O2.1 CHRONIC OBSTRUCTIVE PULMONARY DISEASE AND EXPOSURE TO WOOD DUST

P. A. Demers, H. W. Davies, M. Friesen, C. Hertzman, A. Ostry, K. Teschke.School of Occupational & Environmental Hygiene, University of British Columbia, Vancouver, Canada

Many cross sectional studies have observed an increased prevalence of respiratory symptoms and impaired airflow associated with exposure to wood dust. However, cohort studies have generally observed a decreased risk of death due to chronic obstructive pulmonary disease (COPD) and no evidence of a dose–response relationship among workers exposed to wood dust. This could represent a true difference in the risk of disease or be due to differences in comparison populations (other workers versus the general population), disease endpoint (morbidity versus death), or quality of exposure assessment. To further evaluate the relationship between COPD and wood dust, we examined the relationship between hospitalisation for COPD and quantitative indices of exposure in a retrospective study of sawmill workers. The cohort consists of 11 424 workers from 14 lumber mills employed for one or more years after 1978. Hospital discharge records were available for the period 1985−1998. A quantitative exposure assessment was conducted for non-specific dust, wood dust, and tree species. Exposure metrics for cumulative exposure (duration and mg/m³ years) were developed. Relative risks (RR) were adjusted for sex, race, age group, and time period in internal analyses and time at risk for cases was counted until first hospitalisation for COPD. Overall, 133 members of the cohort were hospitalised with a primary diagnosis of COPD during the follow up period. No association was observed with duration of employment, although an excess risk was observed among workers with 10 to 19 years of employment relative to those with <5 years (standardised relative risk (SRR) 1.6; 95% confidence interval (CI) 0.6 to 4.7). A weak negative association was observed with non-specific dust. Higher incidence was observed among …