Vapor, Dust, and Smoke Exposure in Relation to Adult-Onset Asthma and Chronic Respiratory Symptoms: The Singapore Chinese Health Study

Abstract

Occupational factors contribute to a significant fraction of respiratory disease and symptoms. The authors evaluated the role of occupational exposures in asthma, chronic bronchitis, and respiratory symptoms in the Singapore Chinese Health Study, a population-based cohort of adults aged 45–74 years at enrollment in 1993–1998. Information on occupations and occupational exposures was collected at enrollment for 52,325 subjects for whom respiratory outcomes were obtained via follow-up interviews in 1999–2004. Exposure to dusts from cotton, wood, metal, minerals, and/or asbestos was associated with nonchronic cough and/or phlegm (odds ratio (OR) = 1.19, 95% confidence interval (CI): 1.08, 1.30), chronic bronchitis (OR = 1.26, 95% CI: 1.01, 1.57), and adult-onset asthma (OR = 1.14, 95% CI: 1.00, 1.30). Cotton dust was the major contributor to respiratory symptoms. Vapor exposure from chemical solvents, dyes, cooling oils, paints, wood preservatives, and/or pesticides was associated with nonchronic cough or phlegm (OR = 1.14, 95% CI: 1.03, 1.27), chronic dry cough (OR = 1.55, 95% CI: 1.19, 2.01), and adult-onset asthma (OR = 1.34, 95% CI: 1.15, 1.56). Chemical solvents, cooling oils, and pesticides were the major contributors to respiratory symptoms. These data support the role of occupational exposures in the etiology of respiratory illness in a population-based cohort in Singapore with a low prevalence of atopic illness.

There is growing recognition that occupational exposures make a substantive contribution to adult-onset asthma, chronic bronchitis, and other respiratory symptoms (1, 2). The American Thoracic Society recently estimated that 15 percent of these conditions are work related (3). More than 300 different workplace exposures have been associated with occupational asthma or chronic bronchitis (4–6).

Exposures can be classified according to the possible pathogenic mechanisms. Sensitizing substances cause asthma by inducing specific immunoglobulin E antibodies (1, 7). These are mostly high-molecular-weight allergens of animal or vegetable origin. Others are low-molecular-weight compounds, such as organic and inorganic compounds that act as haptens interacting with a protein to form a complete antigen. Notably, an immunoglobulin E mechanism has not been demonstrated consistently for exposure to agents such as diisocyanates, western red cedar, and acrylates. The reaction between the antigen and specific immunoglobulin E antibodies initiates a cascade of events that result in an allergic inflammatory reaction in the airways. Sensitizing agents induce respiratory symptoms after a latent interval of several months—occasionally weeks or years following exposure (8). Agents that contribute to adverse respiratory symptoms by nonimmunologic mechanisms, mostly irritant vapors such as chlorine, sulfur dioxide, combustion products, and ammonia, may lead to bronchial epithelial cell damage, release of inflammatory mediators, and airway remodeling (9, 10). An irritant-induced respiratory symptom typically develops within hours of acute inhalation, although irritants can lead to chronic respiratory impairment (8).

Much of the data relating occupational exposure to respiratory disease and symptoms come from cross-sectional studies conducted in industrial groups with high levels of exposure (11, 12). However, these studies may suffer from the healthy-worker effect, selection bias due to affected workers leaving jobs involving high levels of exposure, which tends to result in underestimation of the true risk (13, 14). Because workers who have changed employment remain in the general population, this bias may be minimized by sampling subjects from the general population (15). Community-based studies have supported a role for occupational exposures in respiratory disease, but nearly all of these studies have been conducted in Western populations (16). We examined the relation between self-reported exposure to occupational dust (cotton, wood, metal, mineral), smoke (welding, coal burning, wood burning), or vapor (chemical solvents, dyes, cooling oils, paints, wood preservatives, pesticides) and adult-onset, physician-diagnosed asthma and symptoms of cough and phlegm in a large population-based cohort of adult Singaporeans of Chinese ethnicity, the Singapore Chinese Health Study.

