BACKGROUND: A study was undertaken to assess the combined association between urban air pollution and emergency admissions for asthma during the years 1986-92 in Barcelona, Helsinki, Paris and London. METHODS: Daily counts were made of asthma admissions and visits to the emergency room in adults (age range 15-64 years) and children (< 15 years). Covariates were short term fluctuations in temperature and humidity, viral epidemics, day of the week effects, and seasonal and secular trends. Estimates from all the cities were obtained for the entire period and separately by warm or cold seasons using Poisson time-series regression models. Combined associations were estimated using meta-analysis techniques. RESULTS: Daily admissions for asthma in adults increased significantly with increasing ambient levels of nitrogen dioxide (NO2) (relative risk (RR) per 50 micrograms/m3 increase 1.029, 95% CI 1.003 to 1.055) and non-significantly with particles measured as black smoke (RR 1.021, 95% CI 0.985 to 1.059). The association between asthma admissions and ozone (O3) was heterogeneous among cities. In children, daily admissions increased significantly with sulphur dioxide (SO2) (RR 1.075, 95% CI 1.026 to 1.126) and non-significantly with black smoke (RR 1.030, 95% CI 0.979 to 1.084) and NO2, though the latter only in cold seasons (RR 1.080, 95% CI 1.025 to 1.140). No association was observed for O3. The associations between asthma admissions and NO2 in adults and SO2 in children were independent of black smoke. CONCLUSIONS: The evidence of an association between air pollution at current urban levels and emergency room visits for asthma has been extended to Europe. In addition to particles, NO2 and SO2--by themselves or as a constituent of a pollution mixture--may be important in asthma exacerbations in European cities.