INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Various substances of biologic or synthetic chemical origin have been associated with the inception or the aggravation of asthma among individuals who are exposed in occupational settings (1). Although several disease registries and surveillance systems of occupational asthma have been implemented in various countries (2), the epidemiologic evidence with regard to the occurrence of occupational asthma in the community is underdeveloped (5). A limited number of studies have estimated the burden of occupational asthma in the adult general population (6). Estimates of the population attributable risk (PAR) due to occupational exposure range from 2% to 30% and are dependent on the prevalence of exposure and the case definition employed. In the United States, few population studies are available on the association between occupational exposures and asthma or asthma-like symptoms (6, 12); and no study specifically has evaluated the occurrence of occupational asthma among women.

In a community study of elderly women in Northern California, individual information on respiratory health, smoking habits, and occupation had been collected as a part of an ongoing longitudinal project of the effects of aging on physical functioning (13). We used data from the baseline evaluation to evaluate the role of smoking and of occupational factors on the prevalence of self-reported asthma, chronic bronchitis, and asthma-like respiratory symptoms and to estimate the asthma-related risks attributable to smoking and to occupation in this population.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The data presented were obtained as part of the baseline evaluation of subjects who enrolled in the Study of Physical Performance and Age-Related Changes in Sonomans (SPPARCS). SPPARCS is a community-based, longitudinal study of physical activity and fitness in people  55 yr of age who live in the city and environs of Sonoma, California (13). The dominant industries in the community are farming (wine grapes, dairy cattle) and tourism; there is no heavy industry. A community-based census identified 3,057 age-eligible individuals of whom 2,092 (68.4%) agreed to participate in the study and were enrolled between May 1993 and December 1994. Of these, there were 1,246 females who are the subjects for this report. The age distribution of these subjects was similar to that of the 1990 census data for persons  55 who resided in the study community (13). Protocols were approved by the Committee for the Protection of Human Subjects at the University of California, Berkeley and the Committee on Human Research at the University of California, San Francisco.

As part of the initial evaluation, an interviewer-administered questionnaire was completed that contained detailed information on occupational history, respiratory symptoms and illnesses, cigarette smoking history, general medical conditions, and functional capacity. The occupational history included: (1) job title of current or last job; (2) job title during most of subject's working life; (3) specific work performed and duration of employment for #1 and #2; and (4) reasons for retirement and/or status as "currently unemployed." Respiratory symptom and illness questions were those of the ATS/Division of Lung Diseases (DLD), National Heart, Lung, and Blood Institute questionnaire (14), which were supplemented by detailed questions on asthma history (e.g., age of onset, duration, medications, hospitalization, precipitating factors) and wheezing. Complete questionnaire data were available for 1,226 (98.4%) of the women who enrolled in the study (difference due to subjects who could not complete the questionnaire or for whom a proxy interview was obtained).

Subjects performed maximal expiratory maneuvers in the study laboratory in accordance with American Thoracic Society criteria (15) in the seated position with nose clip on a dry, rolling seal spirometer linked to a computer (Plus System; Warren E. Collins, Braintree, MA). Calibration was performed daily with a 3-L syringe. Data were analyzed for subjects with at least two acceptable time-volume tracings for which forced expiratory volumes in one second (FEV₁) were within 10%. Acceptable data were available for 820 (66.9%) women.

Outcome Variables

On the basis of the respiratory questionnaire, three mutually exclusive outcome categories were defined as follows:

1. Adult asthma: report of a physician diagnosis of asthma (at age 18 or more) and wheezing symptoms during the past 12 mo (65 subjects).
2. Chronic bronchitis/emphysema: report of a physician diagnosis of chronic bronchitis/emphysema during the past 12 mo (90 subjects).
3. Asthma-like symptoms: report of shortness of breath with wheeze or chronic wheeze during the past 12 mo in the absence of a physician diagnosis of asthma or chronic bronchitis/emphysema (175 subjects).

All women (68 subjects) with a history of inactive respiratory disorders (report of a physician diagnosis but no symptoms in the past 12 mo) and 14 subjects with childhood asthma (report of a physician diagnosis of asthma before the age of 18) were not included in the above classification. The reference group was composed of 814 women who reported that they never had any of the above respiratory conditions.

