Exposure-Response Relationships between Occupational Exposures and Chronic Respiratory Illness: A Community-based Study1.2
XIPING
xu,
DAVID C. CHRISTIANI, DOUGLAS W. DOCKERY, and LIHUA WANG
Introduction SUMMARY Data from a random sample of 3,606 adults 40 to 69 yr of age residing In BeiJing, China,
Accumulating evidence from both were analyzed to investigate the association of reported occupational exposures to dusts and workforce-based and community-based gaseslfumes with the prevalence of chronic respiratory symptoms and level of pulmonary function. studies consistently suggests that occupaThe prevalence of occupational dust exposure was 32%, and gas or fume exposure, 19%. After tional exposure to dusts, gases, and/or we adjusted for age, sex, area of residence, smoking status, coal stove heating, and education, fumes is a risk factor for chronic respiraan Increased prevalence of chronic phlegm and breathlessness was significantly related to both types of exposures. Chronic cough was significantly related only to dust exposure, and persistent tory diseases (1, 2). Such an association ,. wheeze only to fume exposure. The global estimates of the relative odds of the four symptoms is biologically plausible, Le., occupational were 1.30 (95% CI [confidence interval] 1.09 to 1.48) and 1.27 (95% C11.09 to 1.48), respectively, agents contain a number of toxicants or for dusts and for gaseslfumes. These two occupational exposures are associated with chronic respimodifiers of the effects of toxicants that ratory symptoms Independent of smoking, gender, and each other. There was an Increasing prevaare inhaled, bringing these toxicants inlence of each symptom with Increasing dust and fume exposure, represented by the Index of cumuto intimate contact with the airway lative exposure duration and exposure Intensity. Linear trends for Increased prevalence of chronic and/or parenchyma. Interactions bebronchitis and breathlessness were significant for both exposures, while the linear trend for wheeze tween occupational exposure and cigawas only significant for gaseslfumes. Among subjects who did not report using coal stove heating, rette smoking may act to increase the dust exposure was a significant predictor for FEV1 , FEV1/FVC, FEF25 - 75 , and peak expiratory flow prevalence of chronic respiratory symprete (PEFR). There was also a significant decrease for FEV1 and FVC with Increase of gas/fume exposure levels. Both current and former smokers appeared to be more susceptible to the effect toms or to' accelerate the decline of pulof dusts than the never smokers. There was a significant Interaction between smoking and dust monary function (1). exposure on FEV and PEFR. In addition, dust exposure appears to Interact with gaseslfumes Workforce-based epidemiologic studies to reduce PEFR. These results from a Chinese population add to the body of evidence suggesting allow measurement of the concentration that nonspecific occupational dust and fume exposure Is a risk for the development of chronic obof specific occupational agents in the AM REV RESPIR DIS 1992; 148:413-418 structive puImonary disease. working environment and, therefore, offer the accurate assessment of exposures with less measurement error and misclassification than community-based studies. However, investigations of actively em- analysis and internal comparisons with- from Beijing, China, was analyzed (1) to ployed groups may be biased by the se- in exposed groups (high exposure versus assess the exposure-response relationlection of workers who are less suscepti- low exposure) can help to clarify the is- ships between occupational exposures to ble to the effects of specific exposures. sue and provide evidence for assessing dusts and gases/fumes and chronic reSusceptible subjects may change jobs or causality. spiratory symptoms and pulmonary. leavethe workforce entirely, may develop Several community-based studies in chronic respiratory disease after leaving Europe and the United States have shown the exposed jobs, or may die sooner than that occupational exposures are associatthe "survivor" occupational cohort (3-5). ed with increased prevalence rates of re- (Received in original form November 4, 1991and Subjects in community-based studies are spiratory symptoms and impairment of in revised form March 30, 1992) selected regardless of their current oc- pulmonary