Osteoporosis is highly prevalent in chronic obstructive pulmonary disease (COPD) patients and has been related to several clinical features. However, most studies have been in relatively small COPD cohorts. Therefore, the objectives of this study were to compare bone attenuation measured on low-dose chest computed tomography (CT) between COPD subjects and smoker and nonsmoker controls, and to relate bone attenuation to clinical parameters, inflammatory biomarkers, and outcomes in a large, well-characterized COPD cohort. We studied 1634 COPD subjects, 259 smoker controls, and 186 nonsmoker controls who participated in a large longitudinal study (ECLIPSE). We measured bone attenuation, extent of emphysema, and coronary artery calcification (Agatston score) on baseline CT scans, and clinical parameters, inflammatory biomarkers, and outcomes. Bone attenuation was lower in COPD subjects compared with smoker and nonsmoker controls (164.9 ± 49.5 Hounsfield units [HU] versus 183.8 ± 46.1 HU versus 212.1 ± 54.4 HU, p < 0.001). Bone attenuation was not significantly different between COPD subjects and smoker controls after adjustment for age, sex, and pack-years of smoking. In the COPD subjects, bone attenuation correlated positively with forced expiratory volume in 1 second (FEV1, r = 0.062, p = 0.014), FEV1/forced vital capacity (FVC) ratio (r = 0.102, p < 0.001), body mass index (r = 0.243, p < 0.001), fat-free mass index (FFMI, r = 0.265, p < 0.001), and C-reactive protein (r = 0.104, p < 0.001), and correlated negatively with extent of emphysema (r = −0.090, p < 0.001), Agatston score (r = −0.177, p < 0.001), and interleukin-8 (r = −0.054, p = 0.035). In a multiple regression model, older age, lower FFMI and higher Agatston score were associated with lower bone attenuation. Lower bone attenuation was associated with higher exacerbation (r = −0.057, p = 0.022) and hospitalization (r = −0.078, p = 0.002) rates but was not associated with all-cause mortality. In conclusion, CT-measured bone attenuation was lower in COPD subjects compared with nonsmoker controls but not compared with smoker controls, after adjustment for age, sex, and pack-years of smoking. In the COPD subjects, bone attenuation was associated with age, body composition, and coronary artery calcification but was not associated with all-cause mortality.
Chronic obstructive pulmonary disease (COPD) has, in addition to its pulmonary effects, several extrapulmonary effects that may contribute to the severity of the disease.[1-5] Osteoporosis, which is characterized by low bone mineral density and micro-architectural changes, is recognized as one such extrapulmonary effect. The prevalence of osteoporosis has been suggested to be higher in COPD patients than in control subjects matched for age and sex. However, as described in a meta-analysis, most studies on osteoporosis have been performed in relatively small cohorts.
In COPD patients, osteoporosis is associated with older age, lower body mass index (BMI), more severe airflow limitation, and the use of oral and inhaled corticosteroids.[8-11] More recently, low bone density has been associated with emphysema[12-14] and with cardiovascular disease in patients with chronic kidney disease. Although both osteoporosis and cardiovascular disease are common in COPD patients,[11, 16] research on a relationship between bone health and cardiovascular disease in COPD is sparse.
The mechanism linking osteoporosis with COPD is not entirely clear. Osteoporosis has been suggested to be either a comorbidity, owing to shared risk factors (eg, older age and smoking), or a systemic effect of COPD with a cause-and-effect relationship. Several pathophysiologic mechanisms linking osteoporosis with COPD have been suggested, including systemic inflammation, disturbance of the osteoprotegerin (OPG)/receptor activator of NF-κB (RANK)/RANK ligand (RANKL) pathway, and vitamin D deficiency.
The Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE; study identification number SCO104960; ClinicalTrials.gov identifier NCT00292552) is a longitudinal study in a large cohort of COPD subjects and smoker and nonsmoker controls. As part of this study, a low-dose chest computed tomography (CT) scan was performed from which measurements of bone density of the thoracic vertebrae can be made, in addition to the extent of emphysema and coronary artery calcification. The objectives of this study were: 1) to compare CT-measured bone attenuation of the thoracic vertebrae between COPD subjects and smoker and nonsmoker controls, and 2) to relate CT-measured bone attenuation to clinical parameters, inflammatory biomarkers, and outcomes in a large, well-characterized COPD cohort.
