The Relationship of Educational Attainment with Pulmonary Emphysema and Airway Wall Thickness

Transcription

1 The Relationship of Educational Attainment with Pulmonary Emphysema and Airway Wall Thickness Miriam Gjerdevik 1,2, Thomas B. Grydeland 3,4, George R. Washko 5, Harvey O. Coxson 6, Edwin K. Silverman 7, Amund Gulsvik 4, and Per S. Bakke 4 1 Norwegian Registry of Long-Term Mechanical Ventilation, 2 Norwegian Registry of Chronic Obstructive Pulmonary Disease, 3 Department of Thoracic Medicine, Haukeland University Hospital, Bergen, Norway; 4 Department of Clinical Science, University of Bergen, Bergen, Norway; 5 Division of Pulmonary and Critical Care Medicine, Brigham and Women s Hospital, Boston, Massachusetts; 6 Department of Radiology and Centre for Heart Lung Innovation, University of British Columbia, Vancouver, British Columbia, Canada; and 7 The Channing Division of Network Medicine and Pulmonary and Critical Care Division, Brigham and Women s Hospital and Harvard Medical School, Boston, Massachusetts Abstract Rationale: Low educational attainment is a risk factor of chronic obstructive pulmonary disease(copd). There is limited knowledge on the relationship between educational level and computed tomography measures of emphysema and airway wall thickness (AWT). Objectives: We hypothesized that low educational attainment is associated with increased emphysema and AWT in ever-smokers with and without COPD. Methods: We included 462 and 485 ever-smokers with and without COPD in a cross-sectional study, aged years. The sample was divided into groups reflecting educational attainment: primary, secondary, and university. We performed linear regression to examine associations between educational attainment and both emphysema and AWT separately for those with and without COPD. We adjusted for sex, age, smoking status, age of onset of smoking, pack-years, height, and body mass index. Measurements and Main Results: Compared with university education, in subjects with COPD, primary education was associated with a 68.1% (95% confidence interval = %; P = 0.01) relative increase in emphysema and secondary education was associated with a 50.6% (95% confidence interval = %; P = 0.02) relative increase. There was a nonsignificant trend toward an association between lower educational attainment and increased emphysema among those without COPD (P = 0.18), yet greater age appeared to modify this association (P = 0.01). We did not detect significant linear relationships between educational attainment and AWT in subjects with or without COPD. Conclusions: Lower educational attainment was associated with increased emphysema among adults with COPD. Among those without COPD, this association was more pronounced with increasing age. No significant linear relationship between educational attainment and AWT was found. Clinicians treating adults with emphysema should keep in mind that factors related to low education beyond that of smoking and occupational dust exposure might be of importance to the disease. Keywords: chronic obstructive pulmonary disease; quantitative computed tomography imaging; educational level (Received in original form October 27, 2014; accepted in final form March 30, 2015 ) Supported by GlaxoSmithKline Research and Development Limited, UK (data collection). Author Contributions: M.G. performed the statistical analyses and drafted and revised the manuscript; T.B.G., G.R.W., and E.K.S. revised the draft manuscript; H.O.C. and A.G. took part in the data collection and revised the draft manuscript; P.S.B. took part in the data collection and drafted and revised the manuscript. Correspondence and requests for reprints should be addressed to Miriam Gjerdevik, M.Sc., Department of Thoracic Medicine, Haukeland University Hospital, 5021 Bergen, Norway. address: miriam.gjerdevik@igs.uib.no This article has an online supplement, which is accessible from this issue s table of contents at Ann Am Thorac Soc Vol 12, No 6, pp , Jun 2015 Copyright 2015 by the American Thoracic Society DOI: /AnnalsATS OC Internet address: Low socioeconomic status (SES) is a wellknown risk factor in lung health (1). SES, measured in terms of educational level, occupational status, or income level, has been shown to be related to respiratory symptoms (2), spirometry values (3), and gas diffusion capacity (4), as well as incidence and mortality of chronic obstructive pulmonary disease (COPD) (1, 5). A common characteristic for the aforementioned studies is that they do not provide any definitive data as to where in Gjerdevik, Grydeland, Washko, et al.: The Relationship of SES to Emphysema and AWT 813

