The aim of this study is to evaluate ambulatory respiratory drug sales data as health indicators for the short-term effects of ambient air pollution in the city of Le Havre. Daily respiratory drug sales data were crossed with daily ambient air concentrations of sulfur dioxide (SO2), nitrogen dioxide (NO2), and black smoke (BS) using an autoregressive Poisson regression model adjusting for time trends, seasonal variations, influenza epidemics, and weather. Relative risks (RR) were expressed for an increase of two standard deviations above the mean of each pollutant. Respiratory drug sales were associated with most pollutants studied with lags varying from 1 to 9 days. For daily mean concentrations of BS, RR = 1.037 (95% confidence interval (CI) 1.009-1.066) for lag 1 and RR = 1.052 (95% CI 1.023-1.081) for lag 8. For daily mean concentrations of N02, RR = 1.033 (95% CI 1.001-1.066) for lag 1 and RR = 1.046 (95% CI 1.014-1.079) for lag 8. RR observed with a daily 1 h maximum of SO2 were RR= 1.027 (95% CI 1.004-1.051) for lag 3 and RR= 1.032 (95% CI 1.009-1.056) for lag 9. Our study concludes that ambulatory respiratory drug sales data could be useful for epidemiological surveillance of air pollutant health effects.