In recent years, several studies in Europe have associated within-city contrasts in air pollution with various health end points including mortality in cohort studies of adults, and respiratory morbidity in cross-sectional and cohort studies of children. Many of these studies have used NO2 contrasts as the primary exposure variable, which raises the issue of whether such associations are uniquely found for NO2 per se, or whether NO2 acts as a surrogate for a complex mixture of combustion pollutants primarily derived from vehicular traffic. Exposure assessment in these studies has been based on dispersion modelling, on data from routine monitoring networks, on stochastic models developed from dedicated spatially resolved monitoring, or some combination of these. The results of a number of recent European studies are discussed.