Hilde Notø, Karl-Christian Nordby, �’ivind Skare, Helge Kjuus and Wijnand Eduard

Background and Objectives Respiratory effects have been linked to aerosol exposure in cement production workers. This presentation aims at estimating annual exposure levels to thoracic aerosol during the study period. The thoracic fraction was chosen because dynamic lung function was the main outcome of the international study.

Methods We collected 7120 personal shift measurements of thoracic aerosol contributed by 2866 persons within 8 job types in 24 plants in 2007, 2009 and 2011. Measurements above 150 mg/m³ were excluded as light microscope analysis revealed larger particles than expected from the thoracic convention (N = 63). Measurements with absolute Z-scores above 3.29 in models including job type, plant and year were also excluded (N = 44), as well as samples with technical errors (N = 71). Arithmetic mean (AM) exposures were estimated using mixed regression modelling of the ln-transformed exposure. The final model included plant, job type, plant*job type, year, plant*year and season as fixed effects, individuals as random effect, and plant-specific residuals.

Results Of the workers 86% had been measured more than once, on average 2.4 times. All fixed factors contributed significantly to the models. Plant specific residuals also improved the model significantly. A job exposure matrix was constructed for plant-specific job types for each year averaged across seasons. AM exposure levels were estimated by exponentiation of the sum of the regression coefficients of the fixed effects and the half of the plant-specific residual variances. The estimated exposure varied between job types and plants from 0.13 to 14 mg/m³.