2011 ATS Abstracts

 

Statin use in COPDGene subjects is associated with larger airways and fewer exacerbations

 

P. A. Bercz MD¹, V. Kim MD¹, N. Marchetti MD¹, E. Regan MD², G.B. Vance RN¹, B.J. Make MD², M.K. Han MD³, C.P. Hersh MD, MPH⁴, J.D. Newell MD², D. Stinson², E.K. Silverman MD⁴, PhD, G.J. Criner MD¹, and the COPDGene Investigators

¹Temple University Hospital Pulmonary & Critical Care Medicine - Philadelphia, PA/US, ²National Jewish Health, Denver, CO/US, ³University of Michigan School of Medicine, Ann Arbor, MI, ⁴Brigham and Women’s Hospital, Boston, MA/US

Introduction: Statin use may be beneficial in COPD via their pleiotropic anti-inflammatory effects and have been associated with reduced exacerbation frequency, rate of lung function decline, and mortality in large retrospective reviews. By reducing inflammation, statins might improve airway dimensions in COPD patients. To our knowledge, there has been no large multicenter study investigating the effect of statins on radiographic measurements of airway dimensions of COPD patients.

Methods: Enrollees in COPDGene with spirometrically-proven COPD (GOLD I-IV, n=1085) underwent standardized high-resolution CT scanning with measurement of airway dimensions from the major airways (mainstem and lobar bronchi) to the segmental, subsegmental, subsegmental + 1 bronchi of representative lung segments (RB1, RB4, RB10, LB1, LB4, LB10). Measurements included lumen area on inspiration and expiration, mean inner diameter, wall area percent, and average wall thickness. Wall area percent and average wall thickness measurements in the subsegmental +1 airways were excluded due to possible imprecision. Averages of measurements for each separate airway generation were compared between subjects on statins and those not on statins. Comparisons were made using Student’s t and Fisher exact tests. Multiple linear and logistic regression analysis was performed to control for age, race, BMI, and clinical center.

Results: There were 1,085 subjects (292 on statins and 793 not on statins) with COPD that were included in the analysis. Subject characteristics are presented in Table 1. Detailed results of airway dimensions are presented in Table 2. All measurement differences became more prominent in successive lower-order bronchi. Dimensions of major airways between statin and non-statin groups were similar with the exception of mean inner diameter (11.6 ± 1.43 vs. 11.3 ± 1.38, p=0.03). In segmental airways, lumen area on inspiration (19.2 ± 5.93 vs. 18.4 ± 5.08, p=0.03) and mean inner diameter (4.87 ± 0.73 vs. 4.77 ± 0.65, p=0.027) were significantly greater in statin users compared to non-statin users. In sub-segmental and sub-segmental +1 airways, nearly all measurement differences were statistically significant with the sole exception of wall area percent in subsegmental airways (65.2 ± 2.52 vs. 65.5 ± 2.46, p=0.08). Statin use was associated with fewer exacerbations in the 12-month period prior to enrollment (0.40 ± 0.94 vs. 0.56 ± 1.14, p=0.03).

Conclusion: Statin use in COPD is associated with improved inspiratory airway luminal area in segmental and lower-order airways. Statin use was also associated with a reduced exacerbation frequency.

 

 

The Association of Genome-Wide Significant Spirometric Loci with COPD Susceptibility

Peter J Castaldi^1,2, Michael H Cho^2,3, Augusto A Litonjua^2,3, Per Bakke⁴, Amund Gulsvik⁴, David A Lomas⁵, Wayne Anderson⁶, Terri H Beaty⁷, John E Hokanson⁸, James D Crapo⁹, Nan Laird¹⁰, Edwin K Silverman^2,3 for the COPDGene and Eclipse Investigators

