Prevalence of Chronic Obstructive Pulmonary Disease in the U.S. Working Population: An Analysis of Data from the 1997–2004 National Health Interview Survey

Abstract

To estimate the prevalence and the population attributable fraction of chronic obstructive pulmonary disease (COPD) in the U.S. adult workers, we analyzed data obtained from the National Health Interview Surveys for the period 1997–2004. The overall COPD prevalence was 4.0% (95% confidence interval [CI] 3.9–4.1%). The prevalence was higher in females (5.4%, 95% CI 5.3–5.6%) than in males (2.8%, 95% CI 2.7–2.9%); in Whites (4.2%, 95% CI 4.1–4.3%) than in Blacks (3.4%, 95% CI 3.1–3.7%) and other races (2.4%, 95% CI 2.1–2.8%). Compared with insurance, real estate and other finance industry, the top three industries associated with significantly higher prevalence odds ratios (PORs) (adjusted for age, sex, race, and smoking) were other educational services (POR = 1.5, 95% CI 1.0–2.3); transportation equipment (POR = 1.4, 95% CI 1.1–1.8); and social services, religious and membership organizations (POR = 1.4, 95% CI 1.1–1.7). Compared with managers and administrators, except public administration occupation, the top three occupations with significantly higher PORs were health service (1.8, 95% CI 1.5–2.1), other protective service (POR = 1.6, 95% CI 1.2–2.2), and material moving equipment operators (POR = 1.6, 95% CI 1.1–2.3). The overall population attributable fraction for association of COPD with employment was 12.2% for industry and 17.4% for occupation. Further studies are needed to determine specific risk factors associated with COPD in industries and occupations with elevated prevalence and POR.

Keywords:

• Chronic obstructive pulmonary disease
• Prevalence
• Occupation
• Industry

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is a lung disease that includes chronic bronchitis or emphysema with airflow obstruction ([1], [2]). COPD is a leading cause of chronic respiratory disease in the U.S. and a target for public health prevention ([3]). COPD was the fourth leading cause of death in the U.S. in 2004 ([4]) and is projected to be the third leading cause of death by the year 2020 ([5]). In 2001, an estimated 12.1 million adults ages 25 and older were diagnosed with COPD in the U.S ([5]), while in 2002, the total estimated cost of COPD was $32.1 billion ([5]).

Halbert and colleagues identified several studies with nationally representative estimates of the COPD prevalence in various countries or regions and found that the prevalence ranged from 3.7% in Denmark and Sweden to 11% in Italy ([3]). The reported COPD prevalence in the U.S. was 5.4% for all ages and 6.8% for those aged 17 years and older ([3]).

Cigarette smoking is the primary cause of COPD and accounts for 80 to 85% of all cases ([6]). Associations of COPD with certain occupations and industries have been previously established ([7],8,9). Several industry and community-based epidemiologic studies have reported association of COPD with various occupational exposures including dusts, gases, minerals (coal, oil mist, silica), metals (vanadium, osmium), grain, wood, cotton, vapors, and welding fumes ([10],11,12,13,14,15,16,17,18,19,20,21).

The National Health Interview Survey (NHIS) is a survey of the U.S. civilian non-institutionalized population on multiple health issues and has been undertaken since 1957 ([22]). Approximately 40,000 adults participate in the NHIS annually. The survey uses a multiple probability sample design involving stratification, clustering, and oversampling of certain populations. In this study, we used 1997–2004 NHIS data to identify industries and occupations associated with increased prevalence for COPD and to determine the population attributable fraction (PAF) of COPD associated with employment by industry and occupation.

METHODS

We combined the NHIS data for all adult (aged ≥ 18 years) current-employed workers. Current employment was defined on the basis of a question asking about employment during the previous week. Workers who responded “yes” to “working for pay at a job or business,” or “with a job or business but not at work,” or “working, but not for pay, at a family-owned job or business.” were included in the study ([23]). Individuals with COPD were identified as those who responded “yes” to either of the following questions “During the past 12 months, have you been told by a doctor or other health professional that you had chronic bronchitis?” or “Have you ever been told by a doctor or other health professional that you had emphysema?” Cigarette smoking status was classified into three groups: current smokers were those who had smoked at least 100 cigarettes during their entire life and who currently smoked; former smokers were those who had smoked at least 100 cigarettes in their entire life but did not currently smoke; never smokers were those who never smoked or those who had smoked less than 100 cigarettes during their entire life.

