INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic airway diseases contribute considerably to the burden of chronic morbidity in adults, and occupational exposures play a role in the onset of several chronic airway diseases. Occupational asthma is the most common occupational respiratory disease in industrialized countries (1). Many different occupations and more than 200 specific occupational exposures have been associated with asthma (2). It is also known that chronic obstructive pulmonary disease (COPD) can be induced by occupational exposures (3). There is, however, controversy about the role of cigarette smoking in occupational respiratory disease, and about the overlap with asthma (4, 6, 7).

Cigarette smoking, typically associated with occupation, is the major determinant of COPD. Hence, smoking is likely to be a confounding factor in the relationship between occupational exposures and COPD. Even when adjusting for smoking status, a residual confounding effect can be expected, since both COPD and occupation are also associated with smoking quantity, and not only with smoking status. Thus, it is better to evaluate associations between occupational exposures and COPD separately for smokers and nonsmokers (6) and to adjust for smoking quantity within the group of smokers. In addition, earlier studies have suggested that occupational exposures may interact with smoking (6).

The role of occupational asthma as a concurrent lung disease is another important consideration. It has been suggested that COPD may arise as a consequence of longstanding asthma (6), but this is probably only apparent in older patients. Furthermore, COPD can be attributable to different types of exposures, including agents known to cause occupational asthma. Occupational exposure to these agents, especially respirable irritants, can also lead to chronic productive cough. This so-called industrial bronchitis itself does not lead to airway obstruction, but the concurrent asthma may (6). Moreover, nonspecific bronchial responsiveness (NSBR) without apparent asthma is a risk factor for decline in lung function, and NSBR itself may also be induced by occupational exposures. Thus, it is important to understand whether COPD is related to occupational exposures independently of asthma.

Evaluating relationships between occupational exposures and COPD with stratification by asthma and smoking status requires a large study population. We analyzed data from the European Community Respiratory Health Survey (ECRHS), a multicenter study among young adults, both males and females, from European and other industrialized countries. Results of two national analyses of the ECRHS, dealing with occupational exposures and COPD, have been published so far (8, 9). A study in New Zealand (8) showed increased risks for chronic bronchitis and airway obstruction in workers who had reported occupational exposure and in certain job groups. Analyses of a subset of Spanish patients (9) revealed an association between exposures as assessed with a job exposure matrix (JEM) and chronic bronchitis and lung function.

A drawback of both studies was that analyses were done on relatively small groups, which did not allow subdivision of the study population in order to evaluate modifying factors. In addition, it was not known whether occupational exposures were related to bronchitis symptoms and lung function in other countries participating in the ECRHS. We performed analyses of data from 14 countries in the ECRHS to study the relationships between specific occupations and occupational exposures and symptoms of chronic bronchitis, lung function, and NSBR. All analyses were done with stratification by current asthma and by smoking status.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Population

The aims and methods of the ECRHS have been described elsewere (10). We used data from a random sample of the general population seen at 30 study centers in 14 countries (n = 15,111 subjects). Excluded from analyses were subjects with missing data for symptom questions (n = 283), cigar and pipe smokers and subjects with incomplete data on smoking habit (n = 680), and subjects without data about current or last-held occupation (n = 895; mainly full-time students).

Occupational History

The ECRHS classification was used for coding of occupations (11, 12). Subjects were classified by current occupation or, for subjects reporting a change of job because of problems with respiratory health, by occupation at that time. This was done to avoid selection bias (13). Subjects were further classified into 10 occupational groups with a high probability of exposure to vapors, gas, dust, or fumes, and into one reference group of professional, administrative, and clerical workers (11). Some potentially relevant job categories, such as mining, rubber and plastics work, and glass and ceramics work were too few (< 25) to constitute a separate group, and were not included in this part of the analyses (n = 2,336).

Occupational exposures were assessed directly from self-reported exposure to vapors, gas, dust, or fumes. We also used a JEM that provided a semiquantitative estimate (none, low, or high) of exposure to biologic dusts, mineral dusts, and gases/fumes for each occupation (9, 11).

Health Outcome

Chronic bronchitis was defined as the regular expectoration of phlegm for at least 3 mo in each year. A second definition was a regular cough with phlegm production for at least 3 mo each year. Asthma was defined as an attack of asthma in the previous 12 mo, having been awakened by an attack of shortness of breath at any time in the previous 12 mo, or current use of asthma medication (11).

Airway obstruction was evaluated through measurement of FEV₁. Subjects performed at least three acceptable reproducible lung function maneuvers done according to standard spirometric procedures. NSBR was defined as a 20% decrease in FEV₁ associated with a methacholine dose of  8 µmol, as described elsewhere (10). Values for FEV₁ and NSBR were available for 84% and 62% of the population, respectively.