function (4, 5, 7-10). Of these Prom the Environmental Epidemiology and Occupational status, and, thus, the effect studies, only two addressed the ex- cupational Health Programs, Department of Enof selection bias is decreased. But these posure-response relationship between oc- vironmental Health, Harvard School of Public studies may still be affected by a "healthy cupational agents and chronic respiratory Health; Department of Medicine, Pulmonary and worker effect" (6), because those with symptoms (5, 8). The importance of rec- Critical Care Unit, Massachusetts General Hospibetter health may be selected into the ognition of an exposure-response rela- tal, and Harvard Medical School; the Channing Laboratory, Brigham and Women's Hospital, Bosworkforce by means of a preemployment tionships in establishing causality has ton, Massachusetts; and Department of Environhealth examination. Such a bias would been reemphasized in a recent review by mental Health, School of Public Health, Beijing tend to underestimate the effects of oc- Becklake (2) of the role of occupational Medical University, Beijing, China. 2 Correspondence and requests for reprints cupational agents. Dichotomized oc- exposures on the rise in airway obstrucshould be addressed to Xiping Xu, M.D., Environcupational exposures, Le., exposed ver- tive disease mortality. mental Epidemiology Program, Harvard School sus nonexposed, provide only limited inThis cross-sectional respiratory health of Public Health, 655 Huntington Avenue, Bosformation on effects. Exposure-response study of a general population of sample ton, MA 02115. 1/FVC
1
413
XU, CHRISTIANI, DOCKERY, AND WANG
414
function; and (2) to examine the interactions between cigarette smoking and occupational agents on respiratory symptoms and pulmonary function. Methods Data from a random sample of adults 40 to 69 yr of age residing in three areas of Beijing were used to investigate the relationships between occupational exposures and chronic respiratory health. The methods of selectionand characteristics of this population have been described elsewhere (11). Briefly, this study sample was drawn from the 1982 National Census Records in the following three areas: Dongchen (residential area), Haidian (suburban area), and Shijinshan (industrial area) by using a two-stage sampling method (12).The sampling unit in the first stage was the administrative unit of residence, i.e., the village in Haidian, and the residential committee in the other two areas. In the second stage, the sampling unit was the subject. The survey was conducted between August 1 and September 30, 1986. Information about occupational exposure to dusts and gases/fumes, exposure duration, and exposure intensity was obtained by trained interviewers using a modified Chinese translation of the ATS-DLD questionnaire (11, 13). Intensity of dust exposure was classified into three categories: low (a little but not too much), moderate (when finished work there was a lot of dust on the subject's clothes), and high (one cannot see the person just a few yards away). Intensity of gas/fume exposure was also classified into three categories: low (occasional exposure), moderate (frequent exposure but concentration not perceived to be high), and high (very frequent or daily exposure to high concentrations). A cumulative exposure index was created by grouping subjects into three exposure categories based on both exposure duration and exposure intensity. High cumulative exposure indicates that subjects had been exposed to dusts (or gases/fumes) over 10 yr with moderate or high intensity, and low cumulative exposure for less than 10 yr with low intensity. The others were included in the medium exposure group. This index was used in all exposure-response analyses. Respiratory symptoms were likewise determined using a Chinese translation of the standard questions (11, 13, 14).Chronic cough was defined as cough for 3 or more months of the year, chronic phlegm as sputum production for 3 or more months of the year, breathlessness as shortness of breath when walking with other people of one's own age on level ground, and persistent wheeze as wheezing on most days or nights. Chronic bronchitis was defined as the presence of both chronic cough and chronic phlegm. Pulmonary function measurements were performed in accordance with ATS performance and reproducibility criteria (15). Subjects performed vital capacity (VC) and FVC tests on the electric auto-spirometer (AS-300;