2 the lungs the exposures related to low SES have their effects. Such knowledge may help us to understand the underlying mechanisms in the associations between SES and pulmonary health. Computed tomography (CT) of the thorax offers the potential to subdivide the pathological processes of COPD into emphysema and airway disease, as well as to quantify them (6). As CT is increasingly used to phenotype COPD (7), knowledge of what factors influence the CT indices is important to enhance interpretation of analyses and to allow interstudy comparison of results. Limited data are available as to the relationship between SES and indices of CT thorax. A recent report from a prospective cohort study with participants selected from six centers across the United States surprisingly found that percent emphysema was related to high SES (8). No data are available on the relationship between SES and airway wall thickness (AWT). Based on studies showing that low SES is related to respiratory symptoms indicative of both airway disease (cough, phlegm) and emphysema (dyspnea) (2, 9), we hypothesized that low educational attainment, as a measure of SES, would be associated with both CT-derived indices of increased emphysema and increased AWT. The objective of our present work was to test this hypothesis in ever-smokers with and without COPD in a cross-sectional study. This study offered the opportunity to examine the relationship of educational attainment and the CT indices independent of sex, age, smoking status, age of onset of smoking, pack-years, height, body mass index (BMI), and occupational dust exposure. We also wanted to examine to what extent education interacted with sex, age, smoking status, and pack-years in its relationship with emphysema and AWT. Some of the results of this study have been previously reported in the form of an abstract (10). Methods The study comprised 462 subjects with COPD and 485 subjects without COPD, all ever-smokers aged years, recruited from a hospital patient register and from two general population studies in the period All participants in the current study were part of the Genetic COPD study (11, 12), and constituted approximately one-half of the Genetic COPD study population (947 of 1,909) that received an optional CT scan. Subjects recruited from the hospital patient register comprised 70% of the patients with COPD and 7% of the subjects without COPD. Inclusion criteria were: (1) white; (2) age above 40 years; (3) current or former smoker with at least 2.5 pack-years of smoking history; and (4) no severe a 1 - antitrypsin deficiency. Individuals with asthma were not excluded. A complete list of inclusion and exclusion criteria is given elsewhere (13). The study area, the county of Hordaland on the west coast of Norway, is sparsely populated approximately 33 inhabitants per square kilometer. The largest city, Bergen, has nearly 250,000 inhabitants. Levels of air pollution are low the 24-hour mean concentration of particulate matter less than 10 mm in diameter has not exceeded 50 mg/m 3 on any day in the last 5 years, whereas the annual average of nitrogen dioxide (NO 2 )is40mg/m 3 (14). All participants gave informed consent before inclusion in the study, and the study was approved by the Regional Committee for Medical and Health Research Ethics of Western Norway. All subjects completed a chest CT scan and questionnaires asking for information on educational attainment, smoking status, and occupational dust exposure (for details, see the online supplement). The sample was divided into educational levels: primary (primary and lower secondary school; in Norway, 9 yr of mandatory education); secondary (upper secondary school; not mandatory); and university. Subjects were examined by spirometry, according to American Thoracic Society standards, using a Vitalograph 2,160 Spirometer (Vitalograph Ltd, Buckingham, UK) before and after bronchodilation with 0.4 mg of salbutamol. Subjects were assessed at least 6 weeks after any respiratory infection, but were not asked to withhold regular medication. Norwegian reference values for FEV 1 and FVC were used (15). Subjects with COPD had a postbronchodilator FEV 1 /FVC ratio less than 0.70 and FEV 1 less than 80% of predicted. Subjects without COPD had a postbronchodilator FEV 1 /FVC of 0.70 or greater and FEV 1 of 80% of predicted or greater. The subjects with COPD were classified into Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages A D, as previously described (16). CT The CT scans were acquired using a GE LightSpeed Ultra CT scanner (120 kvp, 200 ma; GE Healthcare, Milwaukee, WI) at suspended full inspiration (apex to base) using 1-mm slice thickness at 20-mm intervals (13). The CT scans were reconstructed using both a low spatial frequency reconstruction algorithm ( standard ) for density measurements and a high spatial frequency algorithm ( bone ) for airway measurements. CT-derived total lung volume was measured. The extent of emphysema was primarily assessed using the percentage of lung voxels with X-ray attenuation values less than 2950 Hounsfield units (percentage low attenuation areas). Percent emphysema for the whole lung was calculated (17). To avoid potential bias from different distribution of airway sizes between subjects, airway wall thickness is presented as the square root of the wall area for a standardized airway with an internal perimeter of 10 mm (AWT-Pi10) (13, 18). Statistical Methodology Tests for trends across the ordered groups of educational attainment was assessed using linear regression for normally distributed data, the nonparametric test by Cuzick for continuous and skewed data (19), and the Chi-squared test for dichotomous outcomes. We conducted linear regression analyses to measure the relationship of educational attainment to emphysema and AWT. The analyses were performed separately for subjects with and without COPD. The level of emphysema was log transformed to improve the linear relationship between the dependent and independent variables. A unit increase for the independent variable was thus associated with an average increase of 100(exp(b) 2 1)% in the dependent variable, while holding all other predictor variables constant, b being the regression coefficient of the linear regression. In the multivariate analyses, we adjusted for sex, age, smoking status, age of onset of smoking, pack-years, height, and BMI (model A). Additional analyses were performed, adding occupational dust exposure (model B) and asthma (diagnosis 814 AnnalsATS Volume 12 Number 6 June 2015

