Our main finding in this cohort of pulp and paper mill workers was the association between irritant peak exposures and respiratory work disability, defined as change of work due to affected breathing. In addition, asthma, but also chronic bronchitis and chronic rhinitis, played an important role in causing work change.
In this cohort, the incidence of respiratory work disability was 10.2/1000 person-years among subjects with asthma. This estimate is almost twice that found in asthmatics from the general population in the ECRHS (5.7/1000 person-years). The incidence rate among all workers in this study was 1.6/1000 person-years, which is quite similar to the estimate in the ECRHS general population sample (1.2/1000 person-years) . This could be explained by the fact that irritant exposure was fairly common; 40% of our study population reported gassings. The longitudinal analysis of the ECRHS study revealed an indication (not significant) of an exposure-response relationship between exposure to gases and fumes and respiratory work disability, both among asthmatics and in the general population. This indicates that higher exposure may increase the risk of respiratory work disability.
Most previous studies in this area were performed using population-based surveys. To our knowledge, this is one of only a few studies of respiratory work disability to examine an occupational cohort, pulp and paper mill workers in this case. Pulp production in the paper industry involves known respiratory irritants such as sulphur dioxide and chlorine dioxide, especially in the digester and bleachery departments, where we found a higher prevalence of work change. The prevalence of gassing in our study population (40%) was similar to that found in previous studies (45-60%); gassing occurs fairly frequently in the pulp industry, but is often not severe enough to require health care (e.g., physician intervention or hospitalization) and sick leave from work [10, 13, 14], so it does not necessarily lead to change of work. Some authors have addressed the issue of increased asthma risk in pulp and paper industry workers, especially in those who have experienced gassing [13, 14]. In this study asthma could be both a confounder (that is related to gassings and independently related to disability without being involved in the causal pathway) and an intermediate between the exposure and the outcome of disability. If asthma mainly acted as an intermediate the risk of disability associated with gassings should differ with and without asthma, but that has not occurred in our analyses. Workers with pre-employment asthma, as well as workers without respiratory diseases, were also at risk for disability. We found gassing to be a strong and independent predictor of changing work due to respiratory problems.
The present study did not suggest a strong effect of smoking habits on work change. This "healthy smoker effect" seems to be amplified in those with respiratory diseases, especially if the onset is in childhood or adolescence, in which case the subject is usually less likely to start smoking . Workplace ETS seems to be a weak risk factor for changing work.
It was not possible to study the influence of sex in this cohort, since only one woman reported changing work due to respiratory problems.
The definition of work change as an indicator of work disability was formulated on the basis of previous surveys, and was preferred to a self-reported work ability scale [3, 16]. In any case, the key questions about exposure may be double-biased by respiratory symptoms and the presence of work change due to respiratory symptoms. Prompted questions, such as those in the present questionnaire, seem less prone to recall bias than open-ended questions, and questions about irritant peak exposure have been used in other epidemiological studies [13, 14, 17].
The response rate was not high and could have influenced the result. The shorter questionnaire version was used for a random non-responders interview (n = 254) in the cohort. Items on work change were not included, but the risk of asthma related to gassings among non-responders was similar to responders despite differences in age and employment . Also the risk estimate and CI are so clearly significant that less risk among non-responders would not change the conclusion.