Monitoring of short-term effects of urban air pollution on mortality. Results of a pilot study in 9 French cities

Background: This study aims at quantifying air pollution effects on mortali ty and at evaluating the feasibility of a standardized epidemiological surv eillance system of air pollution in 9 French cities. Methods: Data collection and analysis followed a standardized protocol. Dat a pollution depended on the development of local air quality surveillance n etworks (number of indicators, number of stations...). The Generalised Addi tive Models (GAM) were used to quantify the association between air polluti on and mortality. Results: In the 9 studied areas, associations between all causes, cardiovas cular and respiratory mortality, and air pollution indicators were observed . These associations were linear without threshold. Depending on the pollut ants, excess in mortality related to an interquartile increase in acid-part iculate pollution varied between 0.3 and 3.5% for total mortality, 0.5 and 6.3% for cardiovascular mortality, and between 0.1 and 12% for respiratory mortality. Photochemical air pollution varied between 0.4 and 7.3% for tota l mortality, 1.4 and 6.7% for cardiovascular mortality, and between 2.7 and 30.4% for respiratory mortality. Conclusion: In spite of a standardized common protocol, some disparities, i nherent to the local characteristics, were noted (length of time series, nu mbers of ambient urban stations selected and pollutants available...). Neve rtheless, this pilot study showed that multicentric epidemiological monitor ing of air pollution effects on health was feasible. Yet, this requires to validate the results obtained through a re-analysis of the mortality data o n a longer period of study. It also requires to study the feasibility and t he relevance of the use of other health indicators, such as hospital admiss ions.