¹Institute for Clinical Research and Health Policy Studies, Tufts Medical Center, Boston, MA, USA; ²Channing Laboratory and ³Division of Pulmonary and Critical Care Medicine, Brigham and Women's Hospital, and Harvard Medical School, Boston, MA, USA, ⁴Haukeland University Hospital, Bergen, Norway; ⁵Department of Medicine, University of Cambridge, Cambridge Institute for Medical Research, Cambridge, UK; ⁶Research Triangle Park, NC 27709, USA; ⁷Johns Hopkins School of Public Health, Baltimore, MD, USA; ⁸Colorado School of Public Health, University of Colorado Denver, Aurora, CO, USA; ⁹National Jewish Medical and Research Center, Denver, CO, USA; ¹⁰Harvard School of Public Health, Harvard University, Boston, MA, USA

Rationale: Two recent meta-analyses of genome-wide association studies conducted by the CHARGE and SpiroMeta consortia identified novel loci yielding evidence of association at genome-wide significance levels in an analysis of FEV₁ and FEV₁/FVC.  We hypothesized a subset of these genomewide significant spirometric (GWS spirometric) markers would also be associated with COPD susceptibility when tested in four COPD case-control study samples.

Objectives: To determine whether eleven previously identified GWS spirometric genes/regions are associated with COPD susceptibility in multiple independent COPD case-control study samples. Twenty-one previously identified COPD candidate genes were also tested for comparison.

Methods: Single nucleotide polymorphisms (SNP) in eleven previously identified GWS spirometric genes/regions were tested for association with COPD status in four independent COPD case-control study samples (NETT/NAS, the Norway case-control study, ECLIPSE, and the first 1,000 subjects in COPDGene - total sample size 3,459 cases/1911 controls). Association analyses were conducted using logistic regression adjusting for age and principal components of genetic ancestry. Adjustment for multiple comparisons was performed via permutation methods. Additional testing from a dense panel of imputed HapMap2 SNP markers was performed for these eleven genes/regions, as well as for twenty-one well-studied COPD candidate genes.

Measurements and Main Results: Of the eleven previously identified GWS spirometric genes/regions, four loci harbored SNPs associated with COPD susceptibility at nominal levels of significance (p<0.05): the 4q24 locus including FLJ20184/INTS12/GSTCD/NPNT, the 6p21 locus including AGER and PPT2, the 5q33 locus including ADAM19, and the 2q33 locus including PID1. After adjusting for multiple comparisons, only the 4q24 locus remained significant. In comparison to well-studied candidate genes, markers in previously identified GWS spirometric genes/regions were more strongly associated with COPD susceptibility.

Conclusions: Targeted association testing of markers in previously identified GWS spirometric genes/regions in data from 4 COPD studies identified a region on 4q24 (nearby genes FLJ20184/INTS12/GSTCD/NPNT) that is associated with COPD susceptibility. Our findings also suggest that markers showing strong association with quantitative measures of pulmonary function are good candidate genes for association with COPD susceptibility.

Support: This work was supported by U.S. National Institutes of Health (NIH) grants R01 HL075478, R01 HL084323, P01 HL083069, U01 HL089856 (Silverman), K08HL102265 (Castaldi), UL1 RR025752,  and U01 HL089897 (Crapo). The National Emphysema Treatment Trial was supported by the National Heart, Lung, and Blood Institute, the Centers for Medicare and Medicaid Services, and the Agency for Healthcare Research and Quality. The Normative Aging Study is supported by the Cooperative Studies Program/ERIC of the U.S. Department of Veterans Affairs, and is a component of the Massachusetts Veterans Epidemiology Research and Information Center (MAVERIC). The Norway cohort and the ECLIPSE study (Clinicaltrials.gov identifier NCT00292552; GSK Code SCO104960) are funded by GlaxoSmithKline.

 

 

Identification of Factors that Motivate Volunteers to Participate in Research Studies

A. Chapman^1, K. Sprenger¹, S. Kraus¹, A. Delsing¹, J. Keating¹, P. Pirotte¹, K. Crawford¹, D. C. Look¹

University of Iowa, Iowa City¹

Rationale:  Identifying why individuals enroll in pulmonary research studies aids in the recruitment and retention of study subjects.