Minato Co., Japan) while in the standing position, with nose clip, either at the central station or at home. A minimum of three acceptable measurements was performed. The measurement with the highest FVC (the FEV. from the highest FVC) was used in the analysis. All of the pulmonary function values are reported at ambient temperature and pressure saturated with water (ATPS). The survey was conducted between August and September, during which the indoor temperatures were relativelystable, ranging from 22° C to 29° C. The pulmonary function parameters analyzed in this report were FEV., FVC, FEV./FVCOJo, FEF2 s - 7S , and FEF2 S- 7S/FVCOJo, and PEFR. FEV., and FVC measurements were standardized for height (HT) by dividing by HT 2 (11, 16). A previous report showed that the association between air pollution and FEV. for subjects who did not report using coal stove heating was stronger than for subjects who reported using coal stove heating, which suggests that the unexposed subjects may have been more sensitiveto airborne contaminants (11). In addition, the coal was burned in household stoves for heating with low thermal efficiency,no desulfurization, inefficient dust collection, and no or low smokestacks. It may not be appropriate to include the subjects who reported using coal stove heating in nonexposure group. Thus, the analyses of pulmonary function wererestricted to subjects who did not report using coal stove heating. Multiple logistic regressions were used to assess the association of occupational exposure with chronic respiratory symptoms. Because the symptoms of interest for each subject were correlated, the quasi-likelihood ideas of Liang and Zeger were used to model simultaneously these multiple binary correlated respiratory symptoms (17-19). The global estimate obtained from estimation equation represents the mean effect 0 f exposure on all symptoms of interest (18).In an exposure-response trend analysis, the values, 0, 1,2, and 3, were assigned to each level of exposure, respectively, to dusts and gases/fumes. The likelihood-ratio test statistic was used to test whether the assumption of the linear trend for symptoms was valid (20, 21). The effects of exposure levels based on assigned values will be reported, if the test results indicate that linearity of the trend with dusts or gases/ fumes is adequate. Multiple linear regressions were used to estimate the impact of occupational agents on pulmonary function. The regression coefficients were estimated by ordinary least-square (OLS) method. The variance was calculated by robust method, which can provide consistent estimator of OLS variance regardless of the presence of heteroskedasticity (22). The product terms of agents were used to assess the interactions between smoking and occupational exposures and between dusts and gases/fumes.
Results
There were 3,606(1,762 men, 1,844wom-
en) out of 3,746 subjects with complete information on respiratory symptoms, occupational exposure history, and covariates. The prevalence of occupational dust exposure was 32070 and gas/fume exposure 19070 for the overall population. Men had higher prevalence of both dust exposure and gas/fume exposure than women (table 1).Of the three areas studied, Shijinshan, the industrial area, had the highest exposures to both dusts and gases/fumes, although women in Dongchen had as much gas/fume exposure as those in Shijinshan. The prevalence of occupational exposure to either dusts or gases/fumes decreased with age except for men aged 40 to 59 yr. Exposure frequency also decreased with education, although subjects in the lowest education level had less gas/fume exposure than those in other groups in both men and women, and for dust exposure among women. Subjects who reported using central heating systems had significantly higher rates of occupational exposure than those who reported using coal stove heating, reflecting the predominant use of central heating systems in the industrialized area, Shijinshan, compared with the residential area, Dongchen. The most frequently reported specific occupational exposures (table 2) for dusts are coal (19.6070) and for gases/fumes were nonspecific gas (34.4070). Exposure to coal dust among these workers was not in mines.
