Objectives There is evidence of adverse associations between short-term exposure to traffic-related pollution and health, but little is known about the relative contribution of the various sources and particulate constituents.

Methods For each day for 2011–2012 in London, UK over 100 air pollutant metrics were assembled using monitors, modelling and chemical analyses. We selected a priori metrics indicative of traffic sources: general traffic, petrol exhaust, diesel exhaust and non-exhaust (mineral dust, brake and tyre wear). Using Poisson regression models, controlling for time-varying confounders, we derived effect estimates for cardiovascular and respiratory hospital admissions at prespecified lags and evaluated the sensitivity of estimates to multipollutant modelling and effect modification by season.

Results For single day exposure, we found consistent associations between adult (15–64 years) cardiovascular and paediatric (0–14 years) respiratory admissions with elemental and black carbon (EC/BC), ranging from 0.56% to 1.65% increase per IQR change, and to a lesser degree with carbon monoxide (CO) and aluminium (Al). The average of past 7 days EC/BC exposure was associated with elderly (65+ years) cardiovascular admissions. Indicated associations were higher during the warm period of the year. Although effect estimates were sensitive to the adjustment for other pollutants they remained consistent in direction, indicating independence of associations from different sources, especially between diesel and petrol engines, as well as mineral dust.

Conclusions Our results suggest that exhaust related pollutants are associated with increased numbers of adult cardiovascular and paediatric respiratory hospitalisations. More extensive monitoring in urban centres is required to further elucidate the associations.