Exposure Variables

Smoking. Smoking habit was categorized into four classes: never smokers, quitters of 10 yr or more, quitters in the previous 10 yr, and current smokers. This classification was based on an analysis of the distribution of lifetime cigarette smoking and forced expiratory volumes, which demonstrated that subjects who quit within 10 yr of interview had FEV₁ that were more like those of subjects who were current smokers than those of other former smokers (data not shown).

Occupational history. The last and the longest occupation were initially coded with codes used by the 1990 U.S. Census (16). Subsequently, the codes for the longest occupation were grouped into 13 major occupational groups. The last and longest occupations were identical for 81.5% of subjects (934 of 1,144). The grouping was done without regard to the health status of the women. We tried to follow the classification scheme of the SWORD projects in the United Kingdom (17) which has been used in two studies on the risk of asthma due to occupation (8, 11). However, our categorization had to consider differences between the U.S. Census and the British codes, as well as the fact that, for several categories of the SWORD classification (e.g., woodworkers, welders, painters, construction workers), there were few or no subjects in our population, given the predominance of males in such occupations. We combined women in executive and managerial occupations (codes: 1-89 in the U.S. Census rubric) and those in administrative occupations (codes: 303-309) into a single category to serve as referents in the evaluation of the association between different occupations and respiratory conditions.

All job descriptions in the individual occupational histories were reviewed blindly with regard to health status by two of us with experience in occupational medicine (F.F., M.S.). Each subject was classified as potentially exposed or unexposed to dust, gas, vapors, fumes, or sensitizers (DGVFS) in the workplace (first pass by F.F.). The assignment to the exposed or unexposed category was based on the existing knowledge about the factors associated with occupational asthma (1, 18), as well as on the prevalence data from the National Occupational Exposure Survey (19). Based on the results of the first pass, a standardized assignment protocol was established. A second, complete, blind review was undertaken by F.F.; and the assignments were made with the standardized assignment protocol. Finally, a third, blind, independent review of the coding was undertaken by M.S. Coding disagreements between M.S. and F.F. were then reconciled by mutual agreement. Duration of exposure to dusts, gas, vapors/fumes, or sensitizers was also recorded.

Data Analysis

Multinomial logistic regression (logistic regression that permits the outcome to take on more than two levels; mlogit of STATA [20]) was used to calculate odds ratios (OR) and 95% confidence intervals (CI) for the association between adult asthma, chronic bronchitis, and asthma-like symptoms with smoking and occupation. In the logistic model, the reference group was the 814 females without any of the respiratory conditions and symptoms of interest. In the evaluation of the association between the three outcome variables and the occupational groups (reference group = executives, managers, administrators), adjustment was made for age and smoking (reference group = never smoked). PAR of asthma and asthma-like symptoms due to smoking and to occupation were estimated (21).

Adjustment of the lung function measures was accomplished by linear regression of the specific lung function measure on mean-centered age, height, dummy variable for smoking, and dummy variables for the respiratory conditions.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Table 1 shows the selected characteristics of the women in the population studied. Most women were white with a relatively high educational level (62% had 13 or more years of education). The mean age (± SD) was 70.3 ± 8.7. Only 2.2% never worked; 70% were retired. A total of 49% were never-smokers, and 9% were current smokers. Subjects with any of the chronic respiratory conditions (39 asthmatics, 59 subjects with chronic bronchitis, and 115 individuals with asthma-like symptoms for whom a lung function test was available) had significantly lower FVC, FEV₁, FEF_25-75, and FEF₇₅ values than women in the reference category (557 subjects) after adjustment for age, height, and smoking habit in multiple linear regression (Table 2). Women with adult asthma had the lowest lung function values.

OR and 95% CI for the association between smoking and respiratory conditions are presented in Table 3. Former smokers had an increased odds of self-reported, doctor-diagnosed asthma, chronic bronchitis, and self-reported asthma-like symptoms; OR ranged from 1.4 to 4.7. There was a strong and significant association between current smoking and both chronic bronchitis (OR = 6.4) and asthma-like symptoms (OR = 3.8), whereas the association between current smoking and asthma was weaker and nonsignificant (OR = 1.6).