Industry and occupation

Data on current industry and occupation were collected from all adult workers who were employed during the week prior to the interview. Due to confidentiality reasons, the National Center for Health Statistics (NCHS) used industry/occupation 2-digit codes and created 42 industry and 41 occupation groupings for public use data files. These recodes were based on industry and occupation groups and subgroups consistent with the 1987 Standard Industrial Classification (SIC) and Standard Occupational Classification (SOC) [http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.htm]. For both industry and occupation groupings, those who had missing information (i.e., refused to answer, were not ascertained, or did not know their industry/occupation) were excluded from the analysis.

Data analysis

We estimated the overall population COPD prevalence for the entire period 1997 to 2004, by demographic characteristics, smoking status, and current industry and occupation. We used SAS® 9.1 (SAS Institute Inc., Cary, NC) software to analyze data and to calculate the estimated frequencies, variances, and prevalence rates with corresponding 95% confidence intervals (CIs). Sample weights provided by NCHS were used to take into account the complex sampling design and non-response. Differences of COPD prevalence by gender, race, and smoking status were evaluated using the F-test statistic. We used multivariate logistic regression to estimate the COPD prevalence odds ratios (PORs) with corresponding 95% CIs, by industry and occupation, adjusting for age, gender, race, and smoking status. The reference industry and occupation used for POR calculations were insurance, real estate and other finance and managers and administrators, except public administration, respectively. These groups were chosen because of reported minimal risk for COPD in these industries and occupations ([9]).

The PAF for COPD was defined as the fraction of cases with COPD associated with industrial (occupational) exposure and was calculated using the following formula ([9], [18]):where k = number of industries (occupations); P_i = cases_i/total cases, where cases_i represents the estimated number of COPD cases in the ith industry (occupation) and total cases represents the total number of cases summed across industries (occupations). For calculating PAF, we included the industries (occupations) with POR greater than 1. Attributable cases (AC) are cases attributable to employment in the ith industry or occupation, and were calculated using the following formula: AC = N_i*P_i (POR_i – 1)/POR_i, where N_i = estimated respondents in the ith industry or occupation; P_i = estimated prevalence in the ith industry or occupation.

RESULTS

The estimated overall COPD prevalence was 4.0% (95% CI 3.9–4.1%) (Table 1); prevalence was significantly higher in females (5.4%, 95% CI 5.3–5.6%, P < 0.005) than males (2.8%, 95% CI 2.7–2.9%); in whites (4.2%, 95% CI 4.1–4.3, P < 0.005) than blacks (3.4%, 95% CI 3.1–3.7) and others (2.4%, 95% CI 2.1–2.8); and in current smokers (6.6%, 95% CI 6.3–6.9, P < 0.005) than former smokers (4.4%, 95% CI 4.1–4.6) and nonsmokers (2.8%, 95% CI 2.7–2.9). Age-specific COPD prevalence increased with age, from 3.5% (95% CI 3.3–3.6%) in the age group 18–44 years to 6.8% (95% CI 5.2–8.5%) in the age group ≥ 75 years.

Table 1 Estimated number of the U.S. working adults (≥ 18 years) and prevalence of COPD by selected characteristic, 1997–2004

			COPD Prevalence % (95% CI)
Characteristics	Workers^†	Cases^†	Overall	Current smokers	Former smokers	Never smokers
Age group
(years)
18–44	81,337	2,822	3.5 (3.3–3.6)	5.7 (5.4–6.1)	3.7 (3.3–4.1)	2.4 (2.3–2.6)
45–64	42,318	2,039	4.8 (4.6–5.0)	8.1 (7.6–8.7)	4.5 (4.1–4.8)	3.5 (3.3–3.8)
65–74	3,221	222	6.9 (6.1– 7.7)	15.5 (12.1–18.8)	7.6 (6.3–8.9)	4.0 (3.0–4.9)
≥ 75	748	51	6.8 (5.2–8.5)	8.0 (0.9–15.1)	9.8 (6.9–12.8)	4.0 (2.2–5.8)
Gender
Male	68,534	1,917	2.8 (2.7–2.9)	4.3 (4.0–4.6)	3.4 (3.1–3.7)	1.8 (1.6–1.9)
Female	59,090	3,217	5.4 (5.3–5.6)	9.6 (9.1–10.1)	5.9 (5.4–6.3)	3.8 (3.6–4.0)
Race
White	106,016	4,475	4.2 (4.1–4.3)	6.9 (6.6–7.2)	4.5 (4.2–4.7)	2.9 (2.7–3.0)
Black	13,903	471	3.4 (3.1–3.7)	5.3 (4.6–6.1)	3.8 (3.0–4.6)	2.7 (2.4–3.0)
Other	7,705	188	2.4 (2.1–2.8)	3.6 (2.8–4.5)	3.3 (2.3–4.3)	1.9 (1.5–2.3)
Total	127,624	5,134	4.0 (3.9–4.1)	6.6 (6.3–6.9)	4.4 (4.1–4.6)	2.8 (2.7–2.9)

†Estimated number in 1,000s.