Materials and Methods
Our analysis was based on the data collected in the ECLIPSE study. The design and aims of the study have been published previously. The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. All subjects provided written informed consent, and the study was approved by the relevant ethics and review boards.
COPD subjects who were between the ages of 40 and 75 years were enrolled in the study if they had a history of 10 or more pack-years of smoking, a forced expiratory volume in 1 second (FEV1) of less than 80% of the predicted value after bronchodilator use, and a ratio of FEV1 to forced vital capacity (FVC) of 0.7 or less after bronchodilator use. In addition, smoker and nonsmoker control subjects who were between the ages of 40 and 75 years were enrolled if they had an FEV1 of more than 85% of the predicted value after bronchodilator use and an FEV1/FVC ratio of more than 0.7 after bronchodilator use. Smoker controls had a history of 10 or more pack-years of smoking, and nonsmoker controls had a history of less than 1 pack-year of smoking. Subjects were excluded if they had a respiratory disease other than COPD; a significant inflammatory disease (eg, rheumatoid arthritis and lupus); cancer in the 5 years before study entry; evidence of alcohol, drug, or solvent abuse; had undergone lung surgery; or used oral corticosteroids chronically.
At baseline, subjects underwent standard spirometry after the administration of 400 µg of inhaled albuterol/salbutamol. The subjects' self-reported respiratory symptoms, medications, smoking history, occupational exposure, and coexisting medical conditions were documented at study entry with the use of the American Thoracic Society-Division of Lung Disease questionnaire. Nutritional status was assessed by the BMI and fat-free mass index (FFMI), the latter measured using bioelectrical impedance analysis (Bodystat 1500, Bodystat Ltd., Isle of Man, UK). In COPD subjects, the 6-minute walk was performed according to the American Thoracic Society guideline.
Bone attenuation, extent of emphysema, and coronary artery calcification were assessed on baseline low-dose CT scans of the chest acquired using multi-detector-row CT scanners (Siemens Healthcare, Erlangen, Germany or GE Healthcare, Milwaukee, WI, USA) with a minimum of four rows that have been shown to produce comparable density values when using similar protocols. Imaging was performed in the supine position, at suspended full inspiration, without administration of intravenous contrast. Exposure settings were 120 kVp and 40 mAs, and images were reconstructed using 1.0-mm (Siemens) or 1.25-mm (GE) contiguous slices and a low spatial frequency reconstruction algorithm (GE: Standard; Siemens: b35f). CT scanners were calibrated regularly using industry and institutional standards.
Bone attenuation was measured on CT using the software VOXAR 3D version 16.0 (Toshiba Medical Visualisation Systems, Edinburgh, UK) as described previously, and has been shown to correlate with bone density measurements by DXA scanning. Briefly, the mean bone attenuation of thoracic vertebrae 4, 7, and 10 (T4, T7, and T10) were determined by placing circular regions of interest in the central parts of the vertebral bodies. The average bone attenuation of these three vertebrae was calculated and expressed in Hounsfield units (HU).
Quantitative assessment of emphysema was performed by attenuation mask analysis (Pulmonary Workstation 2.0, VIDA Diagnostics, Iowa City, IA, USA). The extent of emphysema was expressed as the percent of low attenuation areas (%LAA) with an HU of less than −950.
Coronary artery calcification was measured as the Agatston score on a dedicated postprocessing workstation (Vitrea Fx, version 3.1.0, Vital Images, Minnetonka, MN, USA) using calcium score analysis software (VScore, Vital Images). The Agatston score was calculated as previously described. The calcium score percentile based on age, sex, and ethnicity was calculated using previously published distributions from a cohort of healthy, asymptomatic individuals from the Multi-Ethnic Study of Atherosclerosis (MESA).
Serum and plasma samples for biomarker measurements were collected in the morning, after fasting overnight, at baseline. Blood samples were stored at −80°C until they were analyzed centrally. Circulating white blood cell (WBC) count was measured in a central clinical laboratory. Interleukin-6 (IL-6), interleukin-8 (IL-8), and tumor necrosis factor-α (TNF-α) were measured in serum samples using validated immunoassays on the SearchLight Protein Array Platform (Aushon Biosystems, Inc., Billerica, MA, USA). Fibrinogen (K-ASSAY fibrinogen test, Kamiya Biomedical Co., Seattle, WA, USA) and C-reactive protein (CRP) (Roche Diagnostics, Mannheim, Germany) were measured using immunoturbidometric assays validated for use with EDTA plasma. Biomarker values >95th percentile for healthy nonsmokers were considered to be elevated. Systemic inflammation was defined as two or more elevated biomarker levels (WBC, CRP, IL-6, and fibrinogen) at study entry as previously described.