Respiratory Symptoms The prevalence of respiratory symptoms was calculated by reports of occupational exposures (table 3). The relative odds of chronic respiratory symptoms were calculated for subjects with and without a history of dust exposure, after we adjusted for residential area, age, sex, cigarette smoking status, indoor coal stove for heating, and education level (table 3). Chronic phlegm and breathlessness were significantly associated with both dust and gas/fume exposure, while chronic cough only with dusts and wheeze only with gases/fumes. The global estimate of relative odds of the four symptoms was 1.30 (95070 CI 1.09 to 1.48) and 1.27 (95070 CI 1.09 to 1.48), respectively, for dusts and gases/fumes (table 3). Twenty-three percent of exposed subjects reported exposures to both dusts and gases/fumes. The interaction terms of the exposures with respiratory symptoms were tested using logistic models, but the results were not significant (p = 0.28 to 0.68). When both dust and
RELATIONSHIPS BETWEEN OCCUPATIONAL EXPOSURES AND CHRONIC RESPIRATORY ILLNESS
415
TABLE 1 PERCENTAGE DISTRIBUTION OF OCCUPATIONAL EXPOSURES TO DUSTS, GASES/FUMES BY SEX AND BY AREA, AGE, EDUCATION LEVEL, SMOKING STATUS AND HEATING TYPE, AMONG 3,606 ADULTS IN BEIJING, CHINA Male
Total Area Dongcheng Haidian Shijinshan Age, yr 40-49 50-59 60-69 Education Illiterate Primary sch. High sch. > High sch. Smoking Never Ex Current Heating Coal stove Other
Female
Subjects n (0/0)
Dust Exposure (%)
Gas/Fume Exposure (%)
Subjects n (0/0)
Dust Exposure (0/0)
Gas/Fume Exposure (%)
1,762
37.7
24.0
1,844
27.0
14.4
541 (31) 610 (35) 611 (35)
30.9 30.2 51.4
26.8 12.8 32.6
585 (32) 618 (34) 641 (35)
30.8 13.9 36.2
19.2 4.9 19.2
606 (34) 722 (41) 434 (25)
41.4 41.3 26.7
29.2 22.9 18.4
758 (41) 684 (37) 402 (22)
34.8 25.9 14.2
19.3 13.9 6.0
308 693 604 157
(18) (39) (34) (9)
41.2 38.7 35.9 33.8
15.6 26.0 26.8 20.4
881 515 386 62
(48) (28) (21) (3)
22.4 32.6 29.5 30.7
10.2 17.7 19.2 16.1
392 (22) 209 (12) 1,161 (66)
34.4 41.2 38.2
25.5 23.0 23.6
1,184 (64) 156 (8) 504 (27)
26.5 26.9 28.2
14.9 15.4 12.9
31.7 49.8
20.2 31.4
1,225 (66) 619 (34)
23.0 34.9
13.2 16.6
I"
1,173 (67) 589 (33)
TABLE 2 FREQUENCY OF SPECIFIC OCCUPATIONAL EXPOSURES TO DUSTS AND GASES/FUMES Dusts
Gases/Fumes
Type
n
Coal Lime Metal Cement Cotton Wood, paper, grass Metallurgical Silica Others
228 160 142 128 96 74 59 37 239
19.6 13.8 12.2 11.0 8.3 6.4 5.1 3.2 20.6
1,163
100.0
Total
Type
%
n
%
Gas· Solvent Gasoline, kerosene Acid gas Acetylene, ethylene, methane Pesticide Others
236 102 102 63
34.4 14.9 14.9 9.2
36 30 118
5.2 4.4 17.2
Total
687
100.0
* Nonspecific gas.