*p < 0.05.

COPD by industry

Industries associated with significantly higher risk for COPD were other educational services (POR = 1.52, 95% CI 1.01–2.28); transportation equipment (POR = 1.37, 95% CI 1.07–1.76); social services, religious and membership organizations (POR = 1.37, 95% CI 1.13–1.65); general merchandise stores (POR = 1.32, 95% CI 1.04–1.68); and business services (POR = 1.28, 95% CI 1.07–1.54) (Table 2).

Table 2 Estimated prevalence of COPD, prevalence odds ratio (POR), number of attributable cases (AC), and the population attributable fraction (PAF) by industry—U.S. working adults (≥18 years), 1997–2004

		COPD prevalence
				Never smokers
Industry	Workers^†	% (95% CI)	POR (95% CI)	% (95% CI)	AC^†	PAF%
Other educational services	564	5.9 (3.7–8.1)	1.52 (1.01–2.28)	3.9 (2.0–5.8)	11	0.22
Mining	493	4.2 (2.1–6.3)	1.39 (0.81–2.39)	1.5 (0.0–3.3)	6	0.11
Transportation equipment	2,265	4.4 (3.5–5.2)	1.37 (1.07–1.76)	2.5 (1.6–3.3)	27	0.52
Social services, religious and membership organizations	4,792	5.6 (4.9–6.2)	1.37 (1.13–1.65)	4.6 (3.8–5.4)	72	1.40
General merchandise stores	2,288	5.1 (4.2–6.0)	1.32 (1.04–1.68)	4.4 (3.2–5.6)	28	0.55
Repair services	1,931	4.0 (3.0–4.9)	1.29 (0.97–1.73)	2.9 (1.5–4.3)	17	0.34
Business services	7,129	4.4 (3.9–4.8)	1.28 (1.07–1.54)	2.6 (2.1–3.1)	68	1.32
Primary metal industry	746	3.9 (2.4–5.4)	1.24 (0.80–1.92)	0.9 (0.0–1.8)	6	0.11
Machinery, except electrical	2,032	3.9 (3.0–4.3)	1.23 (0.94–1.61)	3.2 (2.1–4.3)	15	0.29
Utilities and sanitary	1,411	3.7 (2.7–4.7)	1.23 (0.90–1.68)	2.9 (1.6–4.3)	10	0.19
Other and not specified durable goods	2,095	4.1 (3.2–5.0)	1.23 (0.94–1.60)	2.1 (1.2–3.0)	16	0.31
Eating and drinking places	5,851	4.5 (3.9–5.1)	1.22 (1.00–1.48)	2.4 (1.8–3.1)	47	0.92
Other transportation	2,920	4.2 (3.5–4.9)	1.22 (0.97–1.53)	2.8 (2.0–3.6)	22	0.43
Textile mill and finished textile products	1,189	4.6 (3.4–5.8)	1.21 (0.89–1.64)	2.2 (1.2–3.1)	9	0.18
Health services, except hospitals	7,125	5.1 (4.6–5.6)	1.20 (1.00–1.43)	3.2 (2.6–3.7)	60	1.17
Other personal services	3,210	4.9 (4.1–5.6)	1.19 (0.96–1.47)	3.1 (2.3–3.9)	25	0.49
Food bakery and dairy stores	2,843	4.5 (3.7–5.4)	1.19 (0.94–1.50)	2.9 (2.1–3.8)	20	0.40
Forestry and fisheries	163	3.7 (0.0–7.3)	1.18 (0.40–3.46)	2.8 (0.0–8.2)	1	0.02
Other and not specified retail trade	6,384	4.2 (3.7–4.8)	1.16 (0.96–1.40)	3.2 (2.6–3.9)	36	0.71
Elementary and secondary schools and colleges	10,763	4.3 (3.9–4.7)	1.13 (0.96–1.34)	3.4 (3.0–3.8)	54	1.04
Entertainment and recreation services	2,330	3.7 (2.8–4.6)	1.13 (0.84–1.51)	2.5 (1.4–3.6)	10	0.19
Private households	859	5.2 (3.6–6.7)	1.12 (0.80–1.58)	2.3 (1.2–3.4)	5	0.09
Public administration	6,194	3.9 (3.5–4.4)	1.10 (0.92–1.33)	2.9 (2.4–3.4)	23	0.44
Furniture lumber and wood	1,445	3.6 (2.5–4.7)	1.10 (0.76–1.58)	3.6 (1.8–5.4)	5	0.09
Electrical machinery, equipment and supplies	1,619	3.6 (2.6–4.5)	1.08 (0.79–1.48)	1.8 (1.0–2.7)	4	0.09
Banking and credit agencies	2,832	4.0 (3.2–4.8)	1.06 (0.84–1.35)	3.4 (2.6–4.3)	7	0.13
Other non durable goods	1,531	3.6 (2.6–4.6)	1.06 (0.77–1.47)	2.0 (0.8–3.1)	3	0.06
Trucking service and warehousing	2,496	3.3 (2.6–4.0)	1.06 (0.81–1.39)	2.0 (1.1–2.8)	5	0.09
Wholesale trade	4,200	3.5 (2.9–4.0)	1.06 (0.85–1.32)	2.7 (2.0–3.4)	8	0.15
Printing publishing and allied industries	1,623	3.9 (2.9–4.8)	1.06 (0.79–1.42)	3.1 (1.9–4.3)	3	0.07
Chemicals and allied products	1,206	3.4 (2.3–4.5)	1.05 (0.73–1.51)	2.2 (1.0–3.4)	2	0.04
Communications	1,925	3.4 (2.6–4.2)	1.03 (0.78–1.36)	2.0 (1.3–2.8)	2	0.04
Automotive dealers and gasoline stations	2,004	3.4 (2.6–4.2)	1.01 (0.75–1.35)	2.3 (1.3–3.3)	0	0.01
Other industries	5,570	3.1 (2.9–3.4)	0.96 (0.81–1.10)	2.2 (1.9–2.4)
Insurance, real estate and other finance	5,598	3.8 (3.3–4.3)	1.00	2.6 (2.1–3.2)
Total	127,624	4.0 (3.9–4.1)		2.8 (2.7–2.9)	627	12.20

