© 2002 American Thoracic Society
Occupational Airborne Exposure and the Incidence of Respiratory Symptoms and AsthmaDepartment of Thoracic Medicine; Centre for Clinical Research, Haukeland University Hospital; Institute of Medicine; and Section for Medical Statistics, University of Bergen, Bergen, Norway Correspondence and requests for reprints should be addressed to Tomas Eagan, M.D., Department of Thoracic Medicine, Haukeland University Hospital, N-5021 Bergen, Norway. E-mail: tomas.eagan{at}med.uib.no
Several prevalence studies have suggested an association between occupational exposure and respiratory symptoms and asthma, but there has been a lack of incidence studies to verify this. This study examined the incidence of respiratory symptoms and asthma in an 11-year Norwegian community cohort study with 2,819 subjects. Predictors examined were sex, age, educational level, lifetime exposure to quartz, asbestos, and dust or fumes, as well as smoking habits and pack-years. The prevalence of exposure to quartz, asbestos, and dust or fumes was, respectively, 3.7%, 5.0%, and 28.3% at baseline. In those exposed to dust or fumes, the odds ratios (95% confidence intervals) varied between 1.4 (1.1, 1.7) and 2.1 (1.3, 3.2) for developing respiratory symptoms or asthma after adjusting for sex, age, educational level, and smoking. Between 5.7% and 19.3% of the incidence of respiratory symptoms and 14.4% of the incidence of asthma were attributable to dust or fumes exposure after adjustment for sex, age, educational level, and smoking. In conclusion, airborne occupational exposure increases the incidence of respiratory symptoms and asthma, independent of sex, age, educational level, smoking habits, and pack-years.
Key Words: cohort study • incidence • asthma • respiratory symptoms • occupational exposure
When studying the associations of occupational exposures and respiratory symptoms or asthma, both studies of working populations and community samples have been conducted (1–19). A study of a working population usually allows for better classification of the specific exposures associated with different jobs. This can provide important evidence concerning a causal relationship of chemicals, fumes, or dusts to symptoms or asthma, if conducted longitudinally. However, studies of working populations are vulnerable to the healthy worker effect, in which there is a selection of susceptible workers out of the work force (20). This could theoretically underestimate the effects of the exposure. Moreover, the working population usually has high levels of exposure and gives a poor estimate of the impact of exposure on disease in the general population. To characterize the association of occupational exposure and disease in the population at large, a community study is the design of choice. A longitudinal community study conducted in Krakow, Poland between 1968 and 1981 reported an increased risk for developing several symptoms and asthma after exposures to dust, variable temperatures, and chemicals (13). A longitudinal study of a large part of the Finnish working population estimated the attributable fraction of occupational exposures to asthma, being 29% for men and 17% for women (21). However, the study lacked information on individual smoking habits, although an adjustment using smoking habits from the population at large was presented. A recent study from health maintenance organization registers in North America estimated that 21% of asthma cases were caused by occupational exposures (22). The aim of this study, on a Norwegian general population, was to examine the association between occupational exposure to quartz, asbestos, and dust or fumes and the incidence of respiratory symptoms and asthma in 1985–1996/1997. Furthermore, we estimated the fraction of the incidence of respiratory symptoms and asthma that could be attributed to smoking and dust or fumes exposure.
Subjects and Study Design In 1985, a random sample of 3,786 subjects aged 15–70 years, from the city of Bergen and 11 surrounding municipalities in western Norway, received a mailed questionnaire with 40 questions about respiratory symptoms and asthma, smoking habits, and occupational airborne exposure (23, 24). After two reminder letters, 3,370 subjects (89.0%) had responded. In late 1996, 189 of the responders in 1985 were deceased, leaving 3,181 subjects eligible for follow-up. The questionnaire was expanded to 58 questions; however, the wording of the questions regarding symptoms and disease was identical to the baseline study in 1985. A total of 2,819 subjects (88.6%) returned the questionnaire after two reminder letters and a telephone reminder (25).
