Increased mortality in COPD among construction workers exposed to inorganic dust

While there is consensus that cigarette smoking is a specific cause of chronic obstructive pulmonary disease (COPD), the association between occupational exposures and COPD has not been as straightforward. However, in recent years, an increasing consensus about the risk associated with dusty work has emerged 1–4. One problem has been that measuring COPD in epidemiological studies is not simple, as no consensus exists regarding the epidemiological definition of COPD.

Annual decline of lung function has been used as one proxy variable for COPD. In longitudinal studies of occupationally exposed cohorts, the annual decline in forced expiratory volume in one second has been found to be related to gas, dust and fume exposure 5–10. The reported declines were ∼7–8 mL·yr⁻¹, after adjustment for age and smoking.

Studies of coal miners have shown an increased mortality due to bronchitis and emphysema 11–13, especially centrilobular emphysema 14–16. A relationship between dust exposure and degree of emphysema has been found in studies of coal and hard-rock miners 17, 18. Gold miners seem to have a greater risk of dying from COPD than coal miners, which may be due to the higher silica contents in gold mine dust 2.

Regarding other kinds of occupational dust exposure, the literature is scanty. Robinson et al. 19 analysed the mortality among 27,362 US carpenters' union members. The study was not controlled for smoking habits and they observed an increased mortality due to emphysema among construction carpenters. SjÖgren et al. 20, in a register-based study, reported that agricultural workers had increased mortality due to COPD. Hence, there is a need for powerful longitudinal studies evaluating the risk of COPD in groups occupationally exposed to other kinds of dust, not only from gold mines and coal mines. Here, the authors report the results of a prospective cohort study of 317,629 male construction workers exposed to different types of dust. The more specific aim of the study was to elucidate whether occupational exposure to inorganic dust, smoke or fume increased the risk for death from COPD, taking smoking habits into account.

Methods

In 1968, the Swedish Foundation for Occupational Safety and Health (Bygghälsan) was established to coordinate all activities concerning occupational health among Swedish construction workers. Construction workers were invited to health examinations at intervals of 2–5 yrs. Although the programme was voluntary, at least 80% of eligible workers participated at least once. Data from the health examinations, including occupational titles and smoking habits, were registered in a central database established in the early 1970s, as previously described 21–23. Through the personal identity number and a linkage with the National Cause of Death Register, it was possible to identify subjects who died (underlying cause) from COPD. The diagnoses from the International Classification of Diseases (ICD) were used, as follows: from ICD 8, 490–492; from ICD 9, 490–492 and 494–496; and from ICD 10, J43–J44. Each worker was followed from entry into cohort (earliest 1 January 1971) until 31 December 1999. The loss of subjects was low (0.15%) and such persons were excluded from the analysis. Those examined before the age of 15 yrs or after the age of 67 yrs were also excluded.

The smoking habit at the time of the first health examination was used. When information on smoking was lacking, information from a later visit was used. Hence, smoking was classified as never-smokers, exsmokers, smokers and unknown smoking habits. The occupational title at the time of the first health examination was used. If the title included jobs that were often used for replacement of sick persons, e.g. storage worker or cleaner, the individual was excluded from further analysis. Most workers remained in the same occupation. For example, among the workers who participated in repeated examinations, 74% reported the same occupation. The exposed group consisted of 200,735 males. The reference group consisted of all males in jobs that were assessed as nonexposed in the job-exposure matrix (see below), 116,894 males in total. The study was approved by the committee of ethics at Umeå University (Umeå, Sweden).

Results

Altogether 523 deaths due to COPD occurred among workers exposed to any airborne exposure and there were 200 deaths among the controls. Using the person-year method, stratifying for smoking and age, there was a slightly increased RR of death due to COPD (RR 1.12 (95% CI 1.03–1.22)) among workers with any airborne exposure as compared to controls. In the different major exposure groups, except exposure to wood dust, there was increased mortality due to COPD. Formal statistical significance was observed among those exposed to inorganic dust (RR 1.16 (95% CI 1.05–1.28)) and among those exposed to fumes (RR 1.22 (95% CI 1.04–1.42)) (table 3⇓).

When restricting the analyses to never-smokers, the mortality due to COPD among workers with any airborne exposure further increased (RR 2.11 (95% CI 1.43–3.00)). In addition, the RRs increased and remained significant among those exposed to inorganic dust, gases and irritants, and to fumes (table 4⇓).

The results of the Poisson regression analyses are shown in tables 5⇓ and 6⇓. When analysing the whole cohort, the increased mortality risk remained among those exposed to inorganic dust (hazard ratio 1.10 (95% CI 1.06–1.14)), but the risk declined among the groups exposed to gases and irritants, and to fumes. When restricting the analyses to never-smokers, there was an increased risk among those exposed to inorganic dust, gases and irritants, and to fumes.

The fraction of COPD attributable to work among the exposed was estimated as 10.7% overall and 52.6% among never-smokers.