3 confirmed by doctor) as covariates to model A, one at a time. Occupational dust exposure was the only covariate with missing values. As less than 2.0 and 2.5% were missing in the COPD and non-copd groups, respectively, these subjects were simply excluded from model B. Finally, we tested the a priori decided interactions between educational level and the variables, sex, age, smoking status, and pack-years, employing the likelihood ratio test. Each interaction term was added separately to model A. All statistical analyses were conducted using STATA/IC version 12.1 for Windows (StataCorp LP, College Station, TX). Conventionally, the two-sided significance level was set at Results Table 1. Characteristics of the study population Characteristics Table 1 shows the characteristics of subjects with and without COPD. There was a higher proportion of men among the subjects with COPD than among those without. The subjects with COPD were older, had a smaller BMI, were more likely to be current smokers, and had a larger total smoking exposure in terms of packyears than the subjects without COPD. Educational attainment was not evenly distributed the subjects without COPD had more education; they were also less exposed to occupational dust. Of those diagnosed with COPD, almost 40% were classified in the most severe GOLD stage D. No significant trend was observed in the proportion of men and current smoking by educational attainment, either in those with or in those without COPD, whereas the age declined significantly by level of education for both groups (Table 2). The age of onset of smoking increased and the median number of pack-years decreased significantly by educational attainment in the non-copd group. FEV 1 % predicted increased significantly with higher education for subjects diagnosed with COPD. Occupational airborne exposure increased with less education in both subjects with and without COPD, whereas no significant linear trend by education was seen for BMI. The Relationship of Educational Attainment to Emphysema The median extent of emphysema increased with lower education (Table 3), the trend being significant in subjects with COPD Subjects with COPD Subjects without COPD n = 462 (48.8%) n = 485 (51.2%) Men, % Age, yr, mean (SD) 64.3 (9.4) 55.5 (9.4) Current smokers, % Pack-years, median (25th, 26.9 (19.0, 40.0) 16.0 (9.0, 25.4) 75th percentiles) Age of onset of smoking, 18.5 (5.1) 18.4 (4.7) mean (SD) BMI, kg/m 2, mean (SD) 25.5 (4.8) 26.4 (4.0) PB FEV 1 pp, mean (SD) 52.6 (16.9) 94.7 (9.0) Educational attainment, % Primary school Secondary school University education Occupational dust exposure, 12.2 (14.8) 5.6 (10.3) mean (SD) GOLD classification, % Stage A 23.2 Stage B 32.7 Stage C 6.5 Stage D 37.6 Definitions of abbrivations: BMI = body mass index; COPD = chronic obstructive pulmonary disease; GOLD = Global Initiative for Chronic Obstructive Lung Disease; PB FEV 1 pp = post-brochodilatory FEV 1 % predicted. There is missing information on 9 subjects with COPD and 12 subjects without COPD for occupational dust exposure and 13 subjects on GOLD classification. (P = 0.002), but not in those without COPD (P = 0.183). In the COPD group, the median level of emphysema was 12.0% (25th/75th percentiles: 3.2/18.2%) for those with a primary education, 6.1% (25th/75th percentiles: 2.2/17.4%) for those with a secondary education, and 3.4% (25th/75th percentiles: 0.9/11.3%) for those with a university education. The corresponding figures in the non-copd group were 0.7% (25th/75th percentiles: 0.2/1.6%), 0.5% (25th/75th percentiles: 0.2/1.2%), and 0.5% (25th/75th percentiles: 0.2/1.1%) for those with a primary, secondary, and university education, respectively. The univariate regression analyses showed that both primary and secondary education were related to increased emphysema compared with university education (see Table 4). These associations were, however, only significant in subjects with COPD. The strength of these associations was reduced in the multivariate analyses (model A, Table 4). Compared with university education, in subjects with COPD, there was a 68.1% (95% confidence interval [CI] = %; P = 0.009) relative increase in emphysema among those with a primary education, and a 50.6% (95% CI = %; P = 0.023) relative increase among those with a secondary education. In subjects without COPD, there was a 21.5% (95% CI = to 72.3%; P = 0.272) relative increase in emphysema among those with a primary education, and a 21.3% (95% CI = 25.9 to 56.4%; P = 0.135) relative increase among those with a secondary education. Additional analyses also adjusting for occupational dust exposure (model B, Table 4) showed similar, but somewhat weaker, associations to model A. Information on all variables in the multivariate analyses is provided in Table E1 in the online supplement. The coefficients were only marginally affected when additionally adjusting for asthma (data not shown). Linear tests for trends, adjusting for confounders, were also performed (Table E2). The overall results were consistent with those presented in Table 4. In subjects without COPD, we found a significant interaction between educational attainment and age on the extent of emphysema, after adjusting for sex, age, smoking status, age of onset of smoking, pack-years, height, and BMI (P = using the likelihood ratio test; see Figure 1A). The negative association of Gjerdevik, Grydeland, Washko, et al.: The Relationship of SES to Emphysema and AWT 815