TABLE 3 PREVALENCE (0/0) OF CHRONIC RESPIRATORY SYMPTOMS AND ESTIMATED ODDS RATIOS· OF EXPOSURES TO DUSTS AND GASES/FUMES AMONG 3,606 ADULTS Occupational Exposure No
Yes
Odds Ratio
95% Confidence Interval
Dusts Cough Phlegm Breathlessness Wheeze Global
16.1t 22.6 20.9 6.7
20.6 28.0 25.4 7.0
1.32 1.27 1.39 1.02 1.30
1.09-1.59 1.07-1.50 1.16-1.65 0.77-1.36 1.14-1.48
Gases/fumes Cough Phlegm Breathlessness Wheeze Global
17.0 23.3 21.7 6.2
19.9 28.5 25.2 9.2
1.17 1.21 1.31 .; 1.62 1.27
0.94-1.46 1.00-1.48 1.06-1.61 1.18-2.21 1.09-1.48
Symptoms
XIPING
xu,
DAVID C. CHRISTIANI, DOUGLAS W. DOCKERY, and LIHUA WANG
Introduction SUMMARY Data from a random sample of 3,606 adults 40 to 69 yr of age residing In BeiJing, China,
Accumulating evidence from both were analyzed to investigate the association of reported occupational exposures to dusts and workforce-based and community-based gaseslfumes with the prevalence of chronic respiratory symptoms and level of pulmonary function. studies consistently suggests that occupaThe prevalence of occupational dust exposure was 32%, and gas or fume exposure, 19%. After tional exposure to dusts, gases, and/or we adjusted for age, sex, area of residence, smoking status, coal stove heating, and education, fumes is a risk factor for chronic respiraan Increased prevalence of chronic phlegm and breathlessness was significantly related to both types of exposures. Chronic cough was significantly related only to dust exposure, and persistent tory diseases (1, 2). Such an association ,. wheeze only to fume exposure. The global estimates of the relative odds of the four symptoms is biologically plausible, Le., occupational were 1.30 (95% CI [confidence interval] 1.09 to 1.48) and 1.27 (95% C11.09 to 1.48), respectively, agents contain a number of toxicants or for dusts and for gaseslfumes. These two occupational exposures are associated with chronic respimodifiers of the effects of toxicants that ratory symptoms Independent of smoking, gender, and each other. There was an Increasing prevaare inhaled, bringing these toxicants inlence of each symptom with Increasing dust and fume exposure, represented by the Index of cumuto intimate contact with the airway lative exposure duration and exposure Intensity. Linear trends for Increased prevalence of chronic and/or parenchyma. Interactions bebronchitis and breathlessness were significant for both exposures, while the linear trend for wheeze tween occupational exposure and cigawas only significant for gaseslfumes. Among subjects who did not report using coal stove heating, rette smoking may act to increase the dust exposure was a significant predictor for FEV1 , FEV1/FVC, FEF25 - 75 , and peak expiratory flow prevalence of chronic respiratory symprete (PEFR). There was also a significant decrease for FEV1 and FVC with Increase of gas/fume exposure levels. Both current and former smokers appeared to be more susceptible to the effect toms or to' accelerate the decline of pulof dusts than the never smokers. There was a significant Interaction between smoking and dust monary function (1). exposure on FEV and PEFR. In addition, dust exposure appears to Interact with gaseslfumes Workforce-based epidemiologic studies to reduce PEFR. These results from a Chinese population add to the body of evidence suggesting allow measurement of the concentration that nonspecific occupational dust and fume exposure Is a risk for the development of chronic obof specific occupational agents in the AM REV RESPIR DIS 1992; 148:413-418 structive puImonary disease. working environment and, therefore, offer the accurate assessment of exposures with less measurement error and misclassification than community-based studies. However, investigations of actively em- analysis and internal comparisons with- from Beijing, China, was analyzed (1) to ployed groups may be biased by the se- in exposed groups (high exposure versus assess the exposure-response relationlection of workers who are less suscepti- low exposure) can help to clarify the is- ships between occupational exposures to ble to the effects of specific exposures. sue and provide evidence for assessing dusts and gases/fumes and chronic reSusceptible subjects may change jobs or causality. spiratory symptoms and pulmonary. leavethe workforce entirely, may develop Several community-based studies in chronic respiratory disease after leaving Europe and the United States have shown the exposed jobs, or may die sooner than that occupational exposures are associatthe "survivor" occupational cohort (3-5). ed with increased prevalence rates of re- (Received in original form November 4, 1991and Subjects in community-based studies are spiratory symptoms and impairment of in revised form March 30, 1992) selected regardless of their current oc- pulmonary function (4, 5, 7-10). Of these Prom the Environmental Epidemiology and Occupational status, and, thus, the effect studies, only two addressed the ex- cupational Health Programs, Department of Enof selection bias is decreased. But these posure-response relationship between oc- vironmental Health, Harvard School of Public studies may still be affected by a "healthy cupational agents and chronic respiratory Health; Department of Medicine, Pulmonary and worker effect" (6), because those with symptoms (5, 8). The importance of rec- Critical Care Unit, Massachusetts General Hospibetter health may be selected into the ognition of an exposure-response rela- tal, and Harvard Medical School; the Channing Laboratory, Brigham and Women's Hospital, Bosworkforce by means of a preemployment tionships in establishing causality has ton, Massachusetts; and Department of Environhealth examination. Such a bias would been reemphasized in a recent review by mental Health, School of Public Health, Beijing tend to underestimate the effects of oc- Becklake (2) of the role of occupational Medical University, Beijing, China. 2 Correspondence and requests for reprints cupational agents. Dichotomized oc- exposures on the rise in airway obstrucshould be addressed to Xiping Xu, M.D., Environcupational exposures, Le., exposed ver- tive disease mortality. mental Epidemiology Program, Harvard School sus nonexposed, provide only limited inThis cross-sectional respiratory health of Public Health, 655 Huntington Avenue, Bosformation on effects. Exposure-response study of a general population of sample ton, MA 02115. 1/FVC
1
413
XU, CHRISTIANI, DOCKERY, AND WANG
414
function; and (2) to examine the interactions between cigarette smoking and occupational agents on respiratory symptoms and pulmonary function. Methods Data from a random sample of adults 40 to 69 yr of age residing in three areas of Beijing were used to investigate the relationships between occupational exposures and chronic respiratory health. The methods of selectionand characteristics of this population have been described elsewhere (11). Briefly, this study sample was drawn from the 1982 National Census Records in the following three areas: Dongchen (residential area), Haidian (suburban area), and Shijinshan (industrial area) by using a two-stage sampling method (12).The sampling unit in the first stage was the administrative unit of residence, i.e., the village in Haidian, and the residential committee in the other two areas. In the second stage, the sampling unit was the subject. The survey was conducted between August 1 and September 30, 1986. Information about occupational exposure to dusts and gases/fumes, exposure duration, and exposure intensity was obtained by trained interviewers using a modified Chinese translation of the ATS-DLD questionnaire (11, 13). Intensity of dust exposure was classified into three categories: low (a little but not too much), moderate (when finished work there was a lot of dust on the subject's clothes), and high (one cannot see the person just a few yards away). Intensity of gas/fume exposure was also classified into three categories: low (occasional exposure), moderate (frequent exposure but concentration not perceived to be high), and high (very frequent or daily exposure to high concentrations). A cumulative exposure index was created by grouping subjects into three exposure categories based on both exposure duration and exposure intensity. High cumulative exposure indicates that subjects had been exposed to dusts (or gases/fumes) over 10 yr with moderate or high intensity, and low cumulative exposure for less than 10 yr with low intensity. The others were included in the medium exposure group. This index was used in all exposure-response analyses. Respiratory symptoms were likewise determined using a Chinese translation of the standard questions (11, 13, 14).Chronic cough was defined as cough for 3 or more months of the year, chronic phlegm as sputum production for 3 or more months of the year, breathlessness as shortness of breath when walking with other people of one's own age on level ground, and persistent wheeze as wheezing on most days or nights. Chronic bronchitis was defined as the presence of both chronic cough and chronic phlegm. Pulmonary function measurements were performed in accordance with ATS performance and reproducibility criteria (15). Subjects performed vital capacity (VC) and FVC tests on the electric auto-spirometer (AS-300;