All subjects were contacted monthly by telephone to assess the frequency of exacerbations and hospitalizations using a structured interview scheme. Rate of decline in FEV1 was determined as previously described. Survival status was recorded for all subjects at the 3-year visit date. In the case of subjects who withdrew from the study early, their survival status was recorded at the next planned visit and again at the 3-year time point. Date of death or last confirmed alive date were reported.
Results are shown as mean (SD or SE), median (interquartile range [IQR]), or frequency (percentage), as appropriate. Comparisons between subject groups were carried out using analysis of variance (ANOVA), Kruskal-Wallis tests, or Cochran-Mantel-Haenszel tests, and the characteristics of each COPD subpopulation were compared with the full ECLIPSE population. Univariate associations between bone attenuation and clinical factors were assessed by Spearman correlation coefficients. Multiple regression models to estimate bone attenuation were constructed based on predictors identified in the literature:[6, 8, 11, 29] age, sex, pack-years of smoking, FEV1, FEV1/FVC, FFMI, %LAA, CRP, IL-8, and Agatston score. A Cox proportional hazards model was used to determine the effect of bone attenuation on all-cause mortality. Two-sided p values <0.05 were considered significant with no adjustment for multiple testing. Analyses were conducted with SAS Version 9.1 (SAS Institute, Cary, NC, USA).
In total, 2164 COPD subjects, 337 smoker controls, and 245 nonsmoker controls were enrolled in the ECLIPSE study. Six hundred fifty subjects were excluded because their scans were either not available or the scans had not been adequately performed to assess bone attenuation. Seventeen subjects who used oral corticosteroids were also excluded because we did not wish to assess the bone density in these patients with this risk factor for osteoporosis. Thus, our cohort comprised 1634 COPD subjects, 259 smoker controls, and 186 nonsmoker controls (Table 1). Our study cohort was not significantly different compared with the ECLIPSE cohort as a whole (Supplemental Table S1). The Agatston score was measured in a subpopulation of 672 COPD subjects, representative of the total COPD cohort of the ECLIPSE study (Supplemental Table S1).
|COPD subjects||Smoker controls||Nonsmoker controls||COPD subjects versus smoker controls||COPD subjects versus nonsmoker controls||Smoker controls versus nonsmoker controls|
|No. of subjects||1634||259||186|
|Age (years)||63.3 ± 7.1||55.1 ± 8.8||53.4 ± 8.9||<0.001||<0.001||0.053|
|Male (n [%])||1045 (64%)||149 (58%)||68 (37%)||0.047||<0.001||<0.001|
|Pack-years (n)||48.8 ± 27.1||31.2 ± 22.9||0.2 ± 1.2||<0.001||<0.001||<0.001|
|FEV1 (L)||1.35 ± 0.52||3.39 ± 0.76||3.35 ± 0.78||<0.001||<0.001||0.412|
|FEV1 (% predicted)||48.5 ± 15.9||109.7 ± 11.9||116.2 ± 13.6||<0.001||<0.001||<0.001|
|FEV1/FVC ratio||44.6 ± 11.5||79.5 ± 5.0||81.4 ± 5.3||<0.001||<0.001||<0.001|