4 Table 2. Educational attainment by sex, age, airborne exposures, lung function, and body mass index in subjects with and without chronic obstructive pulmonary disease Level of Education Primary School Secondary School University Education P Value* Subjects with COPD n = 140 n = 270 n =52 Men, % Age, yr, mean (SD) 66.5 (9.5) 63.5 (9.0) 62.5 (9.7) Current smokers, % Pack-years, median (25th, 75th percentiles) 27.1 (19.1, 40.0) 26.9 (19.5, 40.0) 26.8 (16.8, 37.1) Age of onset of smoking, yr, mean (SD) 18.0 (4.7) 18.7 (5.1) 18.9 (5.6) Occupational dust exposure, mean (SD) 16.5 (15.8) 11.6 (14.4) 3.5 (8.8),0.001 BMI, kg/m 2, mean (SD) 25.0 (4.4) 25.7 (5.0) 25.6 (4.4) PB FEV 1 pp, mean (SD) 50.0 (15.1) 52.7 (17.2) 59.2 (18.1) Subjects without COPD n =77 n = 289 n = 119 Men, % Age, yr, mean (SD) 58.2 (11.0) 55.3 (9.1) 54.1 (9.0) Current smokers, % Pack-years, median (25th, 75th percentiles) 20.5 (14.0, 31.0) 16.0 (8.6, 25.6) 13.5 (7.5, 22.2),0.001 Age of onset of smoking, yr, mean (SD) 17.9 (4.1) 18.3 (4.5) 19.2 (5.4) Occupational dust exposure, mean (SD) 9.5 (14.3) 6.1 (10.2) 1.7 (5.1),0.001 BMI, kg/m 2, mean (SD) 26.6 (4.2) 26.6 (4.2) 26.0 (3.3) PB FEV 1 pp, mean (SD) 93.3 (8.5) 94.9 (8.9) 95.0 (9.6) Definitions of abbrivations: BMI = body mass index; COPD = chronic obstructive pulmonary disease; PB FEV 1 pp = post-brochodilatory FEV 1 % predicted. Missing information: COPD, occupational dust exposure: 1 primary education, 7 secondary education, 1 university education; non-copd, occupational dust exposure: 2 primary education, 9 secondary education, 1 university education. *Tests for trends across educational attainment: linear regression analysis for measuring linear trends for normally distributed data, nonparametric test for trend for skewed data, and Chi-squared test for dichotomous outcome. lower education on emphysema was more pronounced with increasing age, and this relationship still held after adjustment for occupational dust exposure. Compared with university-educated 40 year olds, primary-educated 80 year olds had a 10-fold relative increase of emphysema, whereas university-educated 40 year olds had a 2.5-fold relative increase. A somewhat similar trend was seen in subjects with COPD, although not to the same extent, and this was not significant (interaction P =0.357; see Figure 1B). None of the other apriori selected interactions were significant. The Relationship of Educational Attainment to AWT The mean AWT-Pi10 did not vary significantly between educational attainment, either in subjects with or in those without COPD (Table 3). In subjects with COPD, the mean AWT-Pi10 was 4.96 mm (SD = 0.35) for the primary educated, 4.92 mm (SD = 0.33) for those with a secondary education, and 5.00 mm (SD = 0.35) for the university educated. In subjects without COPD, the corresponding numbers were 4.77 mm (SD = 0.32), 4.76 mm (SD = 0.28), and 4.77 mm (SD = 0.30) for those with a primary, secondary, and university education, respectively. No significant relationship between education and AWT-Pi10 was observed in the univariate or multivariate analyses (Table 4). Information on all variables in the multivariate analyses is provided in Table E1. We observed, however, a significant interaction between educational attainment and smoking status in relationship to AWT in subjects without COPD (P = applying the likelihood ratio test). In current smokers, compared with university education, there was a 0.08-mm (95% CI = to 0.19; P = 0.159) increase in AWT among those with a primary education, after adjusting for sex, age, age of smoking onset, pack-years, height, and BMI (see Table E3). The increase among Table 3. Extent of emphysema (percentage low attenuation areas) and airway wall thickness by educational attainment %LAA950 AWT-Pi10, mm n Median (25th, 75th Percentiles) n Mean (SD) Subjects with COPD Primary school (3.18, 18.19) (0.35) Secondary school (2.18, 17.43) (0.33) University education (0.95, 11.30) (0.35) P value* Subjects without COPD Primary school (0.22, 1.61) (0.32) Secondary school (0.22, 1.25) (0.28) University education (0.17, 1.10) (0.30) P value* Definitions of abbrivations: %LAA950 = % low attenuation areas,2950 Hounsfield units; AWT-Pi10 = airway wall thickness at an internal perimeter of 10 mm; COPD = chronic obstructive pulmonary disease. *Tests for trends across educational attainment: linear regression analysis for measuring linear trends for normally distributed data, nonparametric test for trend for skewed data. 816 AnnalsATS Volume 12 Number 6 June 2015