Prevalence odds ratio (POR): adjusted for age, sex, race, and smoking status.

PAF%: percentage of COPD cases attributable to employment.

† Estimated number in 1,000s.

*Reference group.

Industries associated with POR < 1 included fabricated metal industries including ordnance; hospitals; agriculture; legal, engineering and other professional services; railroads; construction; and armed forces (excludes reserves and national guard) (not shown). Among never smokers, industries with highest COPD prevalence included social services, religious and membership organizations (4.6%), general merchandise stores (4.4%), and other educational services (3.9%).

COPD by occupation

Of all occupations, health service was associated with highest POR for COPD (POR = 1.78, 95% CI 1.48–2.14) (Table 3). Occupations with POR < 1 included technologists and technicians except health; police and firefighters; other transportation, except motor vehicles; forestry and fishing; health assessment and treating occupations; farm operators and managers; architects and surveyors; military; computer equipment operators; health diagnosing; and construction laborers (not shown). Among never smokers, occupations with highest COPD prevalence included secretaries, stenographers, and typists (4.5%), financial records processing (4.1%), and health service (3.8%).

Table 3 Estimated prevalence of COPD, prevalence odds ratio (POR), number of attributable cases (AC), and the population attributable fraction (PAF) by occupation—U.S. working adults (≥ 18 years), 1997–2004