|RV/TLC ratio||0.5 ± 0.1||0.3 ± 0.1||0.4 ± 0.1||<0.001||0.001||<0.001|
|BMI (kg/m2)||26.5 ± 5.8||26.7 ± 4.3||27.4 ± 5.6||0.258||0.044||0.364|
|FFMI (kg/m2)||17.1 ± 2.9||17.1 ± 2.5||17.2 ± 2.7||0.345||0.575||0.964|
|6MWD (m)||370 ± 122||—||—|
|Inhaled steroid use (n [%])||1186 (73%)||0||0||<0.001||<0.001|
|%LAA||17.7 ± 12.3||2.3 ± 3.0||4.0 ± 4.2||<0.001||<0.001||<0.001|
|Agatston score||129.0 [506.0]||0.0 [77.0]||0.0 [3.0]||<0.001||<0.001||0.003|
|Calcium percentile||71.0 [61.0]||0.0 [78.0]||0.0 [43.0]||<0.001||<0.001||0.006|
|CRP (mg/L)||3.2 [5.7]||1.5 [2.2]||1.3 [1.9]||<0.001||<0.001||0.088|
|Fibrinogen (mg/dL)||453.0 [125.0]||384.5 [87.0]||372.0 [101.0]||<0.001||<0.001||0.067|
|IL-6 (pg/mL)||1.5 [2.2]||0.6 [1.0]||0.4 [0.6]||<0.001||<0.001||<0.001|
|IL-8 (pg/mL)||6.9 [8.8]||8.3 [12.5]||3.8 [4.6]||<0.001||<0.001||<0.001|
|TNF-α (pg/mL)||2.4 [10.4]||2.4 [41.0]||2.4 [0.0]||<0.001||0.001||<0.001|
|WBC (×106/mL)||7.6 [2.6]||7.1 [2.4]||5.7 [2.0]||0.004||<0.001||<0.001|
CT-measured bone attenuation was lower in COPD subjects compared with smoker and nonsmoker controls (164.9 ± 49.5 HU versus 183.8 ± 46.1 HU versus 212.1 ± 54.4 HU respectively, p < 0.001). However, the difference in bone attenuation between COPD subjects and smoker controls was not statistically significant after adjustment for age, sex, and pack-years of smoking (p = 0.206, Fig. 1).
In the COPD subjects, CT-measured bone attenuation correlated positively with FEV1, FEV1/FVC ratio, BMI, FFMI, and CRP, and correlated negatively with %LAA, Agatston score, and IL-8 after correction for age, sex, and pack-years of smoking (Table 2). Fig. 2 shows the correlation between bone attenuation and Agatston score. The significantly positive correlation between bone attenuation and CRP disappeared after adjustment for BMI (r = 0.015, p = 0.558). In a multiple regression model, older age, lower FFMI, and higher Agatston score were associated with lower CT-measured bone attenuation (Table 3).
|Spearman correlation coefficient||p Value|
|FEV1 (% predicted)||0.077||0.002|
|Inhaled steroid use||−0.004||0.874|
|Total COPD cohort||Agatston cohort|
|Estimate (SE)||p Value||Estimate (SE)||p Value|
|Age||−1.807 (0.181)||<0.001||−1.780 (0.283)||<0.001|
|Female||4.517 (3.319)||0.174||−5.901 (5.034)||0.242|
|Pack-years||0.019 (0.047)||0.692||−0.043 (0.076)||0.569|
|FEV1||−2.972 (3.554)||0.403||−6.679 (5.469)||0.223|
|FEV1/FVC||0.244 (0.173)||0.159||0.553 (0.268)||0.039|
|FFMI||4.506 (0.475)||<0.001||4.528 (0.682)||<0.001|
|%LAA||−0.043 (0.131)||0.740||0.075 (0.200)||0.706|
|CRP||0.024 (0.104)||0.819||0.098 (0.145)||0.498|
|IL-8||−0.075 (0.040)||0.062||−0.089 (0.049)||0.070|
|Agatston score||—||—||−0.010 (0.003)||<0.001|
Four hundred forty-eight COPD subjects (27%) had two or more elevated biomarker levels and 707 COPD subjects (43%) had no elevated biomarker levels at baseline. CT-measured bone attenuation was not significantly different between the patients with and without systemic inflammation (165.8 ± 49.4 HU versus 163.8 ± 49.1 HU, p = 0.241).