5 Table 4. Linear regression analyses of change in emphysema level (percentage low attenuation areas) and change in airway wall thickness by level of education Emphysema* Airway Wall Thickness Univariate Analyses Multivariate Analyses Univariate Analyses Multivariate Analyses Model A Model B Model A Model B % Change* (95% CI) % Change* (95% CI) % Change* (95% CI) Change in mm (95% CI) Change in mm (95% CI) Change in mm (95% CI) Subjects with COPD Primary education x (36.6, 229.6) 68.1 jj (14.2, 147.6) 54.5 (3.3, 131.0) (20.16, 0.06) 0.00 (20.10, 0.10) 0.01 (20.10, 0.11) Secondary education 59.3 (5.6, 140.2) 50.6 (5.7, 114.6) 41.4 (21.7, 103.3) (20.19, 0.02) (20.14, 0.04) (20.14, 0.05) Subjects without COPD Primary education 29.5 (210.9, 88.2) 21.5 (214.2, 72.3) 15.5 (219.6, 65.9) 0.01 (20.08, 0.09) (20.11, 0.05) (20.11, 0.05) Secondary education 16.5 (211.9, 53.9) 21.3 (25.9, 56.4) 19.2 (28.2, 54.7) (20.07, 0.06) (20.08, 0.03) (20.09, 0.03) Definitions of abbrivations: CI = confidence interval; COPD = chronic obstructive pulmonary disease. University education is used as the reference category. *The emphysema level is log transformed, and the displayed coefficients are the relative change in percent. Model A: adjusted for sex, age, smoking status (ex- vs. current smoking), pack-years, age-of-onset smoking, height, and body mass index. Model B: adjusted for sex, age, smoking status (ex- vs. current smoking), pack-years, age-of-onset smoking, height, body mass index, and occupational dust exposure. x P, jj P, P, current smokers with a secondary education was 0.03 mm (95% CI = to 0.12; P = 0.441). In ex-smokers, there was a 0.11-mm (95% CI = to 20.01; P = 0.032) decrease in AWT among those with a primary education, and a 0.6-mm (95% CI = to 0.01; P = 0.096) decrease among those with a secondary education. The interaction of education and smoking status on AWT did not reach the level of significance in the COPD group (P = 0.143). Stratified analyses are nevertheless shown in Table E3. Less education was related to increased AWT among current smokers, although not significantly, and no linear dose response relationship was observed in ex-smokers. None of the other apriori selected interactions were significant, either in subjects with or in those without COPD. Discussion The main findings were: (1) in subjects with COPD, those with low educational attainment had more emphysema when adjusting for sex, age, smoking status, height, BMI, and occupational dust exposure (the same tendencies were noted in those without COPD, although not significantly); (2) the inverse association between emphysema and education was stronger among older adults this interaction was only significant in subjects without COPD, although the same associations were observed in the COPD group; and (3) no significant linear relationship was found between educational attainment and AWT. A 10.0 B Fold change of emphysema Fold change of emphysema Primary Secondary University 40 Age Primary Secondary University 40 Age Educational attainment Educational attainment Figure 1. Fold change of emphysema according to age and educational attainment, compared with university-educated 40-year-old subjects (indicated by a fold change of 1). Analyses were adjusted for sex, age, smoking status (ex- vs. current smoking), pack-years, age-of-onset smoking, height, and body mass index. (A) Subjects without chronic obstructive pulmonary disease (COPD), interaction P = (B) Subjects with COPD, interaction P = Gjerdevik, Grydeland, Washko, et al.: The Relationship of SES to Emphysema and AWT 817

6 Educational level is the most frequently used method of assessing SES. Information on educational attainment is easy to obtain, and is often more accurate than information regarding income (20). Our results may therefore, to some extent, be generalized and discussed in the context of SES. Previous studies have found associations between low SES and various indices of COPD, such as respiratory symptoms and impaired lung function. We have extended this knowledge by providing an anatomical correlate to the observed relationships. Our findings on the relationship between educational attainment and emphysema contradict those of a recent prospective cohort study with participants selected from six centers across the United States, showing increased emphysema with higher SES (8). That study did not adjust for COPD status, whereas our analyses were performed separately in subjects with and those without COPD. The severity of emphysema and COPD status is highly correlated. One would thus expect our associations between educational level and emphysema to be even stronger if we did not control for this variable. As stated by the authors themselves, their surprising finding could have been due to subjects with low SES, relative to subjects with high SES, having systematically poorer compliance with technician instructions to hold their breath at maximum inspiration during the CT scanning (8). This would work to underestimate emphysema in subjects with low SES. The MESA (Multi- Ethnic Study of Atherosclerosis) study (8) could have been more subject to such a bias than the present study, as the former was a multicenter study with many technicians involved, whereas the present study is a single-center study with few, but highly trained, technicians. However, as neither of the