Methods: The study protocol was approved by the University of Iowa Institutional Review Board. Two hundred and seventy four participants of the COPDGene Study at the University of Iowa were given a questionnaire with four choices.  Participants were asked to choose one answer that would best reflect their reason for participation.

Q. A) "I choose to take part in the study because my participation could help someone else some day"

Q. B) "I choose to take part in the study because I wanted to have the diagnostic testing done that I knew was part of this study (ie CT scan)"

Q. C) "I choose to take part in the study because I knew that there was compensation of $75 for completing the study visit"

Q. D)" I find medical research interesting"

Results:  The demographics of the participants surveyed resulted in 55% male, 45 % female. Residential demographics showed 69% lived with in a 50mile radius of the Clinical Research site, 31% lived greater then 50miles from the research site and 7% lived greater then 100miles from the research site. Participant age range showed 56% were between 45 -64 years old, 44% were between 65-80 yrs old. Of the 274 questionnaires collected 55% choose option A, 31% choose option B, and 7% choose option C and D.

Conclusions: Study participants take part in studies for multiple reasons. The data collected demonstrated that the majority of participants volunteered for this research study because they hoped their participation could benefit others in the future. The motivation to receive diagnostic testing was the second largest group. These two responses could not be attributed to the subject's age, proximity to site, education, and socioeconomic status. We conclude that recruitment targeted to those who view research as a humanitarian endeavor could improve study participation.

Funding for this research was provided by NIH COPDGene study, grant numbers U01 HL089897 and U01 HL089856.

 

 

A Genome-wide Association Study in COPD Identifies a Susceptibility Locus on Chromosome 19q13

Cho MH, Castaldi PJ, Hersh CP, Demeo DL, Sylvia JS, Klanderman BJ, Ziniti JP, Lange C, Litonjua AA, Sparrow D, Murphy JR, Regan EA, Make BJ, Hokanson JE, Murray T, Hetmanski JB, Pillai S, Kong X, Anderson WH, Tal-Singer R, Lomas D, Bakke P, Gulsvik A, Crapo JD, Beaty TH, Silverman EK, on behalf of the ICGN, ECLIPSE, and COPDGene Investigators

Rationale: The genetic risk factors for chronic obstructive pulmonary disease are still largely unknown.  Modest-sized genome-wide association studies have to date identified several novel risk loci for COPD in CHRNA3/CHRNA5/IREB2, HHIP, and FAM13A; further loci may be identified through larger studies.

Methods: We performed a genome-wide association study using subjects from the Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE); Normative Aging Study (NAS); National Emphysema Treatment Trial (NETT); Bergen, Norway (GenKOLS); and the first phase of subjects from the COPDGene study.  Genotyping was performed on Illumina platforms (Human Hap550, Quad610, and Omni1), with additional markers imputed using MACH using the HapMap2 CEU reference population.  Logistic regression models were used to adjust for age, pack-years of cigarette smoking, and ancestry via principal components, and results were summarized using fixed-effects meta-analysis.

Results: After quality control, a total of 3459 cases and 1911 smoking controls (372, 853, 1736, and 498 cases, and 492, 805, 176, and 501 controls, from NETT/NAS, GenKOLS, ECLIPSE, and COPDGene respectively) were included.  296,201 markers were successfully across all cohorts, and > 2 million markers were imputed.  We identified a new genome-wide significant locus on chromosome 19q13 (top genotyped SNP rs7937, OR=0.74, P value = 4 x 10-9).  Further refining the locus using data from the 1000 Genomes Project identified a marginally association (with imputed marker rs2604894, OR=0.74, P 3.2 x 10-9).  These two SNPs were genotyped in 2859 subjects in nuclear families from the International COPD Genetics Network study and tested for association with COPD affection status using PBAT.  P values for association with COPD for rs7937 and rs2604894 were 0.28 and 0.11 respectively; the significance of the association increased for pre-bronchodilator FEV1 (P=0.08 and 0.04) and when the analysis was limited to severe (GOLD 3&4) COPD (P=0.09 and 0.017). 