CT-measured bone attenuation correlated positively with the rate of decline in FEV1, and correlated negatively with exacerbation and hospitalization rates after adjustment for age, sex, and pack-years of smoking (Table 4). One hundred fifty-four (9%) COPD subjects died during 3 years of follow-up. The mean bone attenuation in the survivors was 165.9 ± 48.8 HU compared with 156.0 ± 55.5 HU in the nonsurvivors (p = 0.223). Cox proportional hazards modeling showed that bone attenuation was not associated with all-cause mortality in COPD subjects (Table 5).
|Spearman correlation coefficient||p Value|
|Decline in FEV1 (mL/year)||0.064||0.010|
|Hazard ratio estimate (95% CI)||p Value|
|Age (years)||1.063 (1.036, 1.091)||<0.001|
|Male versus female||1.128 (0.802, 1.587)||0.489|
|Bone attenuation (HU)||0.998 (0.995, 1.002)||0.282|
Our data showed that bone attenuation of thoracic vertebrae 4, 7, and 10 measured on low-dose chest CT was lower in COPD subjects compared with smoker and nonsmoker controls. In the COPD subjects, older age, lower FFMI, and higher Agatston score were independently associated with lower CT-measured bone attenuation. In addition, lower CT-measured bone attenuation was associated with lower rate of decline in FEV1 and higher exacerbation and hospitalization rates, but was not associated with all-cause mortality.
In line with previous data,[8, 17, 30] COPD subjects had lower bone attenuation compared with smoker and nonsmoker controls. However, the difference in bone attenuation between COPD subjects and smoker controls was not statistically significant after adjustment for age, sex, and pack-years of smoking. Because older age, female sex, and smoking are well-established risk factors of osteoporosis, these factors may contribute to the difference in bone density between COPD subjects and smoker controls. Our data suggest that osteoporosis is a comorbidity rather than a systemic effect of COPD.
In the COPD subjects, higher bone attenuation of thoracic vertebrae 4, 7, and 10 was associated with higher FEV1 and less severe emphysema. In line with these findings, Ohara and colleagues demonstrated that in 65 male COPD patients, the average bone density of thoracic vertebrae 4, 7, and 10 correlated positively with FEV1 (r = 0.286) and correlated negatively with percent of low attenuation areas (%LAA; r = −0.522). The higher correlation coefficients reported in their study may be because of differences in the study populations (eg, race and sex) or the methods used (such as different slice thickness, emphysema quantification in a part of the lung rather than the whole lung, and assessment of bone attenuation with phantom CT). Several other studies have demonstrated a relationship between bone health and lung function parameters.[8, 31] It has even been suggested that the skeletal and pulmonary systems share a common underlying mechanism such as impairment of Wingless tail/β-catenin signaling or disturbance of the RANK/RANKL/OPG pathway owing to systemic inflammation.
In addition to its relation with lung function parameters, bone attenuation decreased with age in the COPD subjects. A previous study showed that bone density of the thoracic vertebrae correlated negatively with age in male COPD patients, and a cross-sectional study with 554 COPD patients showed that patients aged 55 years or older had an increased risk of osteoporosis compared with their younger peers. In the general population, older age is a well-established risk factor of osteoporosis.[6, 34] In a healthy young skeleton, the rate of bone formation and matrix mineralization equals the rate of bone resorption and matrix degradation. However, during the aging process, significant amounts of bone are lost because of enhanced resorption coupled with decreased formation, resulting in osteoporosis.
Bone attenuation correlated positively with BMI and FFMI. This finding is supported by a study in which overweight and obese COPD patients had a decreased risk of osteoporosis, whereas cachectic COPD patients had an increased risk of osteoporosis compared with their normal-weight peers. Low body weight, BMI, and FFMI are well-known risk factors of osteoporosis in COPD patients[8, 36] and in the general population.[6, 34] Low body weight may lead to bone loss because of decreased mechanical loading or direct effects of loss of fat-free mass or fat mass on bone metabolism. Loss of fat-free mass may result in bone loss owing to increased systemic inflammation and protein breakdown, and loss of fat mass may result in reduced bone formation owing to decreased secretion of bone active hormones from the pancreatic beta cells (eg, insulin and amylin) and from adipocytes (eg, leptin).
Moreover, bone attenuation correlated negatively with coronary artery calcification. Although the relationship between bone density and cardiovascular disease has been studied widely in non-COPD populations,[15, 38-40] only one previous study showed a relationship between low bone density and increased arterial stiffness in COPD patients. The relationship between the skeletal and cardiovascular systems may be because of common risk factors such as older age and smoking, or a shared underlying mechanism such as systemic inflammation, disturbance of the RANK/RANKL/OPG system, and reduced bone perfusion resulting from generalized atherosclerosis.
