The Hamilton Study: Estimating Exposure to Ambient Suspended Particles

In the industrial city of Hamilton, Ontario, Canada, we recently carried out an epidemiological study of the effect of environmental factors on respiratory health in 3500 elementary, schoolchildren. The level and size distribution of suspended particles in ambient air was measured from 24-h samples taken at 6-day intervals from a network of 29 hivol TSP samplers, and nine Andersen 2000 4-stage cascade impactors. Exposure was computed by generating a 3-dimensional response surface (in TSP and in easting and northing geographical coordinates) using a linear regression model of the form: TSP = (1 + E + N)², based on monthly geometrical mean data for all sites. From the response surface generated for a given month, TSP levels were predicted by the model for all schools by specifying their geographical coordinates. The yearly exposure for a given child was determined from the arithmetic mean of the predicted values for 12 monthly TSP levels. A similar procedure was employed for calculation of the exposure to the "fine" (≤3.3 μm) and "coarse" (>3.3 μm) size fraction, as well as the aerodynamic mass median diameter of particles from the network of cascade impactors. Results of the measurements showed that gradients for TSP up to approximately 10 μg/m³/km exist over the city covering distances from 5 to 10 km. The range of 1 yr mean exposure values calculated for each child was from 30.5 μg/m³ to 74.5 μg/m³. Comparable figures for particle size were up to 0.3 μm AMMD (aerodynamic mass median diameter)/km and annual mean particle size exposure from 2.69 to 3.53 μm AMMD.