studies used spirometry-gated CT scans, it is hard to examine if dissimilarities in degree of inhalation by SES between the two studies could explain the different study results. The MESA study adjusted for self-reported passive smoking, a history of hay fever, and a family history of emphysema among siblings, in addition to the covariates in the present study. Hence, our study would be more subject to residual confounding than the MESA study. As stated by the MESA study authors, however, the direction of the SES emphysema relationship was not altered after these adjustments. To our knowledge, the current study is the first to examine the relationship of educational attainment to AWT. Education versus Severity of Emphysema The inverse relationship between low education and level of emphysema may have several explanations. As people tend to stay in the same socioeconomic group into which they are born, early-life airborne exposures, more frequent in low than in high SES, may impair alveolar formation during the first 3 years of life (21) and later on during alveolar development. Potential exposures are environmental tobacco smoke (22) and particles emitted from traffic (23). Low SES is related to higher intake of cured meat (e.g., bacon, hot dogs), which was associated with higher incidence of a physician-diagnosed COPD among 71,000 women followed for 16 years in the Nurses Health Study (24). Cured meat contains nitrite as a preservative. Animal and in vitro studies have shown that nitrite intake produces effects that mimic agerelated lung damage, and may work through modification of extracellular matrix proteins, including fragmentation of elastin (24). As there is family clustering of both emphysema (25) and low SES, one could speculate that there is a genetic susceptibility to airborne pollutants other than smoking (and occupational airborne exposure), at least partly explaining the education emphysema relationship. We observed that the negative association of lower education with the level of emphysema was more prominent with increasing age, the interaction being significant in the non-copd group. There are several interpretations for this finding. First, it could indicate that exposures related to low SES become relatively more important as the aging of the lung progresses. Second, it could be a cohort effect. Exposures related to low SES might have been more harmful to the lungs when the oldest members of our study were young as compared with that experienced by the youngest participants. Only a longitudinal design can disentangle the first and second interpretation. Third, the fact that younger subjects have received considerably more formal teaching than the older subjects implies that education might be a poorer means of discriminating between SES in younger and older subjects (26). Finally, recall bias could influence our results. However, if the elderly forget their educational level to a greater extent than the younger persons, this would tend to underestimate the observed relationship. Relationship between Education and AWT No significant linear relationship between educational attainment and AWT was found. However, there was a significant interaction between education and smoking status in the non-copd group. In those without COPD, we observed that low educational attainment was related to increased AWT in current smokers, but not in ex-smokers. The same associations with smoking were seen in those with COPD. This interaction may imply that the combined effect of current smoking and exposures related to low education may cause airway inflammation, reflected as increased AWT on CT. There are several potential exposures related to low SES. First, there are agents from poor indoor and outdoor air quality. These exposures may come into play in adult life, or they may occur in childhood, reducing lung growth and causing disorders that may persist into adulthood or make the adult airways more vulnerable to other air pollutants. Particulate matter exposures in early life are hypothesized to impair the innate immune system and make the airway more susceptible to infection in later life (27). Second, regular physical exercise is more often seen in high SES, and is associated with reduced inflammatory markers, such as IL-6, TNF-a, and C-reactive protein (28). Finally, although there are indications that the airway microbiome is not affected by smoking status (29), one may speculate that the combined low-ses exposures and current smoking status may affect the lung microbiome, causing increased airway inflammation (30). In the non-copd group, we observed an opposite association between educational attainment and AWT in ex-smokers as compared with that described in current smokers. This finding is hard to explain, but interpretations should be approached with caution in a cross-sectional study that could be subject to a selection bias. Strengths and Limitations of the Study This study was a large, single-center study, which allows for extensive adjustment for important confounders. Almost all the 818 AnnalsATS Volume 12 Number 6 June 2015