MATERIALS AND METHODS

Study population

The design of the Singapore Chinese Health Study has been described previously (17). Briefly, the cohort was recruited between 1993 and 1998, drawn from permanent residents or citizens of Singapore who lived in government-built housing (86 percent of the Singapore population resides in such facilities). Men and women of Chinese ethnicity (Cantonese or Hokkien dialect), aged 45–74 years, were eligible. A total of 63,257 persons (∼85 percent of eligible subjects) were enrolled. A baseline questionnaire was administered in person by trained interviewers to collect data on smoking history, diet, medical history, and occupations. A follow-up telephone interview was completed between 1999 and 2004 for 52,325 cohort members. The average time between the baseline and follow-up interviews was 5.8 years. The follow-up interview included detailed information on exposure to environmental tobacco smoke; medical history, including physician diagnosis of hay fever, allergic rhinitis, sinusitis, or eczema; and items on asthma, cough, and phlegm from the American Thoracic Society questionnaire (ATS-DLD-78-A). These items were not obtained on the baseline questionnaire, and items on occupation were not obtained on the follow-up questionnaire.

The institutional review boards at the University of Southern California (Los Angeles, California), the National University of Singapore, and the National Institute of Environmental Health Sciences (Research Triangle Park, North Carolina) approved this study. Informed written consent was obtained from all participants. As of December 31, 2004, 7,722 deaths had occurred in the overall cohort.

We categorized the 52,316 persons for whom respiratory outcome data were complete (for nine, data on duration of symptoms were missing) into mutually exclusive categories consisting of physician-diagnosed adult-onset asthma (n = 1,426; defined as onset of symptoms after age 18 years), physician-diagnosed childhood asthma (n = 878), chronic dry cough (n = 364), chronic phlegm (n = 1,475), chronic cough plus chronic phlegm (chronic bronchitis, n = 417), or nonchronic cough and/or phlegm (n = 2,652). The referent group consisted of subjects who did not report any of the following: cough, phlegm, or asthma (whether childhood or adult onset; n = 45,104). In this paper, we do not present data for the 878 childhood-onset asthmatics because the illness likely began prior to occupational exposures. Chronic was defined as occurring on most days for at least 3 months of the year and lasting more than 2 years in a row. We first classified subjects according to physician-diagnosed asthma and then classified those without this report according to the presence of chronic cough and/or phlegm. Thus, subjects with asthma were classified as asthmatic, even if they also had cough and/or phlegm. Use of mutually exclusive categories enables all subjects to be compared with a common reference category of subjects free from asthma, cough, or phlegm and thus better answers the question of whether occupational exposures are related to respiratory outcomes in general.

Occupation was assessed by 45 questions on the baseline questionnaire regarding current or previous jobs and exposures plus corresponding questions on duration of exposure. Subjects were asked whether they worked in any of the following industries for 1 year or longer: manufacture and repair of boots, shoes, or other leather goods; manufacture of furniture and cabinets; cotton textile; electrical and electronic; lumber and sawmill, carpentry, or joinery; rubber and tire manufacturing; manufacture of dyes or dyestuffs; manufacture of paints; manufacture of plastics; petroleum; metal production or processing; construction work; and urea formaldehyde manufacturing. With respect to specific job categories, subjects were asked whether they had a job lasting 1 year or longer as a welder, textile machine mechanic, other mechanic, cotton spinner or weaver, painter, textile dyer, machinist, printer, tailor or seamstress, janitor or cleaner, vocational driver or food hawker, or cook or other kitchen worker. Subjects were also asked whether they were exposed on the job (lasting for 1 year or longer) to any of the following substances: cotton dust; wood dust; wood preservatives; metal dust or fumes; rock or mineral dust; smoke (all types); smoke from welding; smoke from burning coal or coke; smoke from burning wood; other smoke; pesticides; asbestos; coal tar; soot, pitch; acid or alkali solutions; chemical solvents; dyes or dyestuffs; cutting, cooling, or lubricating oils; paints; or formaldehyde. For each question, duration was also asked in categories of 1–4 years, 5–9 years, 10–14 years, 15–19 years, and ≥20 years.