		COPD prevalence
				Never smokers
Occupation	Workers^†	% (95% CI)	POR (95% CI)	% (95% CI)	AC^†	PAF%
Health service	2,649	7.3 (6.3–8.4)	1.78 (1.48–2.14)	3.8 (2.7–4.8)	85	1.75
Other protective service	967	5.0 (3.7–6.3)	1.63 (1.21–2.20)	2.9 (1.4–4.3)	19	0.36
Material moving equipment operators	1,181	4.3 (2.9–5.7)	1.59 (1.12–2.25)	1.9 (0.6–3.2)	19	0.37
Personal service	2,760	5.8 (5.0–6.7)	1.50 (1.24–1.82)	3.8 (2.9–4.8)	54	1.05
Motor vehicle operators	4,022	4.2 (3.6–4.8)	1.39 (1.15–1.69)	2.3 (1.6–3.0)	48	0.93
Food service	5,380	5.1 (4.4–5.7)	1.39 (1.17–1.64)	2.8 (2.1–3.5)	76	1.49
Fabricators, assemblers, inspectors and samplers	2,653	4.6 (3.7–5.4)	1.38 (1.11–1.72)	3.4 (2.3–4.4)	33	0.65
Cleaning and building service	3,092	4.8 (4.0–5.5)	1.37 (1.13–1.67)	2.9 (2.1–3.8)	40	0.78
Other administrative support	11,355	5.1 (4.7–5.5)	1.35 (1.18–1.55)	3.8 (3.3–4.3)	151	2.96
Financial records processing	2,115	5.8 (4.8–6.9)	1.32 (1.06–1.65)	4.1 (2.9–5.3)	30	0.58
Other sales	6,166	4.7 (4.1–5.3)	1.32 (1.12–1.55)	3.2 (2.5–3.9)	69	1.35
Freight, stock and material handlers	3,847	3.6 (3.0–4.2)	1.29 (1.04–1.58)	2.3 (1.6–2.9)	30	0.59
Secretaries, stenographers, and typists	2,682	5.7 (4.8–6.6)	1.25 (1.03–1.52)	4.5 (3.5–5.6)	31	0.60
Machine operators and tenderers, except precision	4,397	4.0 (3.4–4.6)	1.23 (1.02–1.48)	2.5 (1.9–3.2)	32	0.63
Private household	690	5.1 (3.3–6.9)	1.21 (0.82–1.79)	2.3 (1.1–3.4)	6	0.12
Supervisors and proprietors	3,864	4.1 (3.4–4.8)	1.20 (0.98–1.46)	3.3 (2.4–3.1)	26	0.50
Mechanics and repairers	4,691	3.2 (2.7–3.8)	1.19 (0.96–1.48)	1.6 (1.0–2.2)	24	0.48
Health technologist and technicians	1,867	4.6 (3.7–5.5)	1.18 (0.93–1.49)	2.7 (1.8–3.6)	13	0.26
Precision production	3,401	3.7 (3.0–4.4)	1.17 (0.95–1.45)	1.8 (1.2–2.4)	18	0.36
Teachers, librarians, and counselors	6,940	4.1 (3.6–4.6)	1.17 (1.00–1.38)	3.4 (2.8–3.9)	41	0.81
Writers, artists, entertainers, and athletes	2,320	3.7 (2.9–4.4)	1.12 (0.88–1.42)	3.1 (2.2–4.0)	9	0.18
Officials and administrators, public administration	778	4.0 (2.7–5.3)	1.09 (0.76–1.56)	3.7 (2.1–5.4)	3	0.05
Sales representatives, commodities and finance	4,237	3.6 (3.0–4.1)	1.08 (0.89–1.30)	2.6 (2.0–3.2)	11	0.21
Management related	5,120	3.9 (3.3–4.4)	1.08 (0.91–1.28)	3.1 (2.5–3.7)	14	0.27
Other professional specialty	2,949	3.6 (2.9–4.3)	1.07 (0.86–1.34)	3.3 (2.4–4.2)	7	0.14
Mail and message distributing	891	3.5 (2.3–4.8)	1.02 (0.70–1.49)	2.0 (0.8–3.2)	1	0.01
Natural mathematical and computer scientists	2,649	2.6 (2.0–3.2)	1.01 (0.78–1.30)	2.2 (1.5–2.8)	1	0.01
Engineers	1,987	2.4 (1.6–3.1)	1.01 (0.73–1.41)	1.7 (0.8–2.5)	1	0.01
Other occupations	17,283	3.4 (2.4–2.9)	0.89 (0.78–1.03)	1.7 (1.4–2.0)
Managers and administrators, except public administration	12,508	3.4 (3.1–3.7)	1.00	2.5 (2.1–2.8)
Total	127,433^‡	4.0 (3.9–4.1)		2.8 (2.7–2.9)	892	17.4%

Prevalence odds ratio (POR): adjusted for age, sex, race, and smoking status.

PAF%: percentage of COPD cases attributable to employment.

†Estimated number in 1,000s.

*Reference group.