Several statistically significant associations were observed between the respiratory conditions and occupational groupings (Table 4). Artists, writers, decorators, and photographers (OR = 3.1) as well as the group of women in service occupations (OR = 2.4) had a significantly increased risk of asthma. There were nonsignificant elevated risks for asthma for the following occupational groups: technologists and technicians (OR = 2.5, CI 0.79 to 9.1), food preparation (OR = 3.5, CI 0.7 to 17.6), precision production (OR = 3.4, CI 0.7 to 17.1) and homemakers (OR = 2.4, CI 0.5 to 11.5). The odds of chronic bronchitis was significantly elevated (OR = 5.2) in the group of machine operators, assemblers, and inspectors. Asthma-like symptoms were significantly more prevalent among nurses and other nonphysician health care workers (~ 80% nurses, OR = 2.9), social workers (OR = 2.9), and housewives (OR = 2.4) than in the referent category. The data were reanalyzed with "asthma" and "asthma-like" symptoms combined into a single group. The results for nurses, social workers, and homemakers continued to show statistically significant OR (2.3, 3.2, and 2.6, respectively). The results for artists and photographers (1.6, 95% CI = 0.7 to 3.3), food preparation workers (1.7, 95% CI = 0.6 to 5.1), and "other service occupations" (1.4, 95% CI = 0.8 to 2.6) showed OR 1.4 or greater, but none was statistically significant.

The overall proportion of the female population previously exposed to DGVFS was 21.2%. The proportion of exposed subjects was highest in the following occupations: nurses, 92%; food services, 77%; precision production, 58%; other services, 54%; machine operators, 50%; and technicians, 46%. Exposure to DGVFS was associated with a significant increased odds of asthma (OR = 1.8), and with a marginally significant increased odds of asthma-like symptoms (OR = 1.4, CI 0.9 to 2.0) (Table 5). The odds of asthma due to occupation increased with increasing duration of exposure both when duration of exposure was treated as a categorical variable (Table 5) and when it was considered as a continuous variable (asthma: age-smoking-adjusted OR = 1.26, CI = 1.02 to 1.56 per 10 yr duration of exposure). A similar, but less strong trend was observed for asthma-like symptoms (Table 5; OR = 1.13, CI = 0.96 to 1.3, per 10 yr duration of exposure). When subjects were stratified by smoking history, an increased odds of asthma with exposure to DGVFS was observed only in lifelong nonsmokers (OR 5.2) (Table 5).

Given the observed prevalence of ever-smoking in this population (51.1%) and the observed overall odds for asthma (OR = 2.3) and asthma-like symptoms (OR = 2.1) in subjects who ever smoked cigarettes, the estimated PAR for smoking were 40.5% and 35.0%, respectively (Table 6). PAR due to occupational exposure to dusts, gas, vapors, fumes, or sensitizers were 15.1% and 7.5% for asthma and asthma-like symptoms, respectively. The estimates were slightly higher (20.0% and 10.2%, respectively) when all the occupational groups that were associated with respiratory conditions (Table 4: nurses, social workers, artists, technicians, food and other services occupations, precision production and machine operators) were merged to form an a posteriori "high-risk" group (reference group = all other occupations). The risk of asthma and asthma-like symptoms was increased among the few subjects (4.0%) who were still working and whose current occupation was classified as "high-risk" (asthma: OR = 2.7, CI = 0.98 to 7.6; OR = 1.6, CI = 0.7 to 3.6 for asthma-like symptoms), but the estimates of the attributable risk (6.3% and 2.4%, respectively) had wide confidence intervals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

No previous study has tried to quantify the burden of occupational factors on the distribution of respiratory conditions among elderly women. Because asthma and chronic respiratory symptoms are relatively frequent in older age and are an important cause of disability and health care costs, we considered it of interest to evaluate the contribution of occupational factors for these conditions. Permanent impairment and disability have been described among patients with occupational asthma, due to a persistence of the airways inflammatory process even in absence of continued exposure (1).