Dust exposure included cotton dust, wood dust, metal dust, or mineral dust (including asbestos) in a job lasting 1 year or longer. Smoke derived from welding, coal burning, wood burning, or “other” sources was used as the composite smoke variable. Vapor exposure included wood preservatives; pesticides; chemical solvents; dyes; cutting, cooling, or lubricating oils; or paints. We did not ask information on the year in which exposure began or ended.

Statistical analyses

We examined exposures at baseline in relation to the outcome of respiratory symptoms reported on the follow-up questionnaire. Adjusted odds ratios and 95 percent confidence intervals were calculated from polytomous logistic regression models by using PROC CATMOD (version 9.1; SAS Institute, Inc., Cary, North Carolina) (18), except that when the expected cell number was less than 5, odds ratios were not calculated. We adjusted all models for age, gender, and dialect, although these adjustments did not necessarily alter the exposure-outcome associations. For most of the respiratory outcomes, subjects were more likely to be current smokers and to smoke more than those in the referent group; therefore, we carefully controlled for potential confounding by including terms in adjusted models for smoking status (never, current, past) and, for smokers, age at which smoking started and number of cigarettes smoked (19). The inclusion of current age and age at which smoking started captures smoking duration. Subjects were categorized as never smokers if they had never smoked at least one cigarette per day for 1 year or longer. We created a variable for atopy based on yes-or-no responses to questions on ever diagnosis of allergic rhinitis, sinusitis, hay fever, or eczema. Inclusion of terms for education, for exposure to environmental tobacco smoke as a child or adult (at work and/or at home), or for atopy did not alter the odds ratios (change of less than 10 percent) and thus were not included in the final models.

RESULTS

Study demographics

At follow-up, 2.7 percent of the cohort reported a diagnosis of adult-onset asthma, 0.7 percent reported chronic dry cough, 2.8 percent reported chronic phlegm, 0.8 percent reported the combination of chronic cough and phlegm (the definition of chronic bronchitis), and 5.1 percent reported cough and/or phlegm that did not meet the definition of chronic (table 1). Compared with participants without asthma, cough, or phlegm, those in all outcome categories were more likely to be current smokers, to have started smoking before the age of 20 years, and to smoke more cigarettes per day (table 1). Level of education differed modestly by case status. Not surprisingly, symptomatic and asthmatic participants were more likely to be atopic, although the prevalence of atopy was low, a finding consistent with other studies in Chinese populations (20).

Association between exposure and respiratory symptoms

Exposure to dust, a composite variable combined from reported exposure to cotton, wood, metal, minerals, and/or asbestos, was associated with nonchronic cough and/or phlegm (odds ratio (OR) = 1.19, 95 percent confidence interval (CI): 1.08, 1.30), chronic bronchitis (OR = 1.26, 95 percent CI: 1.01, 1.57), and adult-onset asthma (OR = 1.14, 95 percent CI: 1.00, 1.30) (table 2). Participants exposed to smoke from welding, coal burning, wood burning, or other sources, compared with controls, had an increased odds of nonchronic cough or phlegm (OR = 1.12, 95 percent CI: 1.01, 1.25) (table 2). Vapor exposure, derived from chemical solvents; dyes; cutting, cooling, or lubricating oils; paints; formaldehyde; wood preservatives; and/or pesticides, was associated with nonchronic cough and/or phlegm (OR = 1.15, 95 percent CI: 1.03, 1.27), chronic dry cough (OR = 1.55, 95 percent CI: 1.19, 2.01), and adult-onset asthma (OR = 1.34, 95 percent CI: 1.15, 1.56) (table 2). No differences were observed when persons reporting “never worked” were excluded from the analyses (data not shown).