‡Missing occupations were not included.

Attributable cases (AC) and population attributable fraction (PAF)

Approximately 627,000 COPD cases were attributed to employment considering industries, resulting in the PAF of 12.2% (95% CI 12.1–12.5%) (Table 2). The social services, religious and membership organizations industry group was associated with the greatest number of AC (n = 72,000). This industry group include various industries such as social services; museum, art galleries, and zoos; labor unions; and membership organizations (e.g., Automobile Association).

Approximately 892,000 COPD cases were attributed to employment considering occupations, resulting in the PAF of 17.4% (95% CI 17.1–17.6%) (Table 3); PAF ranged from 0.01% in engineers to 3.96% in other administrative support. Other administrative support occupation was associated with the greatest number of AC (n = 151,000).

DISCUSSION

This population-based study has identified industries and occupations associated with a high prevalence of self-reported COPD. Several industries (transportation equipment; repair service; utilities and sanitary; textile mill and finished textile products; health services, except hospitals; other personal services; food bakery and dairy stores; and chemicals and allied products) had an increased PORs when compared with insurance, real estate and other finances. This finding is consistent with results of an analysis of the Third National Health and Nutrition Examination Survey (NHANES III) data by Hnizdo and colleagues ([9]). The authors reported elevated ORs for COPD for the food products manufacturing, utilities, and textile mill products manufacturing industries ([9]). A recent study reported that occupational exposure to organic solvents in the repair and maintenance and transportation industry was associated with increased risk for COPD ([24]). The U.S. Work-Related Lung Disease Surveillance Report 2000 listed trucking service; automobile repair and related services; nursing and personal care facilities: and hospitals industries with significantly elevated proportionate mortality ratio for COPD ([25]). The elevated, although not significantly, COPD prevalence in mining industry in our study is consistent with other studies showing increased prevalence of COPD in coal miners ([26],27,28) and hard-rock miners ([29]). A recent report indicated significantly elevated proportionate mortality ratio for COPD in coal mining ([30]). Our study also indicated an elevated prevalence of COPD in industries such as eating and drinking places; general merchandise store; social services; religious and membership organizations; and business services which were not previously reported. Smoking may be an important risk factor for COPD in these industries. For example, smoking prevalence in eating and drinking places was very high (39.7%) in the U.S. working adults in 2000 ([25]). Other risk factors and occupational exposures associated with COPD in these industries need to be investigated.

Among occupations, an increased COPD prevalence and POR in motor vehicle operators; freight, stock and material handlers; and machine operator and tenderers is consistent with result of an analysis of the NHANES III data ([9]) and other studies on motor vehicles ([31], [32]). Several other not previously reported occupations were also associated with elevated POR and included health service; food services; fabricators, assemblers, inspectors, and samplers; and cleaning and building service. There is limited information on specific exposures associated with the development of COPD among workers in these occupations. For example, a population-based study in New Zealand has reported an increased OR of COPD in bakers under food services and spray painters under mechanics and repairers ([33]). A recent study of Chinese metal fabricators indicated an increased risk of COPD ([34]). Exposures to fabricating chemicals, dusts, and irritating agents among fabricating occupation were associated with COPD ([21], [34]). Studies reported excess annual declines in forced expiratory volume in 1 second among individuals with COPD, which could be explained, in part, by the occupational exposures to dusts, gases, and fumes ([11], [14], [15]). Personal service occupations including hairdressers, barbers, and cosmetologists have high COPD prevalence. A New Zealand study reported that hairdressers complained of shortness of breath, which may be related to COPD ([33]). This may be due to exposures to cleaning solutions and indoor dusts. An elevated POR in cleaning and building services is consistent with studies reporting an association of COPD in cleaners with the use of bleach and other cleaning products ([35], [36]).

The estimated PAF indicates that 17.4% of all COPD cases can be associated with occupations. This estimate is similar to that of 15% reported by the American Thoracic Society ([18]) and of 19% reported from the NHANES study ([9]). The differences may be due to the sample population and reference groups. The overall COPD prevalence (4%) among workers in this study is lower than the prevalence (8.5%) reported for the U.S. general population ([8]), which may be due, in part, to healthy worker effect (i.e., severely ill and disabled persons may be excluded from employment). A previous study based on the NHANES III data reported COPD prevalence in workers of 7.0% ([9]). The lower estimate may be due to the age criteria used for the NHANES III data analysis (NHANES inclusion criteria of 35–75 compared to ≥ 18 years for our study) and the definition of COPD. Their COPD definition was based on Global Initiative for Chronic Obstructive Lung Disease (GOLD) working group spirometric criteria ([37]) compared to questionnaire-based criteria in our study. The reported prevalence of COPD in adults (aged ≥ 18 years) in other studies ranges from 4% in North America to 10% in Europe ([3], [38]).

