Introduction About 15% of adult-onset asthma is attributed to occupational exposures.

Objectives We examined whether prevention policies focusing on high-risk occupations adequately identify occupational asthma risks at a population level. We estimated, in a prospective population-based study, the distribution of asthma risk by occupation, and examined whether asthma risk in prior defined high-risk occupations were distinguished from those of prior defined non-high-risk occupations.

Methods ECRHS is a multicentre cohort study; 9409 participants (52% female) from 13 countries were followed for 10 to 20 years. Incident asthma was assessed by repeated questionnaires on asthma symptoms and medication. We examined the incidence of asthma for each occupation relative to all other occupations during follow-up using Generalized Estimating Equation Poisson regression. On the basis of prior evidence, we classified occupations during follow-up into two asthma risk groups (high and non-high risk). All high-risk occupations had documented exposure to at least one high-risk asthmagen (Occupational Asthma-specific Job Exposure Matrix [JEM]). We compared the distributions of the estimated log Relative Risks (logRR) for each group limiting the main analysis to 90 occupations with more than five incident cases.

Results The median logRR for the 14 prior defined high-risk occupations such as bakers, cleaners, and welders was higher (logRR=0.37) than the non-high risk occupations (logRR=0.06) (Mann-Whitney p-value=0.02). There was considerable overlap in the distributions of the logRRs by risk group (high risk, InteQuartileRange logRR 0.14–0.49; non-high risk IQR -0.21 to 0.36). Several non-high risk group occupations had significantly increased logRRs, including architects and receptionists.

Conclusion The significant overlap in risk curves by prior risk indicates that asthma-related exposures are prevalent and occur in multiple occupations. Current programs for occupational asthma prevention focus correctly on occupations at higher risk, but should also consider exposures in occupations not identified, a priori, as high risk.