This study shows that employment in a number of occupations is associated with an increased odds of self-reported, doctor-diagnosed asthma and self-reported asthma symptoms among older women. Moreover, employment in occupations with a high probability of exposures to dusts, gas, vapors, fumes, or sensitizers was a significant contributor to the burden of asthma PAR = 15 to 20%).

The increased risk of asthma in the category of artists, writers, decorators, and photographers is interesting, since few reports are available on this group with a high potential for exposure to several chemicals that may cause asthma. Artists and decorators may be exposed to low-molecular-weight agents, including diisocyanates, anhydrides, dyes, resins, glues, and metals. Occupational asthma has been described among photographers exposed to ethylenediamine (22) and to sodium metabisulfite (23).

Women in "other service occupations" in this study were at an increased risk of asthma. Most of the subjects in this group were caretakers or cleaners, occupations in which the use of detergents and solvents that have been associated with asthma is common (1). Cleaners also were found to be at high risk of asthma in a recent population study in Spain (8). Homemakers may share the same type of exposure to detergents and solvents; and, therefore, it is not surprising that they have been found to be at increased risk for both asthma and respiratory symptoms in this study and in the study reported from Spain (8).

We found a significantly increased risk of asthma-like symptoms among nurses and other nonphysician health workers, approximately 80% of whom were nurses, with the others falling into the categories of pharmacists, occupational, physical and speech therapists, and dietitians. These results are similar to other reports that have documented occupational asthma among nurses (17) and respiratory therapists (7). Workers in these occupations may be exposed to several biologic agents and to chemicals, (e.g., formaldehyde and glutaraldehyde). Allergy caused by exposure to powdered latex gloves (24, 25) also may contribute to the observed increased odds of symptoms. Although the results for nurses and other health workers are easily interpretable and have a potentially large impact for primary prevention, we have no good explanation for the excess of asthma-like symptoms among social workers.

The other occupations, for which we observed nonsignificant but highly suggestive associations with asthma (technicians, food workers, and farmers) have been associated with asthma in other studies (8, 11). Other investigations have indicated a strikingly elevated risk of asthma among farmers. In a study among males and females in New Zealand, an odds ratio of 4.3 was observed for the association between farming and asthma (11). Our results for farmers, although suggestive of an increased risk, are less striking than those from New Zealand. Dairy farming and grape growing constitute the principal agriculture of the study community and may represent a different set of exposures than those observed in the New Zealand study. Moreover, women may have less intense exposures than men who live and work in agricultural settings.

The significant association that we observed between exposure to gas, fumes, dust, vapors, and sensitizers and asthma parallels findings on asthma or wheeze in other population studies that include both men and women (9, 10, 26). In our study, subjects with the longest duration of exposure to these agents had the highest risk. Xu and Christini (10) found an association between asthma and cumulative exposure but not with duration of exposure. Flodin and colleagues found a significant relation with self-reported probability of exposure (26). In contrast to most published studies that have used self-reports of exposure to DGVFS (10, 12, 27), we assigned exposure to these agents after a blinded review (with regard to health status) of the individual occupational history, as has been done previously (28, 29) in two studies of asthma. Although not widely used in respiratory epidemiology, this approach has been applied extensively in epidemiologic studies of occupational cancer (30). Validity studies tend to suggest that self-report of exposure is probably an inadequate tool for the collection of information on occupational exposure in community studies, and that a detailed job description together with expert assessment would be preferred (31).

Recently, investigators have speculated that occupational asthma is more common in women than had been previously believed (8). Several European studies have identified high risks for respiratory symptoms among women exposed in a variety of occupational settings (17, 27, 32). A higher PAR was identified in females in comparison to males in a recent population study in Spain (8). Viegi and coworkers speculated that adult women are more susceptible than men to chemical substances that can irritate airway receptors (27). Our PAR estimate for asthma (15 to 20%) and asthma-like symptoms (7.5 to 10%) may appear surprisingly high for a population of women who live in a nonindustrialized area. While the use of an a posteriori classification may have led to a somewhat increased estimate, nonetheless, the estimates from this study are consistent with previous work in the U.S. Based on data from the 1978 Survey of Disability and Work, Blanc estimated that 15% of asthma among adults could be related to occupational exposures (6). Similar estimates have been provided in a study from Michigan (10 to 18%) (33) and in another study from the San Francisco Bay Area (6 to 14%) (29).