Discussion

The main finding in this study was an increased mortality from COPD among construction workers exposed to inorganic dusts, as compared to other unexposed construction workers. In this cohort of male construction workers, the proportion of COPD that was attributable to work was considerable, with an attributable fraction of almost 11%. The meaning of this is that reduction of workplace exposures to dust, gases and fumes would prevent one out of ten deaths due to COPD, regardless of other risk factors, such as cigarette smoking.

The main strengths of the present study are that it is a prospective, longitudinal study of a large cohort and that smoking habits have been recorded. In addition, the cohort is based on a register with a high complete coverage of the Swedish construction industry. Finally, an internal control group with known smoking status has been used.

There may be some misclassification of diagnoses between asthma and COPD. According to ICD 8, priority regarding the underlying cause of death was given to COPD if it was mentioned together with asthma, regardless of the cause of death on the death certificate 29. This priority rule has been excluded in subsequent classifications, meaning that since 1986 the underlying cause of death has been coded according to the death certificate. As a result of this, for the period before 1987, some asthmatics may have been coded as having died from COPD, but this misclassification is probably independent of the subject's occupation.

It is probable that the workers in the control group had been exposed to dust, albeit at lower levels than in the exposed group, which may underestimate the risks. Furthermore, the current authors have not been able to estimate the cumulative exposure index, since the exposure is only based upon occupation reported at the first health examination. It should, however, be noted that, in general, Swedish construction workers remain in the same occupation within the construction industry. In those examined at least twice, >70% had the same occupation after 5 yrs. The job-exposure matrix was based on exposure estimations from the 1970s. Several of the workers (in particular those who were older) had probably worked in the construction industry since the 1930s, 1940s or 1950s. Therefore, exposure to asbestos, for example, may have occurred without being taken into account in the job-exposure matrix. In particular, the current authors believe that previous asbestos exposure among plumbers may be the reason for the increased risk of dying from COPD among workers exposed to metal fume, as the group of workers exposed to metal fume consisted mainly of plumbers (80%). There was also a considerable overlap between those classified as exposed to inorganic dust and those exposed to fumes, and to gases and irritants. In the Poisson regression model it was only exposure to inorganic dust that remained associated with COPD mortality when other major exposure groups were introduced into the model.

The classification of smoking habits in the study was not time dependent. Hence, the results from the whole group can be criticised, as different exposure groups may have different cumulative exposure to tobacco smoke, i.e. different pack-years. However, this has been overcome by only analysing never-smokers, which was possible because of the size of the data set. Nevertheless, smoking habits were only reported at the start of follow-up and, therefore, there is a possibility that the fraction of never-smokers who started to smoke was different in the different exposure categories. However, if the analysis among never-smokers was restricted to those with smoking information from repeated health examinations, the relative risk for any airborne exposure did not change (RR 2.11).

Although there is a large body of scientific literature describing studies of COPD concerned with impaired lung function and/or symptoms, this is, to the best of the authors' knowledge, the largest study investigating the relationship between COPD mortality and occupational airborne exposures. The power of the study made it possible to separately analyse different smoking categories. Hnizdo et al. 30 recently found an increased risk for COPD among US construction workers (odds ratio (OR) 1.2), especially among never-smokers (OR 3.4). In the Dutch Zutphen study, an increased risk for nonspecific lung disease was found among construction and cement workers 31. The present results corroborate those from previous studies, indicating an increased risk for COPD among workers exposed to dust, such as asbestos, man-made mineral fibres, mineral dust and concrete dust. However, it should be stressed that the dust exposure is construction related, i.e. the results cannot be directly applied to the general population.

The mechanism for the relationship between dust exposure and COPD remains unclear. Churg and coworkers 32–33, in two separate studies, have previously reported the presence of fibrosis in the small airways among autopsy cases with a previous history of occupational dust exposure. Increased elastic recoil pressure among workers exposed to man-made mineral fibres has also been described 34. Taken together, these findings indicate that occupational exposure to dust, especially mineral dust, causes a scarring or fibrosis in the small airways.

When analyses were restricted to never-smokers, the increased mortality remained, which to some extent is contrary to other studies 2, although it is in accordance with others 30. The attributable fraction among never-smokers with any airborne exposure was high, 53%. Hnizdo et al. 30 also found a high attributable fraction, 31%, among never-smokers. However, their study was based on the general population and the attributable fraction was a population-based estimate. In South African studies of gold miners, the results indicated a multiplicative interaction between smoking and dust exposure 35, 36. The data presented in this study rather suggest an additive component in the relationship between smoking and occupational dust exposure. However, it should be noted that none of the nonsmokers in the South African cohort died of COPD. It should be stressed that the estimate for COPD among never-smokers in the current study is based on 31 exposed and deceased subjects.

Two studies from the US general population have recently been published and show a population-attributable fraction of ∼20% for occupation and COPD 30, 37. This is higher than was found in the current study, especially considering only construction workers were included here. The disparity could be due to differences in exposure between the countries, but also due to the fact that other “nonexposed” construction workers were used as controls. In the general population, “unexposed” subjects probably have a lower exposure to dust than the controls used in this study. Furthermore, the US studies may also include other types of occupational exposure not occurring among construction workers.

In conclusion, occupational exposure to dust increases mortality due to chronic obstructive pulmonary disease, even in never-smokers.