Conclusions: We identified a locus for COPD susceptibility on chromosome 19q13.  This region includes genes RAB4B, EGLN2, and CYP2A6, and this region has previously been identified in association with cigarette smoking behavior.  Further investigation will be required to determine if this region confers COPD susceptibility through nicotine addiction or some other mechanism.

Funding: R01 HL075478, R01 HL084323, P01 HL083069, U01 HL089856 (E.K.S.), K12HL089990 and K08 HL097029 (M.H.C), U01 HL089897 (J.D.C.). NETT was supported by the NHLBI and CMS; NAS was supported by MAVERIC. The Norway cohort, the ECLIPSE study (Clinicaltrials.gov identifier NCT00292552; GSK Code SCO104960) and the ICGN study are funded by GlaxoSmithKline.

 

 

Genome-wide DNA Methylation as a Peripheral Biomarker of Emphysema

D. DeMeo¹, W. Qiu², H. Bacherman¹, G.R. Washko³, B.J. Make⁴, J. Crapo⁴, E.K. Silverman⁵

¹Brigham and Women's Hospital, Channing Laboratory - Boston, MA/US, ²Brigham and Women's Hospital - Boston, MA/US, ³Brigham & Womens Hosp Div of Pulmonary and Critical Care Medicine - Boston, MA/US, ⁴National Jewish Health - Denver, CO/US, ⁵Brigham and Women's Hospital - Boston/US

Background: Epigenetic variation may impact intermediate phenotypes of complex human disease that are not fully explained by DNA sequence variation. Age-of-onset of COPD and variable susceptibility to tobacco smoke are characteristics which may be influenced by epigenetic events such as DNA methylation. We hypothesized that alterations in DNA methylation may associate with quantitative CT scan measures of emphysema in cohort of smokers with and without COPD.

Methods: COPDGene is a study focused on identifying genes influencing COPD susceptibility in non-Hispanic whites and African-Americans (AA). This report analyzed epidemiologic and DNA methylation data from 79 subjects with complete verified data. Genomewide DNA methylation data were available for buccal and blood DNA samples. We used the Illumina HumanMethylation27 Beadchip to investigate genome-wide methylation patterns at 27,578 CpG sites for each subject. We used the ratio of the fluorescence signals from methylated and unmethylated DNA to quantify methylation level. We excluded the sex chromosomes from analysis. We focused on CT phenotypes of percent emphysema as quantified by the program SLICER. For our analysis, we used a binary outcome of percent emphysema greater than or less than 20%. All models were adjusted for current and pack-years of smoking.

Results: DNA methylation and CT scan data were available on 79 subjects. 62% of the subjects were women. The mean age of subjects was 64 years with a mean of 44 pack-years of cigarette smoking; 28% were current smokers. These subjects represented a range of GOLD stages (controls=27 subjects, GOLD 1=4 subjects, GOLD=2,3,4=48 subjects) and a range of CT percent emphysema (0-55%). 638 CpG sites were differentially methylated in buccal DNA in the presence of a percent emphysema greater than 20(P<0.05).The top associated CpG sites were in ANXA9 (chr 1) and GNAS (chr20) (P<0.005). 362 CpG sites were differentially methylated in blood source DNA; the top associated CpG sites were in DARC

(chr 1) and SFTPB (chr 2) (P<0.002). 13 genes had CpG sites associated with higher percent emphysema (P<0.05) that were recapitulated in buccal and blood source DNA. These genes included ABCG1 (chr 21), which has high gene expression in the mammalian lung.

Conclusions: These observations support the hypothesis that DNA methylation in the buccal and blood compartments may be a peripheral biomarker of CT defined emphysema. These data support the importance of integration of gene expression and DNA methylation marks in defining biomarkers that impact intermediate phenotypes of COPD.