We found that cigarette smoking clearly is associated with asthma, chronic bronchitis, and asthma-like symptoms. Current smokers, however, were at increased risk of chronic bronchitis and wheeze but not of asthma. A recent report from the Nurses' Health Study in the U.S. (11) has shown an increased risk among former smokers but not among current smokers. Sex differences in the role of smoking in the occurrence of asthma and chronic wheeze has been controversial because of problems of diagnostic bias. In our data, females were more likely to report a doctor's diagnosis of asthma or chronic bronchitis than males; and males tended to have more asthma-like symptoms without a specific doctor's diagnosis. Unfortunately, we have no data that would permit us to determine whether a physician's knowledge of a woman's occupation influenced the diagnosis of asthma, chronic bronchitis, or emphysema.

Our study did not have the power to study properly how smoking and occupational exposure may interact in relation to the occurrence of asthma and chronic asthma-like symptoms. However, we found that the increased odds of asthma due to occupational exposure were mainly present among nonsmokers. At least two studies (12, 27) have addressed the issue of the combined effect of smoking and occupational exposure on chronic respiratory symptoms. In data from the Six Cities Study, Korn and associates (12) found an independent effect of current occupation and smoking on symptom prevalence. On the other hand, Viegi and coworkers (27) reported that the effect of occupational exposure on several respiratory categories (including wheeze and asthma) was higher among nonsmokers than among smokers.

There is no universally accepted definition of asthma for use in epidemiologic studies. However, in the absence of a measure of bronchial reactivity to supplement a report of wheezing (34), a self-report of a doctor's diagnosis of asthma is used most frequently to classify subjects as asthmatic. We followed this convention. Obviously, we cannot be certain that some subjects in the group with "asthma-like" symptoms, in fact, have asthma. We believe that our choice is supported by the pulmonary function data that are presented in Table 2, which show that, for all measures, subjects with a doctor's diagnosis of asthma have substantially lower levels than those with asthma-like symptoms. These findings are consistent with other epidemiologic studies (34). Because "asthma" and "asthma-like symptoms" do not necessarily measure exactly the same pathophysiologic entities, we separated them in our classification. We adopted a "mutually exclusive" classification because we wanted to isolate asthma, at least as diagnosed by physicians and reported by subjects, from "chronic bronchitis/emphysema," which is more likely to be related to smoking. The data in Table 3 indicate that those with "asthma" and "asthma-like symptoms" were significantly less likely either to be current smokers or to have quit smoking in the past 10 yr than subjects who were labeled as having chronic obstructive pulmonary disease, which would help to justify our a priori decision about the classification.

A limitation of this cross-sectional study is that the exposure and the outcome depended on answers to questionnaires. In addition, we did not have information on all the of the jobs that might have been held by the subjects, nor did we have information that would have permitted us to determine the relationship between the onset of adult asthma and the duration of employment in a particular occupation at the time of onset. Only the last job and the longest occupation in which the subject was employed were ascertained. These limitations were imposed by the length of the questionnaire as it related to the main focus of the study. Therefore, misclassification of exposure could be an issue, since we cannot rule out the possibility that diseased subjects may have recalled their occupation better than healthy individuals. However, the study was focused on physical function and performance rather than on occupation, making it unlikely that recall bias occurred with respect to the relationship between occupation and chronic respiratory diagnoses and/or symptoms. Finally, the lack of data on time of onset in relationship to employment leaves open the possibility that the magnitude of the reported association could be biased upward. However, the consistency of our findings with others suggests that any such bias is likely to be small.

In conclusion, these data indicate that occupational exposures may be an important contributor to the occurrence and/ or persistence of asthma in older women. Prospective population-based studies of younger adults should include detailed occupational histories to clarify the validity and strength of the relationships reported herein.