Abstracts: ISEE 22nd Annual Conference, Seoul, Korea, 28 August-1 September 2010: Air Pollution - Exposure Characterization and Health Effects
1French Institute for Public Health Surveillance, Saint Maurice, France; and 2Ile de France Health Regional Observatory, Paris, France.
Abstracts published in Epidemiology have been reviewed by the societies at whose meetings the abstracts have been accepted for presentation. These abstracts have not undergone review by the Editorial Board of Epidemiology.
Many studies have established the relationship between ozone and short-term mortality. In the present study, we used several strategies to control for time trends and temperature, and we explored the interaction between ozone and season in 9 French cities during the 1998–2006 period.
For each city, the association between daily max-8 hour ozone and daily number of nonaccidental deaths was analyzed using 3 statistical methods: (a) a Poisson generalized additive model (GAM) controlling for long-term trends, seasonality, and day of the week, (b) a time-stratified case-crossover design, (c) a temperature-stratified case-crossover design. The models included ozone lag 0–1, minimum temperature at lag 0, and mean maximum temperature over lags 1–7. The modifying effect of season was studied with ozone–season interaction terms in the GAM models and by stratified analyses in case-crossover models. City specific results were then combined using random effect models.
Results of the 3 modeling strategies were very similar. In the whole study period, the daily number of deaths increased by 0.64% (95% CI: 0.46–0.82) for an increase of 10 μg/m3 in the max-8 hour ozone level in the GAM model. The increase was slightly higher in the time-stratified (0.69% [95% CI: 0.52–0.86]) and in the temperature-stratified case-crossover models (0.73% [95% CI: 0.42–1.04]).
In the 3 models, the effect of the ozone was higher during summer (June-July-August), with an increase by 1.17% (95% CI: 0.81–1.54) in the GAM model, 1.00% (95% CI: 0.73–1.27) in the time-stratified case-crossover design, and 1.22% (95% CI: 0.87–1.57) in the temperature-stratified case-crossover design.
This study provides evidence that the association between ozone and mortality is not sensitive to the way time trends and temperature are controlled. The effect of ozone varies with season and is higher in summer.