Minato Co., Japan) while in the standing position, with nose clip, either at the central station or at home. A minimum of three acceptable measurements was performed. The measurement with the highest FVC (the FEV. from the highest FVC) was used in the analysis. All of the pulmonary function values are reported at ambient temperature and pressure saturated with water (ATPS). The survey was conducted between August and September, during which the indoor temperatures were relativelystable, ranging from 22° C to 29° C. The pulmonary function parameters analyzed in this report were FEV., FVC, FEV./FVCOJo, FEF2 s - 7S , and FEF2 S- 7S/FVCOJo, and PEFR. FEV., and FVC measurements were standardized for height (HT) by dividing by HT 2 (11, 16). A previous report showed that the association between air pollution and FEV. for subjects who did not report using coal stove heating was stronger than for subjects who reported using coal stove heating, which suggests that the unexposed subjects may have been more sensitiveto airborne contaminants (11). In addition, the coal was burned in household stoves for heating with low thermal efficiency,no desulfurization, inefficient dust collection, and no or low smokestacks. It may not be appropriate to include the subjects who reported using coal stove heating in nonexposure group. Thus, the analyses of pulmonary function wererestricted to subjects who did not report using coal stove heating. Multiple logistic regressions were used to assess the association of occupational exposure with chronic respiratory symptoms. Because the symptoms of interest for each subject were correlated, the quasi-likelihood ideas of Liang and Zeger were used to model simultaneously these multiple binary correlated respiratory symptoms (17-19). The global estimate obtained from estimation equation represents the mean effect 0 f exposure on all symptoms of interest (18).In an exposure-response trend analysis, the values, 0, 1,2, and 3, were assigned to each level of exposure, respectively, to dusts and gases/fumes. The likelihood-ratio test statistic was used to test whether the assumption of the linear trend for symptoms was valid (20, 21). The effects of exposure levels based on assigned values will be reported, if the test results indicate that linearity of the trend with dusts or gases/ fumes is adequate. Multiple linear regressions were used to estimate the impact of occupational agents on pulmonary function. The regression coefficients were estimated by ordinary least-square (OLS) method. The variance was calculated by robust method, which can provide consistent estimator of OLS variance regardless of the presence of heteroskedasticity (22). The product terms of agents were used to assess the interactions between smoking and occupational exposures and between dusts and gases/fumes.
Results
There were 3,606(1,762 men, 1,844wom-
en) out of 3,746 subjects with complete information on respiratory symptoms, occupational exposure history, and covariates. The prevalence of occupational dust exposure was 32070 and gas/fume exposure 19070 for the overall population. Men had higher prevalence of both dust exposure and gas/fume exposure than women (table 1).Of the three areas studied, Shijinshan, the industrial area, had the highest exposures to both dusts and gases/fumes, although women in Dongchen had as much gas/fume exposure as those in Shijinshan. The prevalence of occupational exposure to either dusts or gases/fumes decreased with age except for men aged 40 to 59 yr. Exposure frequency also decreased with education, although subjects in the lowest education level had less gas/fume exposure than those in other groups in both men and women, and for dust exposure among women. Subjects who reported using central heating systems had significantly higher rates of occupational exposure than those who reported using coal stove heating, reflecting the predominant use of central heating systems in the industrialized area, Shijinshan, compared with the residential area, Dongchen. The most frequently reported specific occupational exposures (table 2) for dusts are coal (19.6070) and for gases/fumes were nonspecific gas (34.4070). Exposure to coal dust among these workers was not in mines.