The Questionnaires In addition, the symptoms were grouped into three clinically relevant complexes: (1) a chronic bronchitic symptom–complex, including those with incidence of chronic cough and phlegm cough, (2) a chronic obstructive symptom–complex, including those with incidence of chronic cough, phlegm cough, and dyspnea Grade 3; and (3) an asthmatic symptom–complex, including those with incidence of attacks of dyspnea and wheezing. The wording of the questions on occupational exposures was as follows: (1) Have you ever had a workplace with much dust or fumes in the air? (2) Have you ever been exposed to asbestos dust in your work, and (3) Have you ever been exposed to quartz dust or stone dust with quartz at work? These questions have been validated against a structured work history interview (27). We used educational level as an indicator of socioeconomic status (28), with three categories: those with up to 9 years of schooling (primary), those with a degree requiring 12 years of schooling (secondary), and finally, those with a higher degree of education (university). The definitions of smoking habits and pack-years, as well as their distribution during follow-up, have been given recently (29).
Statistical Analysis All first-order interactions between the exposure(s) in question and the other explanatory variables were estimated through separate analyses. Because of the large number of interactions tested, a nominal significance level of 0.01 was chosen for each. The attributable fraction was defined as the proportion of the incidence in the total population attributable to the exposure (31). Adjusted attributable fractions (32) with confidence intervals were estimated according to the method described by Greenland and Drescher (33). The statistical software SPSS 10.0 was used for the logistic regression analyses and STATA 6.0 for the analyses of attributable fractions (34).
In 1985, more than one in four subjects of this general population sample reported past or present occupational dust or fumes exposure (Table 1) . The exposure was reported more than three times as often in men as in women and more commonly in ever-smokers compared with never-smokers. Asbestos and quartz exposures were mainly reported by men and also more often in ever-smokers than in never-smokers. For asbestos and dust or fumes, the most frequent exposures were noted in the middle-aged group, whereas the highest exposure of quartz was seen in the middle-aged and oldest age group (Table 1).
For all symptoms and asthma, except attacks of dyspnea among the quartz-exposed, the crude cumulative incidences were higher among those exposed for all three exposures (Table 2) . For dust or fumes, the differences in cumulative incidences between those exposed and not exposed were highest for phlegm cough, dyspnea Grade 3, and asthma. The cumulative incidences of the symptom complexes were roughly twice as high for those exposed compared with the nonexposed. Those exposed to quartz or asbestos had higher incidences of the cough symptoms. However, those reporting prior asbestos exposure had a marked increase in the cumulative incidence of the symptom-complex wheezing and attacks of dyspnea, as well as asthma (Table 2).
The adjusted odds ratios for developing respiratory symptoms or asthma, if exposed compared with if not exposed, are shown in Table 3 . The odds ratio for developing asthma was 1.6 when exposed to dust or fumes compared with the nonexposed and varied for the respiratory symptoms between 1.4 for attacks of dyspnea and wheezing to 2.1 for dyspnea Grade 3.