We examined possible effect modification of the association of dust, smoke, and vapor exposure with respiratory symptoms by gender, smoking (ever, never), atopy, and fiber intake, a protective factor for chronic bronchitis symptoms in this cohort (21). No appreciable differences were found for gender or fiber intake (data not shown). Given the low prevalence of atopy, power to detect differences by atopic status was limited, and none were observed (data not shown) (20). Smoking status (never vs. ever) significantly modified the association between exposure to dust, smoke, or vapors and chronic dry cough (table 3). For chronic dry cough, the effects of all three broad exposure groups were greater in nonsmokers than in smokers, with p values for interaction all less than 0.04.

Although numbers became small, we looked at individual sources of dust and found the strongest association for the cotton dust variable (table 4). Cotton dust exposure was associated with nonchronic cough or phlegm (OR = 1.40, 95 percent CI: 1.19, 1.64), chronic phlegm (OR = 1.36, 95 percent CI: 1.09, 1.69), and chronic bronchitis (OR = 1.71, 95 percent CI: 1.13, 2.60). Duration of exposure was associated with a significant trend for all three outcomes (p < 0.03). Persons reporting exposure to cotton dust more often indicated work in the cotton textile industry (14 percent) or as a tailor or seamstress (71 percent) than in the other industries or occupations listed on our questionnaire. We found an association between being a tailor or seamstress and chronic bronchitis (OR = 1.35, 95 percent CI: 0.88, 2.06), chronic phlegm (OR = 1.39, 95 percent CI: 1.14, 1.70), or chronic dry cough (OR = 1.32, 95 percent CI: 0.88, 1.99).

Even though we did not find any statistically significant relation between the individual smoke components and respiratory symptoms (table 5), chronic phlegm was associated with coal burning (OR = 1.45, 95 percent CI: 0.95, 2.20), and adult-onset asthma was associated with wood burning (OR = 1.48, 95 percent CI: 0.92, 2.40). Of note, persons reporting exposure to smoke from coal or wood were frequently employed as food hawkers (41 percent and 26 percent, respectively) who cook food in covered outdoor or indoor stalls.

Given the association with vapor exposure, we examined the individual sources of vapors (table 6). Exposure to chemical solvents was associated with an increased odds of nonchronic cough or phlegm (OR = 1.20, 95 percent CI: 1.05, 1.37), chronic dry cough (OR = 1.63, 95 percent CI: 1.17, 2.27), and adult-onset asthma (OR = 1.44, 95 percent CI: 1.19, 1.74). The duration of exposure was associated with a significant trend in all cases (p < 0.03). Subjects reporting exposure to vapors were often in the construction (9 percent), metal processing (7 percent), and electronic (9 percent) industries and/or had occupations such as a mechanic (18 percent), vocational driver (9 percent), or painter (7 percent). In addition, being a cleaner or janitor, an occupation often involving exposure to chemical solvents, was significantly associated with adult-onset asthma (OR = 1.24, 95 percent CI: 1.03, 1.50).

Another source of vapors—cutting, cooling, or lubricating oils—was associated with chronic dry cough (OR = 1.46, 95 percent CI: 1.02, 2.08; p for trend = 0.05). Of those exposed to cutting, cooling, or lubricating oils, 12 percent reported working as a machinist and an additional 9 percent had worked in the metal processing industries.

Occupational pesticide exposure was limited in this general population cohort, yet we observed a moderate association with adult-onset asthma (OR = 1.69, 95 percent CI: 1.13, 2.52). Among persons reporting pesticide exposure, the most commonly reported types of work were construction (22 percent), vocational driver (10 percent), and/or janitor/cleaner (9 percent). We observed an increased odds, albeit of borderline statistical significance, between exposure to dyes/dyestuffs and nonchronic cough or phlegm (OR = 1.46, 95 percent CI: 0.97, 2.20), chronic phlegm (OR = 1.57, 95 percent CI: 0.94, 2.61), and adult-onset asthma (OR = 1.61, 95 percent CI: 0.92, 2.82). Persons exposed to dyes were most often printers (16 percent).