Our study showed higher prevalence of COPD in women than men, and in whites than blacks and other races. These findings are consistent with other studies ([9], [39],40,41,42). Increased prevalence of COPD in women could be due to increase in smoking prevalence in women and to increased indoor air pollution due to biomass fuels (i.e., natural gas) used at home ([42]). For racial differences in COPD, an extensive review of data from NCHS and from population-based studies reported that the prevalence of COPD was lower in blacks compared with whites ([40]). A study on COPD in a New Mexico population sample of Hispanics and whites reported that the prevalence of COPD in Hispanics was lower than the prevalence in whites ([41]).

Cigarette smoking remains the major risk factor for COPD ([6], [20], [43]). However, the presence of COPD in non-smokers indicates association of other factors (e.g., occupational exposures) with the disease development. In this study, never smoking secretaries, stenographers, and typists had the highest COPD prevalence. A recent study of office workers (e.g., clerks, secretaries, and other administrative workers) reported that exposures to paper dust and fumes from photocopiers and printers in office environments were associated with increased risk of chronic bronchitis ([44]). Health service occupation also had high COPD prevalence. Health service workers may be exposed to organic and inorganic dusts, isocyanates, and irritant gases which are associated with COPD ([16], [45]). Nonsmoking teachers, librarians, and counselors and fabricators, assemblers, inspectors and samplers were also associated with an elevated COPD prevalence in this study. A study in art teachers has reported an increased COPD possibly associated with the regular use of water-based paints, lacqers, solvent-based products, and glues ([46]). Fabricators are usually exposed to fabricating chemicals and irritating substances associated with COPD ([21], [34]). In addition, studies have reported that the development of COPD could result from exposures to second-hand smoking at home or workplaces ([47]), or chronic exposures to air pollutants like ozone, nitrogen dioxide, and sulfur dioxide ([48], [49]).

The strengths of the present study include a large sample size over eight years that allowed for increased precision of the prevalence estimates by industry and occupation. However, there are some limitations to this study. First, the NHIS was not designed to be representative of the U.S. employment patterns and is not focused on occupational hazards. Second, the COPD attribution to occupations and industries is subject to misclassification over a subject's lifetime. NHIS collected information on occupations and industries for adults who were working in the week prior to the survey interview. Industries and occupations collected may not be homogeneous with respect to potential exposures because some people could have changed jobs and thus be exposed to different substances over time. It is possible that the relevant occupational exposures associated with the development of COPD in an individual occurred in a different occupation to that reported. Third, there could be a self-reporting bias for COPD. Respondents may have under- or over-reported COPD and no pulmonary function data were available to confirm COPD, and the accuracy of self-reported physician's diagnosis of COPD collected in NHIS is not known. A study based on NHANES III data reported a good relationship between self-reported respiratory symptoms and lung function, and potential misreporting bias did not have a large effect on the results ([8]). Diagnosis of COPD without confirmation by pulmonary function testing may lead to some over-diagnosis ([19]). Moreover, the cross-sectional study design limits the ascertainment of causal exposure. Finally, PAF underestimates the proportion of COPD cases causally related to employment ([50]).

In conclusion, the elevated PORs observed for certain industries and occupations suggest that workplace exposures may be placing individuals at risk for occupational COPD. Consequently clinicians and healthcare policy-makers may need to implement or strengthen current prevention strategies. Primary prevention should include the elimination or reduction of exposures to substances known to be associated with work-related COPD as well as smoking cessation intervention in the workplace ([51], [52]). Secondary prevention can be achieved through epidemiologic surveillance and medical screening based on spirometry (Spirometry Longitudinal Data Analysis at http://www.cdc./niosh/topics/spirometry/spirola.html) for early case detection. Further epidemiologic studies are needed to identify the specific exposures and/or other factors associated with COPD among workers in these industries and occupations. The estimated total industry and occupation PAF provides useful information on occupationally attributed burden of COPD and indicates occupations and industries that could benefit from public health interventions.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the National Institute for Occupational Safety and Health.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

This article is not subject to United States' copyright law.