In our study, bone attenuation was similar among the COPD subjects with and without systemic inflammation. Notably, bone attenuation was negatively correlated with IL-8 but positively correlated with CRP. The positive correlation between bone attenuation and CRP disappeared after adjustment for BMI, suggesting an interaction with BMI. A previous study in COPD patients indeed showed that an obese BMI was associated with elevated CRP levels.
Although a meta-analysis demonstrated that the use of inhaled corticosteroids was associated with increased risk of fractures in COPD patients, our data did not show a relationship between bone attenuation and the use of inhaled corticosteroids. However, the lack of a relationship may be the result of the design of our study because we did not record cumulative doses of inhaled corticosteroids.
Furthermore, no relationship was found between bone attenuation and the 6-minute walking distance. In line with this finding, a cross-sectional study with 95 COPD patients found no relationship between the 6-minute walking distance or the International Physical Activity questionnaire total activity score and the T-score of the femoral neck or the lumbar spine. Although research demonstrating a relationship between physical activity and bone density is sparse in COPD patients, reductions in ground reaction forces, weight-bearing activities, and muscular contraction are suggested to result in disuse osteoporosis, and physical exercise has shown to have a positive effect on bone density in postmenopausal women.
Regarding clinical outcomes, CT-measured bone attenuation correlated positively with rate of decline in FEV1 and correlated negatively with exacerbation and hospitalization rates, but was not associated with all-cause mortality. The positive correlation between bone attenuation and rate of decline in FEV1 may be because of a higher rate of decline in FEV1 in the subjects with a higher FEV1 at baseline. Previous data have indeed demonstrated that the rate of decline in FEV1 is higher in subjects with less severe COPD. Although bone attenuation was correlated with exacerbation and hospitalization rates, it was not associated with all-cause mortality. In the general population, data on the relationship between bone mineral density and mortality are conflicting.[47, 48]
This study has several limitations. First, bone health was assessed using bone attenuation on CT rather than bone density on dual-energy X-ray absorptiometry (DXA), which is the gold standard to diagnose osteoporosis. However, CT-measured bone attenuation has been shown to be strongly correlated with DXA-scanned bone mineral density in COPD patients. Second, vertebral fractures were not assessed, although vertebral fractures can occur in COPD patients with normal bone density. Third, bone attenuation was only assessed on the baseline CT scans. Because rate of bone loss has been associated with mortality in elderly men and women, it would be interesting to correlate changes in bone attenuation to clinical parameters and outcomes in COPD patients. Acknowledging these limitations, this study is important for several reasons. First, we compared bone attenuation in a large cohort of COPD subjects, smoker and nonsmoker controls, while previous studies reported comparisons of bone density in relatively small numbers of COPD subjects and control subjects.[17, 30] Second, our analyses were adjusted for age, sex and pack-years of smoking, whereas previous data were not adjusted for these confounders.[17, 30] Third, our study contributes to the insight in the systemic complexity of COPD because it showed, as far as we know, for the first time a relationship between bone attenuation and coronary artery calcification in COPD subjects. Fourth, our study demonstrated that CT-measured bone attenuation was associated with clinical parameters in COPD subjects. These data suggest, in addition to our previous data, that CT-measured bone attenuation can be used to assess bone health in COPD subjects.
Our data showed that bone density is lower in COPD subjects compared with control subjects. Chest physicians should be aware of the risk of poor bone health in their COPD patients. Because bone attenuation was independently associated with age, body composition, and coronary artery calcification, these factors may contribute to or protect against poor bone health in COPD patients. However, longitudinal studies are needed to study the factors that contribute to osteoporosis or are protective against osteoporosis in COPD and to study their underlying mechanisms.
In conclusion, CT-measured bone attenuation of the thoracic vertebrae was lower in COPD subjects compared with smoker and nonsmoker controls. However, bone attenuation was not significantly different between COPD subjects and smoker controls after adjustment for age, sex, and pack-years of smoking, suggesting osteoporosis may be a comorbidity rather than a systemic effect of COPD. In the COPD subjects, CT-measured bone attenuation was independently associated with age, body composition, and coronary artery calcification but not with all-cause mortality.
The ECLIPSE study was supported by GlaxoSmithKline. LE is an employee of GlaxoSmithKline and owns stock in GlaxoSmithKline. EAR was the recipient of a European Respiratory Society Fellowship (STRTF 381-2011). The work described in this article was performed at the University of Edinburgh as part of this Fellowship.