DISCUSSION

These data from the Singapore Chinese Health Study support the role of occupational exposures in the etiology of physician-diagnosed adult-onset asthma; symptoms of chronic bronchitis, chronic dry cough, or phlegm; as well as nonchronic cough and/or phlegm. Exposure to dust and vapors was especially important. Population-based studies from westernized countries have consistently demonstrated an increased relative risk of respiratory symptoms associated with occupational exposure to dusts, gases, and/or fumes (22–31); there are few studies in Asian populations (32).

Cotton dust exposure was the primary source of the association with dusts in this cohort. Work as a tailor/seamstress was much more common than work in the textile industry among subjects reporting exposure to cotton dust. Extensive literature links exposure to cotton dust, from working in the textile industry, to chronic respiratory symptoms and illness (33). Our study extends these observations to include an increased risk of chronic bronchitis, chronic phlegm, and chronic dry cough for workers exposed primarily as tailors and seamstresses, who would be expected to have much lower levels of exposure than workers in the textile industry.

Although sensitizations to wood dust, metal dust, or mineral dust are well-recognized causes of respiratory symptoms, our population included few exposed subjects, and thus we were not well powered to explore these associations. It is also possible that exposures may be lower in this population than in others, the component exposures may differ, and/or our questions on occupational history may have lacked specificity to capture these exposures.

In our Singapore cohort, we found an association between chemical solvent exposure and nonchronic cough or phlegm, chronic dry cough, and adult-onset asthma. We did not ask about the type of solvent used; therefore, we were limited in our ability to identify the potential etiologic agents. However, chemical solvents are known respiratory irritants affecting the mucosa (34, 35). Chemical solvents have a wide variety of industrial applications, including the manufacture of paints, inks, cleaning products, adhesives, and petrochemicals. Of note, we found that persons working as janitors or cleaners had higher odds of adult-onset asthma compared with those without that job history. A modest increased risk from working as a janitor or cleaner is supported in other population-based studies (36) from the European Community Respiratory Health Survey study (37), the United States (38), and Finland (11). In a community-based study involving seven French cities, asthma was not related to the occupation of cleaner, but exposure to industrial cleaning agents was (15). Our results add to the growing body of evidence for a role of cleaning agents in asthma.

Exposure to cooling, lubricating, and cutting oil aerosols occurs at the highest levels in the automotive industry and has been found to be associated with a variety of respiratory symptoms, including cross-shift changes in pulmonary function, asthma, cough, and phlegm (39–41). Our self-reported exposure category no doubt was subject to a high level of misclassification. Nonetheless, in our Singapore cohort, persons reporting exposure to cutting, cooling, or lubricating oils mentioned an increased prevalence of symptoms of nonchronic cough or phlegm, chronic dry cough, and physician-diagnosed adult-onset asthma. Microbial growth in the predominant aqueous component of cooling and lubricating oils may drive the association with respiratory illness (39, 42).

Pesticide exposure in this cohort was associated with adult-onset asthma. We did not have information regarding the use of specific pesticides; however, various pesticides, including organophosphates, carbamates, fungicides, fumigants, and paraquat, can produce adverse respiratory effects (43). Pesticide exposure contributes to wheeze in agricultural workers (44) and other outdoor workers (45). Agriculture is not a major industry in the small city state of Singapore: persons exposed to pesticides most commonly worked in the construction industry or as a janitor/cleaner. Exposure to pesticides in the construction industry may be a result of chemical pesticide treatment of wood for outdoor construction projects (45). We suggest that construction workers and cleaners may be exposed as a result of pesticide treatment of homes or buildings.