To study whether the associations between occupational exposures and the respiratory symptoms were present also in subjects without asthma, additional analyses were performed after exclusion of subjects with asthma. No substantive changes were found in the odds ratios for the incidence of any of the respiratory symptoms. At follow-up, 11% of the baseline smokers reported having stopped, whereas 5% of the baseline never- or ex-smokers had taken up smoking. Further logistic regression analyses were performed adjusting for changes in smoking habits during the follow-up period. No overt differences in the exposure–disease relationships were noted, as compared with when adjusting only for smoking habits and number of pack-years at baseline. To examine the possibility of residual confounding by smoking, the analyses were also performed on the never-smokers only. For all symptoms except attacks of dyspnea, the odds ratios for the incidences were slightly higher among the never-smokers, and for asthma, the odds ratio nearly doubled to 3.0 (from 1.6, as shown in Table 3). Those reporting prior exposure to asbestos had significantly higher risks of developing phlegm cough, attacks of dyspnea, and the symptom–complexes chronic cough with phlegm cough and attacks of dyspnea with wheezing. For those reporting prior quartz exposure, there was a significantly higher risk for developing the symptom–complex chronic cough with phlegm cough (Table 3). When estimating all first-order interactions between the exposure in question and the other explanatory variables (sex, age, educational level, smoking habits, pack-years), none were found to be statistically significant (p = 0.01). In addition, separate analyses for the sexes revealed no significant differences in the adjusted odds ratios for dust or fumes exposure to the symptoms or asthma between males and females. The attributable fractions caused by occupational dust or fumes exposure varied from 5.7% for the incidence of wheezing to 19.3% for the incidence of dyspnea Grade 3 (Table 4) . For the symptom–complexes, the attributable fractions varied from 13.5 to 36.3% (Table 4). Except for dyspnea Grade 3, and the symptom–complex including dyspnea Grade 3, chronic cough, and phlegm cough, the attributable fraction was higher for smoking than occupational dust or fumes exposure. For the incidences of all the symptoms and asthma, the combined attributable fractions of smoking and occupational exposure were less than the sum of the separate contributions of the two exposure variables (Table 4).
Among the single symptoms, the attributable fractions for quartz and asbestos were highest for chronic cough, being 2.2% and 3.0%, respectively. The attributable fractions for quartz and asbestos to the symptom–complex chronic cough and phlegm cough were 4.0% and 6.4%, respectively.
In this study, there was an increased risk of developing respiratory symptoms and asthma if exposed to dust or fumes at baseline, even after adjusting for sex, age, educational level, smoking habits at baseline and during follow-up, and pack-years smoked before baseline. Both asbestos and quartz exposure were associated with a higher risk of the cough symptoms and asbestos exposure with a higher risk of attacks of dyspnea and asthma. The adjusted attributable fraction for developing asthma if exposed to dust or fumes was 14.4%. There are some methodologic issues to consider. First, in any epidemiologic study, nonresponse is associated with a risk of selection bias. In this study, the response rate was 89% at both baseline and follow-up. This corresponds to a total response rate for both baseline and follow-up of 79%. A high response rate such as this reduces the potential importance of selection bias (25). Second, we adjusted for smoking in terms of smoking habits and pack-years at baseline and changes in smoking habits during follow-up. Additional analyses of the never-smokers were performed. We consider the possibility of residual confounding by smoking small, due to the extensive smoking adjustment, but it cannot be excluded entirely. Third, we did not have data on occupational exposure during the follow-up period. However, within a community cohort, people tend to move out of air-polluted jobs and into cleaner jobs to a higher extent than vice versa (35). Hence, the occupational exposure during the follow-up in this study was probably less than indicated by the baseline occupational exposure. Still, we observed a significant exposure–disease relationship. The occupational exposure characterization is dichotomous without any indication of degree of exposure. This prevents an assessment of dose–response relationships. However, the question of exposure shows a high specificity (27), which is preferable when assessing a prevalent risk factor. Finally, because of the wide confidence intervals of the attributable fractions, these must be interpreted with caution. Also, the validity of the information on smoking and occupational exposure may differ and could hamper direct comparison of the attributable fractions caused by smoking and occupational exposure. Several previous prevalence studies from general populations have suggested a higher risk of respiratory symptoms (1, 4, 10, 12, 15, 17) and asthma (5, 16) if exposed to dust or fumes. The magnitude of the increased risk generally varies with odds ratios of 1.2 to 3.7 for respiratory symptoms (1, 4, 10, 12, 15) and 1.3 to 2.3 for asthma (10, 16). This is generally in accordance with the increased risk for the incidence of respiratory symptoms and asthma found in this