Statistical Analyses

Analyses were performed separately for subjects with asthma (n = 1,099), and those without asthma (n = 12,154). We further stratified these two subgroups into never-smokers, ex-smokers, and current smokers. We calculated prevalence ratios (PRs) to assess the relative risks (RRs) for the categorical variables bronchitis and NSBR (14). PRs adjusted for potential confounders were calculated with Cox's proportional hazards model (15), as modified by Breslow (16), with the PHREG procedure of the Statistical Analysis System (SAS) (SAS Institute, Inc., Cary, NC) (17). Relationships between occupational exposures and FEV₁ were investigated through multiple linear regression analyses (REG procedure of the SAS). All models were adjusted for gender, age, height squared, and country. Analyses of data for current smokers were additionally adjusted for number of cigarettes smoked per day. In order to investigate geographic variations, we grouped countries into six regions; these consisted of two or three adjacent countries and one region that comprised the three non-European countries in the ECRHS.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Occupational exposure to vapors, gas, dust, or fumes was more common among males than among females (Table 1). Current smokers were more often exposed than were ex-smokers or never-smokers. There was a considerable range in exposure between countries, of from 33% to 58% for self-reported exposure and from 30% to 53% for any exposure according to the JEM. The age distribution was similar for never-smokers (mean: 32.5 yr) and current smokers (33.2 yr), whereas ex-smokers (35.2 yr) were somewhat older.

Among nonasthmatic subjects, chronic bronchitis was present in 1% to 3% of never-smokers and ex-smokers, and in 5% to 9% of current smokers, depending on the definition used (Table 2). An excess risk for chronic bronchitis was found among agricultural workers in never-smokers and current smokers. Textile, wood, and food processing workers showed an excess risk predominantly for ex-smokers and current smokers, and paper and chemical workers for current smokers only. Among cleaners, an excess risk was found for ex-smokers, and possibly for never-smokers, but not for smokers. Self-reported exposure to vapors, gas, dust, or fumes was related to chronic bronchitis in all smoking groups. Occupational exposure according to the JEM was associated with chronic bronchitis in current smokers only. The interaction between occupational exposure and smoking, however, was not statistically significant (p > 0.1). When applying the JEM separately for the three types of exposure, we found comparable PRs for biologic dust, mineral dust, and gases/fumes (results not shown).

Figure 1 shows associations between any occupational exposure according to the JEM and cough with phlegm among current smokers in six regions. PRs for cough with phlegm ranged from 1.2 and 2.0, with a trend toward lower risks in the northern European countries. The figure further shows that the risk for bronchitis associated with occupational exposures was higher in males, although the interaction of exposure with gender was not statistically significant. For ex-smokers and never-smokers, the association between occupational exposure and chronic bronchitis was similar for males and females (results not shown).

View larger version (8K): [in this window] [in a new window]   Figure 1.   Associations between occupational exposure to vapors, gas, dust, or fumes (JEM) and cough with phlegm among nonasthmatic smokers. Prevalence ratios and 95% confidence intervals shown are by geographical area and by gender.

To investigate possible residual confounding by level of cigarette consumption within the group of current smokers, we performed analyses in six strata according to the number of cigarettes smoked per day (1 to 5, 6 to 10, 11 to 15, 16 to 20, 21 to 25, and > 25). PRs for bronchitis associated with any occupational exposure assessed with the JEM were consistent among the different strata (results not shown). We found no trend toward an association of cigarette consumption with RR.

Mean FEV₁ in most job categories tended to be lower than in the reference group of office workers (Table 3), although differences were not statistically significant. A consistent trend across the three smoking groups could not be observed. A high level of exposure to vapors, gas, dust, or fumes according to the JEM was associated with lower lung function only in current smokers. This decrease of 61 ml in FEV₁ differed significantly (p < 0.01) from the decrease seen both in never-smokers and in ex-smokers. This was also true when biologic dusts, mineral dusts, and gases/fumes were considered separately. Self-reported occupational exposure was not associated with lung function. NSBR was not related to any of the occupational exposure indices.

Analyses were repeated for asthmatic subjects (Table 4). The prevalence of chronic bronchitis was three to six times higher in this group than among nonasthmatic subjects. Occupational exposure according to the JEM was similar to that of nonasthmatic subjects (42.3% versus 42.2%), but more asthmatic (52.6%) than nonasthmatic subjects (44.4%) had self-reported exposure. RRs for both bronchitis and NSBR, and regression coefficients for FEV₁, were not essentially different for asthmatic and nonasthmatic subjects. Self-reported exposure was associated with chronic bronchitis in all smoking groups, whereas exposure according to the JEM was associated with bronchitis in current smokers only. Subject numbers were too small to evaluate associations according to occupational groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this general population study, occupational exposures were related to chronic bronchitis. Associations were most obvious among cigarette smokers, but the interaction of occupational exposures with smoking was not statistically significant. Significantly increased risks for chronic bronchitis were found for agricultural workers among both smokers and never-smokers. Increased risks were also observed for other occupational groups including textile, wood, food, and paper and chemical processing workers, principally among smokers. Occupational exposures were not clearly related to either lung function or bronchial responsiveness, although smokers with a high occupational exposure had lower levels of FEV₁. Results of the study suggest the occurrence of industrial bronchitis without pronounced airflow limitation.