7 subjects without COPD were sampled from a community study with high response rates, and the sample is representative of the population at large with respect to sex, age, and SES as measured by educational attainment (31). All CT scans were performed with the same scanner. Finally, information on educational attainment was used throughout our analyses as a measure of SES, which is preferable to using income and occupational status (26), although at least two measures of SES are recommended. In most subjects, formal education was finished before the onset of disease (20). Hence, the problem of whether SES is a cause or a consequence of the disease is reduced. Respiratory symptoms in childhood seem to have only a minor adverse effect on educational attainment (32). There are some limitations to the present study. First, although subjects with a total smoking consumption as low as 2.5 pack-years were included in the study, we did not include never-smokers. We assessed smoking as pack-years, current smoking status, and age of onset of smoking, but other aspects of smoking behavior, such as depth of inhalation or brand of cigarettes smoked, could influence the relationships between education and CT indices. Second, self-reported occupational exposure to dust is only a crude indicator of exposure, and there may be residual confounding after adjusting for this information. However, the occupational data have a high specificity (33). Third, the CT protocol used for this study did not allow us to determine the generations of the airways measured: the generations might have differed between subjects with small lungs and those with large lungs. Adjustment for lumen size and number of airways measured did not alter the relationships observed in Table 4. Fourth, the method for measuring AWT prevented us from measuring airways smaller than 2 mm in diameter. This is suboptimal, as COPD is primarily a smallairway disease. It has, however, been shown that measurements of larger airways can be used as a surrogate for the processes in the smaller airways (34), and that the inflammatory response in smaller airways can also be seen in larger airways (35). Finally, it should be noted that our findings might be less generalizable due to local particularities, such as climate or sociodemographic characteristics. Implications of the Study Factors related to low education beyond those of smoking and occupational exposure to dust might be of importance to emphysema. Clinicians treating adults with emphysema should consider screening the patients for these risk factors. Second, emphysema and AWT assess different components of the COPD syndrome. As educational level is a measure of SES, one should consider controlling for SES in the same way as smoking in case control studies of environmental and genetic effects on these phenotypes, to avoid bias. This requirement is further strengthened by indications that, in such studies, the cases are recruited from lower SES groups than the control subjects (11), whereas, in other studies, the source of the participants is not given (36). Conclusions This cross-sectional study of ever-smokers with and without COPD has shown that low educational attainment is related to increasing extent of emphysema among those with COPD. This association is more pronounced with increasing age among those without COPD. No significant linear relationship between educational attainment and AWT was found. Clinicians treating adults with pulmonary emphysema should keep in mind that factors related to low education beyond that of smoking and occupational dust exposure might be of importance to the disease. n Author disclosures are available with the text of this article at References 1 Prescott E, Vestbo J. Socioeconomic status and chronic obstructive pulmonary disease. Thorax 1999;54: Eagan TM, Gulsvik A, Eide GE, Bakke PS. The effect of educational level on the incidence of asthma and respiratory symptoms. Respir Med 2004;98: Johannessen A, Eagan TM, Omenaas ER, Bakke PS, Gulsvik A. Socioeconomic risk factors for lung function decline in a general population. Eur Respir J 2010;36: Welle I, Eide GE, Gulsvik A, Bakke PS. Pulmonary gas exchange and educational level: a community study. Eur Respir J 2004;23: Prescott E, Godtfredsen N, Vestbo J, Osler M. Social position and mortality from respiratory diseases in males and females. Eur Respir J 2003;21: Coxson HO, Mayo J, Lam S, Santyr G, Parraga G, Sin DD. New and current clinical imaging techniques to study chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2009;180: Han MK, Bartholmai B, Liu LX, Murray S, Curtis JL, Sciurba FC, Kazerooni EA, Thompson B, Frederick M, Li D, et al. Clinical significance of radiologic characterizations in COPD. COPD 2009;6: Lovasi GS, Diez Roux AV, Hoffman EA, Smith LJ, Jiang R, Carr JJ, Barr RG. Socioeconomic status is positively associated with percent emphysema on CT scan: the MESA lung study. Acad Radiol 2011;18: Bakke PS, Hanoa R, Gulsvik A. Educational level and obstructive lung disease given smoking habits and occupational airborne exposure: a Norwegian community study. Am J Epidemiol 1995;141: Gjerdevik M, Grydeland T, Gulsvik A, Bakke P. The relationship of socioeconomic status to emphysema and airway wall thickness: ERS International Congress [abstract]. Eur Respir J 2014: P Sørheim IC, Johannessen A, Grydeland TB, Omenaas ER, Gulsvik A, Bakke PS. Case control studies on risk factors for chronic obstructive pulmonary disease: how does the sampling of the cases