Respiratory Symptoms The prevalence of respiratory symptoms was calculated by reports of occupational exposures (table 3). The relative odds of chronic respiratory symptoms were calculated for subjects with and without a history of dust exposure, after we adjusted for residential area, age, sex, cigarette smoking status, indoor coal stove for heating, and education level (table 3). Chronic phlegm and breathlessness were significantly associated with both dust and gas/fume exposure, while chronic cough only with dusts and wheeze only with gases/fumes. The global estimate of relative odds of the four symptoms was 1.30 (95070 CI 1.09 to 1.48) and 1.27 (95070 CI 1.09 to 1.48), respectively, for dusts and gases/fumes (table 3). Twenty-three percent of exposed subjects reported exposures to both dusts and gases/fumes. The interaction terms of the exposures with respiratory symptoms were tested using logistic models, but the results were not significant (p = 0.28 to 0.68). When both dust and
RELATIONSHIPS BETWEEN OCCUPATIONAL EXPOSURES AND CHRONIC RESPIRATORY ILLNESS
415
TABLE 1 PERCENTAGE DISTRIBUTION OF OCCUPATIONAL EXPOSURES TO DUSTS, GASES/FUMES BY SEX AND BY AREA, AGE, EDUCATION LEVEL, SMOKING STATUS AND HEATING TYPE, AMONG 3,606 ADULTS IN BEIJING, CHINA Male
Total Area Dongcheng Haidian Shijinshan Age, yr 40-49 50-59 60-69 Education Illiterate Primary sch. High sch. > High sch. Smoking Never Ex Current Heating Coal stove Other
Female
Subjects n (0/0)
Dust Exposure (%)
Gas/Fume Exposure (%)
Subjects n (0/0)
Dust Exposure (0/0)
Gas/Fume Exposure (%)
1,762
37.7
24.0
1,844
27.0
14.4
541 (31) 610 (35) 611 (35)
30.9 30.2 51.4
26.8 12.8 32.6
585 (32) 618 (34) 641 (35)
30.8 13.9 36.2
19.2 4.9 19.2
606 (34) 722 (41) 434 (25)
41.4 41.3 26.7
29.2 22.9 18.4
758 (41) 684 (37) 402 (22)
34.8 25.9 14.2
19.3 13.9 6.0
308 693 604 157
(18) (39) (34) (9)
41.2 38.7 35.9 33.8
15.6 26.0 26.8 20.4
881 515 386 62
(48) (28) (21) (3)
22.4 32.6 29.5 30.7
10.2 17.7 19.2 16.1
392 (22) 209 (12) 1,161 (66)
34.4 41.2 38.2
25.5 23.0 23.6
1,184 (64) 156 (8) 504 (27)
26.5 26.9 28.2
14.9 15.4 12.9
31.7 49.8
20.2 31.4
1,225 (66) 619 (34)
23.0 34.9
13.2 16.6
I"
1,173 (67) 589 (33)
TABLE 2 FREQUENCY OF SPECIFIC OCCUPATIONAL EXPOSURES TO DUSTS AND GASES/FUMES Dusts
Gases/Fumes
Type
n
Coal Lime Metal Cement Cotton Wood, paper, grass Metallurgical Silica Others
228 160 142 128 96 74 59 37 239
19.6 13.8 12.2 11.0 8.3 6.4 5.1 3.2 20.6
1,163
100.0
Total
Type
%
n
%
Gas· Solvent Gasoline, kerosene Acid gas Acetylene, ethylene, methane Pesticide Others
236 102 102 63
34.4 14.9 14.9 9.2
36 30 118
5.2 4.4 17.2
Total
687
100.0
* Nonspecific gas.
TABLE 3 PREVALENCE (0/0) OF CHRONIC RESPIRATORY SYMPTOMS AND ESTIMATED ODDS RATIOS· OF EXPOSURES TO DUSTS AND GASES/FUMES AMONG 3,606 ADULTS Occupational Exposure No
Yes
Odds Ratio
95% Confidence Interval
Dusts Cough Phlegm Breathlessness Wheeze Global
16.1t 22.6 20.9 6.7
20.6 28.0 25.4 7.0
1.32 1.27 1.39 1.02 1.30
1.09-1.59 1.07-1.50 1.16-1.65 0.77-1.36 1.14-1.48
Gases/fumes Cough Phlegm Breathlessness Wheeze Global
17.0 23.3 21.7 6.2
19.9 28.5 25.2 9.2
1.17 1.21 1.31 .; 1.62 1.27
0.94-1.46 1.00-1.48 1.06-1.61 1.18-2.21 1.09-1.48
Symptoms