Associations between occupational exposures and chronic bronchitis in our study are consistent with findings in previous studies. Excess risks for working in agriculture, wood and textile and paper processing confirm results of earlier, work force-based (18) and general population studies (19). Associations between self-reported exposures and chronic bronchitis were also found in analyses of national subsamples of the ECRHS (8, 9) and in other general population studies (20- 23). Relationships between chronic bronchitis and occupational exposures assessed with a JEM were also observed in the Spanish subsample of the ECRHS (9) and in a longitudinal Dutch study (19).

Results obtained with the JEM in our study were consistent with excess risks in the occupational groups examined. Exposure to biologic dust is likely in agriculture, and the paper, wood, and food processing industries, whereas mineral dust exposure occurs in agriculture (24), and exposure to gases/ fumes occurs in the chemical industry, in agriculture, and in woodworking. The JEM used in the study was not gender- or country-specific. We observed nonsignificant differences in risks between males and females, and between northern Europe and other countries. This was probably due to quantitative or qualitative differences between genders and between countries in occupational exposures within particular jobs.

In contrast with earlier general population studies (20), we were unable to detect a clear-cut association between occupational exposures and lung function. A statistically significant effect on FEV₁ was found only for smokers with a high exposure according to the JEM. The overall lack of effects on lung function can probably be explained by our having studied a young sample, among which a considerable decline in lung function was not yet evident. A general population study in Norway (23) showed that the prevalence of chronic airflow limitation (FEV₁/FVC < 0.7) ranged from 2% to 4% in subjects aged 18 to 44 yr, against 9% to 12% in the age group from 45 to 73 yr. Young age also implies a relatively low cumulative occupational exposure, and probably a more recent and cleaner work environment than that of older subjects exposed in earlier decades and included in earlier studies. In the two national subsamples of the ECRHS analyzed earlier (8, 9), however, effects of occupational exposures on lung function were suggested. This might be due to differences in occupational exposures between countries, as suggested in our study (Figure 1).

The lack of association between occupational exposures and lung function may also be due to healthy worker bias. Workers with respiratory disorders are less likely to enter high-exposure jobs, or when employed are more likely to leave high-exposure jobs. It can be expected that this effect would be strongest among nonsmokers.

We performed all analyses separately in cigarette smokers, ex-smokers, and never-smokers. In earlier studies (8, 9, 20- 23), relationships between occupational exposures and chronic bronchitis were evaluated with adjustment for smoking status. When this was done in our data set, overall risks for exposure to vapors, gas, dust, or fumes on chronic bronchitis were close to those observed within the group of current smokers only. This indicates that smoking would have caused residual confounding in unstratified analyses. Risks of chronic bronchitis were more apparent in current smokers, but we could not formally demonstrate effect modification by smoking. In analyses done on occupational groups, this was probably due to limited numbers of subjects. In analyses done with the JEM, the groups were bigger, but overall RRs were low (< 2), and differences in RR between smokers and nonsmokers were small (the largest difference being 1.7 versus 1.0, respectively). Thus, we cannot exclude the presence of effect modification by smoking; however, if it exists, it is relatively small.

Limitations of this study include misclassification in exposure assessment and possible interference by socioeconomic status (SES). The main aim of the ECRHS was not to study occupational respiratory effects. Information on occupational exposures was based on current (or last held) job and on self-reported exposure. It is unlikely that reporting of current job was influenced by respiratory health status. Therefore, the analyses by occupational group and by JEM, both of which were based on the subjects current job, were probably not biased by differential misclassification. On the other hand, it is probable that subjects with respiratory symptoms were more aware of occupational exposures in the present or in the past. This might have led to differential misclassification when using self-reported exposure. This is probably reflected by the increased RRs of bronchitis symptoms associated with self-reported exposure (Table 2).

We did not adjust for SES. Occupation and SES are strongly interrelated. Therefore, adjustment for SES may remove part of the risk that could be attributed to occupational exposures that are widespread in some SES groups (25). After we adjusted for SES in our data set (using the age of completion of full-time education), RRs did not markedly change, although standard errors increased because of colinearity. This indicated that SES did not act as a strong confounding variable within the association under study.

In conclusion, occupational exposures were associated with the occurrence of chronic (industrial) bronchitis in young adults, being more obvious among cigarette smokers. Occupational exposures were not clearly related to lung function. This may have been due to the relatively young age and therefore the limited duration of exposure among these individuals, to more recent and cleaner work environments, and to a healthy worker bias. A follow-up of this population is currently underway, and will allow us to assess whether fixed airflow limitation (COPD) develops at an older age, especially among persons suffering from chronic bronchitis.