The assessment of the health impacts of air pollution in the Lyon metropolitan area is part of the development of the City of Lyon’s Air Protection Plan, launched in 2003. This plan follows the completion of the 2001–2005 Regional Air Quality Plan. The latter plan sets out guidelines aimed at preventing, reducing, or mitigating the effects of air pollution, drawing in particular on an assessment of the effects of air quality on public health. In this study, the short-term health impact of air pollution is calculated in terms of premature mortality and morbidity (hospital admissions) attributable to air pollution. The long-term health impact is estimated by the number of deaths attributable to air pollution. The study area selected corresponds to an urban zone where the population’s exposure to air pollution can be considered homogeneous based on defined criteria. It includes Lyon and 18 surrounding municipalities, representing a total population of 963,248 inhabitants (1999 census). The study period covers the years 2001 and 2002. This study is based on the methodological principles of the Urban Air Pollution Health Impact Assessment (EIS), whose methodology comprises four steps: hazard identification, exposure estimation, selection of exposure-risk relationships, and risk characterization. The selected air pollution indicators are based on four pollutants routinely measured in the area: NO2, SO2, O3, and PM10, for which the exposure-risk relationships used are derived from epidemiological studies conducted in the general population, with a focus on multicenter and European studies. A daily indicator of pollutant levels was calculated for each of the selected indicators. The total number of projected deaths attributable to air pollution for the years 2001 and 2002 amounts to 424 deaths, including 159 from cardiovascular disease and 32 from respiratory disease, which represents 22 deaths per year per 100,000 inhabitants, including 8 from cardiovascular causes and 2 from respiratory causes. Calculations of the health benefits associated with various air pollution reduction scenarios show that the most effective are those corresponding to 25% reductions in daily levels of the pollutant in question over the entire study period. Regarding long-term health benefits, in 2001 and 2002 the European standard set for 2005 (40 μg/m³ as an annual average) was already being met for PM10 concentrations. Compliance with the European standard set for 2010 would have prevented 97 deaths out of the total number of deaths recorded in a year. Once again, the most effective scenario is the one corresponding to a 25% reduction in the annual average exposure to the pollutant in question (PM10), which would prevent 154 deaths per year. Given the uncertainties and limitations of the methodology used, the results should be interpreted as orders of magnitude of the impact of air pollution on the health of the population in the study area. However, this study shows that even though the relative risks associated with air pollution are low, the large proportion of people exposed results in a significant collective impact on mortality. It also shows that health effects are already apparent at pollution levels well below those for which measures are currently taken. Consequently, reducing emissions at the source on a daily basis is preferable to controlling annual pollution peaks. (R.A.)