to 12 averted deaths per 1000 cyclists (95% CI: 5-19) in Beijing. Similarly, for everyday walking, the reductions in annual all-cause mortality ranged from 23 averted deaths per 1000 pedestrians (95 CI: 16-31) in Sao Paolo to 10 averted deaths per 1000 pedestrians (95%CI: 5-16) in Beijing. Restricting active commuting during days with PM2.5 levels above specific air quality thresholds would not decrease all-cause mortality risk in any examined city. On the contrary, all-cause mortality risk would increase if walking and cycling are restricted in days with PM2.5 concentrations below 150 mu g/m(3) in highly polluted cities (Beijing, New Delhi). In all six cities, everyday active commuting reduced all-cause mortality when benefits of PA and risk or air pollution were combined. Switching to working from home or using public transport on days with high air pollution is not expected to lead to improved all-cause mortality risks.