This study enhances the existing literature because there are few data on occupational exposure in general population studies in Asia. Singapore provides a unique opportunity within Asia; ambient air pollution levels are lower than in most Asian cities and are comparable to those in US cities (46, 47). Workplace smoking is illegal in Singapore (48), and there is a relatively high proportion (72 percent) of never smokers. In addition, since Singapore is tropical, our cohort was not exposed to coal or wood burning in the home for heating, unlike in China. Thus, confounding by these exposures was minimized.

The magnitude of the effects we observed was modest, and there is always the theoretical concern that unknown biases may contribute to modest associations. Nonetheless, the large size of this data set gave us excellent power to detect associations of the modest size that we would expect between occupational exposure and chronic respiratory symptoms.

We lacked information on the precise timing of exposure in relation to the start of respiratory symptoms; therefore, we were unable to address the impact of exposures on work-aggravated asthma, which is a significant component of the effects of occupation (49). In addition, we were unable to address whether workers may have selected “cleaner” occupations because of the early presence of respiratory symptoms. If such a bias were operating, the impact would be that we underestimated the associations.

Although we lacked data on the precise timing of exposures, occupational exposures were assessed on the baseline questionnaire and respiratory symptoms were assessed 5 years later at follow-up, thereby reducing biased recall of occupational exposure contingent on respiratory symptoms, or vice versa. Differential misclassification of self-reported exposures may occur by asthma diagnosis (50). However, in this cohort, we found associations for symptoms not closely related to asthma rather than simply an asthma diagnosis.

A total of 5,339 subjects died prior to the follow-up interview, creating the possibility of bias. However, a strength of this study is that we had cause-of-death data on decedents and found that, for subjects who died without follow-up, exposure to dust, fumes, or vapors was not related to death from respiratory disease as opposed to other causes. This finding suggests that we did not have significant nonresponse bias due to death.

We first categorized subjects on the basis of asthma and then secondarily on cough and phlegm. Thus, there will be a slight loss of power for analyses of cough and phlegm. However, the overlap was minimal; among the asthmatics, 6 percent reported chronic bronchitis, 2 percent reported chronic dry cough, and 6 percent reported chronic phlegm.

A self-report of physician-diagnosed asthma was used to classify subjects as asthmatic. It has been found that questions on “reported asthma” and “physician-diagnosed” asthma have good positive-predictive value and specificity in predicting current asthma (51). We reinterviewed 331 of 406 persons who reported incident asthma on the follow-up questionnaire (82 percent), and 313 of them (95 percent) reconfirmed an asthma diagnosis or asthma symptoms.

There is growing evidence from large population-based studies that a sizable proportion of chronic obstructive pulmonary disease (COPD) may be attributable to workplace exposures (52). Although we did not investigate COPD per se because of the unreliability of self-report of this disease in the general population (53), we included chronic bronchitis, a symptom complex of COPD defined by questionnaire responses alone. However, some people with COPD may not have symptoms and thus were included in the referent group with “no symptoms.” Inclusion of people with COPD in the referent group would tend to attenuate associations.

In this cohort of Chinese Singaporeans, we found significant associations, independent of tobacco smoking or exposure to environmental tobacco smoke, between occupational exposures and adult-onset asthma, chronic bronchitis, chronic cough and/or phlegm, and nonchronic cough and/or phlegm. Specific exposures included cotton dust, chemical solvents, cooling and lubricating oils, and pesticides. Given that a substantial proportion (42.5 percent) of our subjects were older than the common Singaporean retirement age of 62 years, our data suggest that occupational exposures can have long-lasting effects on respiratory symptoms.

This research was funded by the National Institutes of Health (NCI RO1 CA80205, NIEHS Z01 ES43012, and NIEHS KO1 ES00386).

The authors thank Siew-Hong Low of the National University of Singapore for supervising the fieldwork of the Singapore Chinese Health Study, Kazuko Arakawa of the University of Southern California for developing and managing the cohort study database, and Marsha Shepherd of Westat, Inc., for programming.

Conflict of interest: none declared.

References