The authors thank Drs. Nestor Müller and Paola Nasute Fauerbach for their radiological expertise with the assessment of emphysema, and Dr. Harvey Coxson, Tara Candido, Sebastian Cogswell, Heather Davis, Nima Farzaneh, Lukas Holy, Natasha Krowchuk, Helena Lee, Evan Phillips, Claudine Storness-Bliss, Nerissa Tai, Anh-Toan Tran, Nghia Tran, Eugene Wang, and Tomonori Yokogawa for technical assistance with the CT analysis and data management.
The authors acknowledge the principal investigators, steering committee, and scientific committee of the ECLIPSE study:
Principle investigators: Bulgaria: Yavor Ivanov (Pleven), Kosta Kostov (Sofia); Canada: Jean Bourbeau (Montreal), Mark Fitzgerald (Vancouver, BC), Paul Hernandez (Halifax, NS), Kieran Killian (Hamilton, ON), Robert Levy (Vancouver, BC), Francois Maltais (Montreal, QC), Denis O'Donnell (Kingston, ON); Czech Republic: Jan Krepelka (Praha); Denmark: JØrgen Vestbo (Hvidovre); Netherlands: Emiel Wouters (Horn); New Zealand: Dean Quinn (Wellington); Norway: Per Bakke (Bergen); Slovenia: Mitja Kosnik (Golnik); Spain: Alvar Agusti, Jaume Sauleda (Palma de Mallorca); Ukraine: Yuri Feschenko (Kiev), Vladamir Gavrisyuk (Kiev), Lyudmila Yashina (Kiev), Nadezhda Monogarova (Donetsk); United Kingdom: Peter Calverley (Liverpool), David Lomas (Cambridge), William MacNee (Edinburgh), David Singh (Manchester), Jadwiga Wedzicha (London); United States of America: Antonio Anzueto (San Antonio, TX), Sidney Braman (Providence, RI), Richard Casaburi (Torrance, CA), Bart Celli (Boston, MA), Glenn Giessel (Richmond, VA), Mark Gotfried (Phoenix, AZ), Gary Greenwald (Rancho Mirage, CA), Nicola Hanania (Houston, TX), Don Mahler (Lebanon, NH), Barry Make (Denver, CO), Stephen Rennard (Omaha, NE), Carolyn Rochester (New Haven, CT), Paul Scanlon (Rochester, MN), Dan Schuller (Omaha, NE), Frank Sciurba (Pittsburgh, PA), Amir Sharafkhaneh (Houston, TX), Thomas Siler (St. Charles, MO), Edwin Silverman (Boston, MA), Adam Wanner (Miami, FL), Robert Wise (Baltimore, MD), Richard ZuWallack (Hartford, CT).
Steering committee: Harvey Coxson (Canada), Lisa Edwards (GlaxoSmithKline, USA), Katharine Knobil (co-chair, GlaxoSmithKline, UK), David Lomas (UK), William MacNee (UK), Edwin Silverman (USA), Ruth Tal-Singer (GlaxoSmithKline, USA), Jørgen Vestbo (co-chair, Denmark), Julie Yates (GlaxoSmithKline, USA).
Scientific committee: Alvar Agusti (Spain), Peter Calverley (UK), Bartolome Celli (USA), Courtney Crim (GlaxoSmithKline, USA), Gerry Hagan (GlaxoSmithKline, UK), William MacNee (chair, UK), Bruce Miller (GlaxoSmithKline, USA), Stephen Rennard (USA), Ruth Tal-Singer (GlaxoSmithKline, USA), Emiel Wouters (The Netherlands), Julie Yates (GlaxoSmithKline, USA).
Authors' roles: Study design: EAR, EvB, JM, EW, and WM. Study conduct: EAR, JM, EPR, and WM. Data collection: EAR, LE, DM, and MW. Data analysis: EAR, LE, EPR, and WM. Data interpretation: EAR, JM, EPR, FS, EW, and WM. Drafting manuscript: EAR, EPR, and WM. Revising manuscript content: EAR, LE, EvB, DM, JM, EPR, FS, MW, EW, and WM. Approving final version of manuscript: EAR, LE, EvB, DM, JM, EPR, FS, MW, EW, and WM. EAR and LE take responsibility for the integrity of the data analysis.