study. In the study by Krzyzanowski and Jedrychowski (13), the association of dust, variable temperatures, and chemicals to the incidence of respiratory symptoms and asthma was examined in a population aged 19–60 years in Krakow, Poland. Those exposed to dust had a higher risk of developing chronic cough in both sexes, and among men, there was a higher risk for developing phlegm cough and attacks of dyspnea. Among women, there was a higher risk for developing wheeze in those exposed to dust compared with those unexposed. It is difficult to compare the odds ratios from this study to ours, as a log linear model was used, based on whether the subjects were exposed 8 or 13 years before the follow-up. Generally, the odds ratios for the incidence of the symptoms varied between 1.1 and 2.4. No significant association was found between dust exposure and the incidence of asthma or dyspnea when walking on level ground (13). A Dutch longitudinal community study (17) observed a higher incidence of chronic nonspecific lung disease in those occupationally exposed. However, this study comprised only older men of whom only a minority were nonsmokers, and thus, its generalizability is limited. Occupational agents may cause asthma through several mechanisms (36, 37). High molecular agents may act as complete antigens and cause an allergic reaction. Low molecular agents can cause an allergic reaction by acting as haptens, binding with proteins to become complete antigens. Low molecular agents may also bind directly to the ß2 adrenergic receptor to cause bronchial constriction, whereas others may act by stimulating sensory nerves or work through T lymphocyte activation. Studies on working groups have shown associations between mineral dust exposure and airflow limitation (38–41). In the workplace, the most common situation is one of mixed exposures. The questions on quartz and asbestos can be interpreted as indicators of an unhealthy working environment. Inferences of a causal relationship between quartz and asbestos and the incidence of respiratory symptoms and asthma must therefore be drawn with caution. However, at least one previous study from a general population has shown an association between self-reported quartz exposure and airflow limitation (42). Previous analyses from our sample have shown higher incidences of respiratory symptoms in women than men, after adjusting for age and smoking (29). In this study, we were not able to find sex differences, suggesting that the effect of occupational exposure is similar in both men and women. However, this must be interpreted with caution because of the small number of exposed women and the lack of data on degree and duration of exposure. Most prevalence studies have reported no sex difference in the effect of occupational exposure (1, 4, 12, 15, 16). We observed no interaction between smoking and airborne exposure on the incidence of the symptoms and asthma. Neither the Krakow study (13) nor previous cross-sectional community studies have observed such an interaction, with information from questionnaires. Of the prevalence studies examining the effect of a possible interaction on lung function, three out of four studies have not found an interaction (4, 24, 43). However, the study of 1,094 residents of Beijing, China observed an interaction between dust exposure and former smoking on lung function measured by FEV1/FVC and peak expiratory flow rate (15). Differences in both smoking habits and exposure between various studies could help explain the different findings regarding the interactive effects of smoking and occupational exposure. The attributable fraction in a population is interpreted as the proportion of the new cases of symptoms or asthma that would have been prevented if the exposure in question had not occurred. It is important to note that the effect of preventing one exposure in the situation where several exposures are involved will depend on the order in which the exposures are prevented (44). Two prior prospective studies have given estimates of the attributable fraction of asthma caused by occupation (21, 22). The study by Karjalainen and colleagues (21) estimated the attributable fractions caused by occupational factors to be 29% of asthma in men and 17% in women. This was, however, not adjusted for smoking on an individual level or for socioeconomic status. The study by Milton and colleagues (22) gave an overall estimate of 21% of asthma attributed to occupational factors in a register study of a large population of health maintenance members. A recent review of previously published estimates of the attributable fractions estimated that 9% of all cases of adult-onset asthma could be attributed to occupational factors (45). This study gives a somewhat higher estimate. However, in the mentioned review (45), the 12 highest scored studies gave a median attributable fraction of 15%, which is in accordance with this study. In conclusion, this study shows that occupational airborne exposure, measured in a general population setting, increases the risk for developing respiratory symptoms and asthma. This effect is independent of sex, age, educational level, and smoking habits. The fraction of new cases of asthma attributable to dust or fumes exposure was 14%.
Supported by the Norwegian Research Council and the Norwegian Asthma and Allergy Research Foundation. Received in original form March 25, 2002; accepted in final form June 28, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
of smoking or dust/fumes on the 11-YEAR cumulative incidence of respiratory symptoms and asthma