and controls affect the results? Clin Respir J 2010;4: Zhu G, Warren L, Aponte J, Gulsvik A, Bakke P, Anderson WH, Lomas DA, Silverman EK, Pillai SG; International COPD Genetics Network (ICGN) Investigators. The SERPINE2 gene is associated with chronic obstructive pulmonary disease in two large populations. Am J Respir Crit Care Med 2007;176: Grydeland TB, Dirksen A, Coxson HO, Pillai SG, Sharma S, Eide GE, Gulsvik A, Bakke PS. Quantitative computed tomography: emphysema and airway wall thickness by sex, age and smoking. Eur Respir J 2009;34: Norwegian Environment Agency. Central database for air quality 2014 [accessed 2014 Jun 1]. Available from: 15 Gulsvik A, Tosteson T, Bakke P, Humerfelt S, Weiss ST, Speizer FE. Expiratory and inspiratory forced vital capacity and one-second forced volume in asymptomatic never-smokers in Norway. Clin Physiol 2001;21: Johannessen A, Nilsen RM, Storebø M, Gulsvik A, Eagan T, Bakke P. Comparison of 2011 and 2007 global initiative for chronic obstructive lung disease guidelines for predicting mortality and hospitalization. Am J Respir Crit Care Med 2013;188: Martinez FJ, Foster G, Curtis JL, Criner G, Weinmann G, Fishman A, DeCamp MM, Benditt J, Sciurba F, Make B, et al.; NETT Research Group. Predictors of mortality in patients with emphysema and Gjerdevik, Grydeland, Washko, et al.: The Relationship of SES to Emphysema and AWT 819

8 severe airflow obstruction. Am J Respir Crit Care Med 2006;173: Grydeland TB, Dirksen A, Coxson HO, Eagan TM, Thorsen E, Pillai SG, Sharma S, Eide GE, Gulsvik A, Bakke PS. Quantitative computed tomography measures of emphysema and airway wall thickness are related to respiratory symptoms. Am J Respir Crit Care Med 2010; 181: Cuzick J. A Wilcoxon-type test for trend. Stat Med 1985;4: Liberatos P, Link BG, Kelsey JL. The measurement of social class in epidemiology. Epidemiol Rev 1988;10: Svanes C, Sunyer J, Plana E, Dharmage S, Heinrich J, Jarvis D, de Marco R, Norbäck D, Raherison C, Villani S, et al. Early life origins of chronic obstructive pulmonary disease. Thorax 2010;65: Lovasi GS, Diez Roux AV, Hoffman EA, Kawut SM, Jacobs DR Jr, Barr RG. Association of environmental tobacco smoke exposure in childhood with early emphysema in adulthood among nonsmokers: the MESA-lung study. Am J Epidemiol 2010;171: Lopes FD, Pinto TS, Arantes-Costa FM, Moriya HT, Biselli PJ, Ferraz LF, Lichtenfels AJ, Saldiva PH, Mauad T, Martins MA. Exposure to ambient levels of particles emitted by traffic worsens emphysema in mice. Environ Res 2009;109: Jiang R, Camargo CA Jr, Varraso R, Paik DC, Willett WC, Barr RG. Consumption of cured meats and prospective risk of chronic obstructive pulmonary disease in women. Am J Clin Nutr 2008;87: Patel BD, Coxson HO, Pillai SG, Agustí AG, Calverley PM, Donner CF, Make BJ, Müller NL, Rennard SI, Vestbo J, et al.; International COPD Genetics Network. Airway wall thickening and emphysema show independent familial aggregation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2008;178: Bakke PS, Rönmark E, Eagan T, Pistelli F, Annesi-Maesano I, Maly M, Meren M, Vermeire Dagger P, Vestbo J, Viegi G, et al.; European Respiratory Society Task Force. Recommendations for epidemiological studies on COPD. Eur Respir J 2011;38: Grigg J. Particulate matter exposure in children: relevance to chronic obstructive pulmonary disease. Proc Am Thorac Soc 2009;6: Garcia-Aymerich J, Lange P, Benet M, Schnohr P, Antó JM.Regular physical activity modifies smoking-related lung function decline and reduces risk of chronic obstructive pulmonary disease: a population-based cohort study. Am J Respir Crit Care Med 2007; 175: Morris A, Beck JM, Schloss PD, Campbell TB, Crothers K, Curtis JL, Flores SC, Fontenot AP, Ghedin E, Huang L, et al.; Lung HIV Microbiome Project. Comparison of the respiratory microbiome in healthy nonsmokers and smokers. Am J Respir Crit Care Med 2013; 187: Jaspers I. Cigarette smoke effects on innate immune mechanisms in the nasal mucosa: potential effects on the microbiome. Ann Am Thorac Soc 2014;11:S38 S Eagan TM, Eide GE, Gulsvik A, Bakke PS. Nonresponse in a community cohort study: predictors and consequences for exposure disease associations. J Clin Epidemiol 2002;55: Olivieri M, Mirabelli MC, Plana E, Radon K, Antó JM, Bakke P, Benke G, D Errico A, Henneberger P, Kromhout H, et al. Healthy hire effect, job selection and inhalation exposure among young adults with asthma. Eur Respir J 2010;36: Bakke PS, Hanoa R, Gulsvik A. Relation of occupational exposure to respiratory symptoms and asthma in a general population sample: self-reported versus interview-based exposure data. Am J Epidemiol 2001;154: Nakano Y, Wong JC, de Jong PA, Buzatu L, Nagao T, Coxson HO, Elliott WM, Hogg JC, Paré PD. The prediction of small airway dimensions using computed tomography. Am J Respir Crit Care Med 2005;171: Tiddens HA, Paré PD, Hogg JC, Hop WC, Lambert R, de Jongste JC. Cartilaginous airway dimensions and airflow obstruction in human lungs. Am J Respir Crit Care Med 1995;152: Agusti A, Calverley PM, Celli B, Coxson HO, Edwards LD, Lomas DA, MacNee W, Miller BE, Rennard S, Silverman EK, et al.; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) investigators. Characterisation of COPD heterogeneity in the ECLIPSE cohort. Respir Res 2010;11: AnnalsATS Volume 12 Number 6 June 2015