Study design and subjects
In this study, we examined a cohort of Swedish pulp and paper mill workers from four sulphite mills and one sulphate (Kraft) mill established to study different effects of irritants. The main occupational exposures in pulping are wood dust, terpenes, and bleachery chemicals such as chlorine and chlorine dioxide. Sulphite pulping involves exposure to sulphur dioxide, especially in the digester, and sulphate pulping also involves exposure to hydrogen sulphide and other reduced sulphur compounds [6].
The cohort was established by identifying workers from the personnel files in each mill. The initial cohort comprised 14175 subjects employed in the mills for six months or more at any time between 1940 and 2000. A questionnaire was mailed to all members of this cohort who had been employed at any time between 1 January 1970 and 1 July 2000, who were aged under 80 years, and who were alive on 1 July 2000 (n = 7389). Most of the items on this questionnaire had been validated and used in previous studies [7, 8].
The response rate was 52.7% (n = 3894). Currently employed were more prone to respond (62.2%) compared to not currently employed (44.1%). Workers employed more than 10 years and workers aged 50-64 years, answered more frequently than the others. Of the responders 9% answered to the last reminder, a shorter version of the questionnaire survey containing the key questions about respiratory diseases and occurrence of gassing, but not on work change. In addition to the questionnaire data, information for all subjects was extracted from personnel files regarding time period of employment, job title, and mill department. The final population (n = 3226) for this study comprised subjects identified in the personnel files with complete data regarding both employment periods and respiratory work disability.
The local ethics committee at University of Gothenburg reviewed and approved the study protocol.
Definitions
Respiratory work disability was defined as an affirmative answer to the question "Have you ever had to change your job or task due to respiratory problems?"; the reported year in which this change occurred was also recorded [9].
Asthma was defined as self-reported physician-diagnosed asthma [7].
Childhood asthma was defined asthma with reported debut before 16 years of age.
Chronic bronchitis was defined as self-reported cough and phlegm for at least three months for two consecutive years or longer [10].
Chronic rhinitis was defined as rhinitis after the age of 15 that lasted more than a month [11].
Allergic status was defined as an affirmative answer to questions about allergy in childhood and/or ever having had hay fever.
Participants were divided according to smoking status, into current smokers, ex-smokers, and never-smokers. Ever-smokers were defined as workers who had ever smoked daily for at least one year. The years of starting and stopping smoking were used to create a time-dependent variable for smoking habits. Exposure to environmental tobacco smoke (ETS) in the workplace was defined as self-reported exposure to ETS several times a week over the years worked at the pulp mill.
Exposure to irritant gassings was defined either as "gassings" or as having worked in those departments with a high probability of exposure to irritant gases, that is, the digester and bleachery departments. Gassing was defined as a positive answer to at least one of the three sections of the question [8]: "Have you been exposed to i) sulphur dioxide, ii) chlorine/chlorine dioxide, or iii) any other irritant substances at work, resulting in coughing, breathlessness, wheezing, or chest pain?". When analyzing gassings, the unexposed workers were those who did not report any gassing episode at work. Workers in the digester and bleachery departments were compared with workers in the other departments, such as paper production departments.
Statistical methods
We only considered the periods when the subjects were employed in the pulp mills. The observation period was retrospective: 1970-2000. Person-years were calculated from start of employment until end of employment period (according to data from personnel files) or until the year of first occurrence of respiratory work disability. Incidence of respiratory work disability was calculated for the various disease categories; the incidence rate ratio (IRR) and its 95% test-based confidence intervals (CI) for respiratory work disability were computed by comparing workers reporting gassings (irritant peak exposures) with unexposed workers [12].
In the final analysis, the hazard ratios (HR) and 95% CI for respiratory work disability were modelled using Cox proportional hazards regression models. The outcome event was respiratory work disability, and time until event was determined by reported date of first occurrence of respiratory work disability. The exposure was analyzed either as gassings or as work in exposed departments adjusting for asthma, chronic bronchitis, chronic rhinitis, and passive smoking at work. Time-dependent smoking habits, time-dependent age more than 50 years, and allergic status were not included in the final model, and exclusion from the model had minimal impact on coefficients for the exposures of interest.
Sex was not included in the model, as only one woman reported respiratory work disability. All statistical analysis was performed using the statistical software packages SAS version 9.2 (SAS Institute, Cary, NC, USA).
