Occupational exposures estimated by means of job exposure matrices in relation to lung function in the PAARC survey.

OBJECTIVES The aim of this analysis of the French Cooperative PAARC (Pollution Atmosphérique et Affections Respiratoires Chroniques) survey, was to test whether occupational exposures to dusts, gases, or chemical fumes or to specific hazards, estimated by job exposure matrices, were related to a decrease in forced expiratory volume in one second (FEV1). METHODS The most recent occupation was recorded in adults, aged 25-59, from non-manual worker households. Analysis was restricted to 10,046 subjects whose occupation was encountered at least 10 times in the study and who performed good FEV1 tracings. From occupational title, exposures to dusts, gases, and chemical fumes, and to specific hazards were classified in three categories (no, low, and high) with a British, a French, and an Italian job exposure matrix. Specific hazards were analysed for the British and French job exposure matrices for the same 42 specific dusts, gases, and chemical fumes. To limit spurious associations, a selection of seven hazard groups and 12 specific hazards was set before the start of the analysis. Based on the consistency of the relations according to sex and the British and French job exposure matrices, associations of age, height, city, and smoking adjusted FEV1 score with occupational exposures were classified as very likely, possible, or unlikely. RESULTS For the three job exposure matrices and both sexes clear exposure-response relations between the level of exposure to dusts, gases, and chemical fumes, and a decrease in FEV1 were found. Associations with FEV1 were classified as very likely for known hazards such as organic dusts and textile dusts, and not previously recognised hazards such as polycyclic aromatic hydrocarbons (PAHs) and detergents, and as possible for solvents, waxes and polishes, and diesel fumes. Associations found for PAHs and solvents were confirmed by the Italian job exposure matrix. Associations remained significant in women, but not in men, after adjustment for educational level. CONCLUSIONS Hypotheses have been generated for exposure to detergents, PAHs, and solvents, but they need to be interpreted with caution before replication. Significant associations found for known risk factors with a decrease in FEV1 are arguments for the validity of the matrices. Despite the expected limitations of job exposure matrices, these results encourage further work to improve exposure assessment by job exposure matrices.