 

 

Airway Dimensions by Computed Tomography in COPD Patients with Bronchodilator Reversibility

P. Desai¹, P.A. Bercz¹, M.K. Han², G.B. Vance¹, B.J. Make³, C.P. Hersh⁴, J. Newell³, D. Stinson³, E.K. Silverman⁴, G.J. Criner¹, V. Kim¹, &. COPDGene Investigators⁵

¹Temple University Hospital - Philadelphia, ²University of Michigan - Ann Arbor, ³National Jewish Health - Denver, ⁴Brigham and Women's Hospital- Boston, ⁵COPDGene Study - Various/US

RATIONALE: Bronchodilator reversibility (BDR) is a common but variable phenomenon in patients with COPD. In COPD patients, BDR has been associated with small airway smooth muscle mass and epithelial thickness. Hence, BDR is thought to be a manifestation of airways disease and not of emphysema. The radiographic characteristics of airway dimensions that differentiate COPD patients with BDR from those without BDR have not been well described.

METHODS: We analyzed 904 patients with GOLD 1 to 4 disease in the COPDGene® study who had radiographic airway analysis. We divided patients into two groups based on BDR: BDR+ (post-bronchodilator change in % predicted FEV1≥10%) and BDR- (post-bronchodilator change in % predicted FEV1<10%). Quantitative image analysis to calculate percent emphysema, and percent gas-trapping was performed using VIDA software (VIDA Diagnostics, http://www.vidadiagnostics.com) and 3D SLICER (http://www.slicer.org). Airway disease was quantified in each group as wall area percent (WA%: (wall area / total bronchial area) x 100), airway wall thickness (AWT), and the inspiratory lumen area (LA). The mean WA%, AWT, and LA were calculated as the average of the values for six segmental and sub-segmental bronchi in each subject. Using 3D SLICER, airway thickness was also expressed as the square root of the wall area of a theoretical airway of 10mm lumenal perimeter (Pi10).

RESULTS: 627 patients in the BDR- group and 377 in the BDR+ group formed our analysis. Tables 1 and 2 display the results. The groups in general were similar in demographics and lung function, except that the BDR+ group had a lower %FEV1, greater BMI, and less current smokers. The BDR+ group had a greater Pi10, greater segmental and sub-segmental airway WA%, and smaller segmental and sub-segmental airway lumen area compared to the BDR- group. There also existed significant differences in measurements of airway dimensions amongst the different study centers. In a multiple logistic and linear regression model, when adjusted for race, BMI, current smoking status, % predicted FEV1, % gas trapping, and the study centers, BDR remained an independent predictor of greater WA% and lumen area.

CONCLUSION: The presence of BDR in COPD is related to radiographic airway pathology independent of race, BMI, % predicted FEV1, current smoking status, and % gas trapping. These data give us further insight to the underlying patho-physiology of airflow obstruction in COPD.

This study was supported by the NHLBI U01 HL089856 and U01 HL08989

 

 

 

Evaluation of Airway Distensibility Using Paired Expiratory and Inspiratory Volumetric CT scans in Patients with COPD

Alejandro A. Diaz¹, M.D., James C. Ross², M.S.,  Carolyn E. Come¹, M.D., Raúl San José Estépar², Ph.D., MeiLan K. Han³, M.D., Edwin K. Silverman⁴, M.D., Ph.D., George R Washko¹, M.D.

¹Pulmonary and Critical Care Division, Brigham and Women's Hospital, Harvard Medical School, Boston MA; ²Surgical Planning Laboratory, Department of Radiology, Brigham and Women's Hospital, Boston MA; ³ Division of Pulmonary and Critical Care, University of Michigan Health Systems, Ann Arbor MI; ⁴Channing Laboratory, Brigham and Women's Hospital, Boston MA;

Background: In COPD, the ability of airways to increase in diameter with lung inflation (airway distensibility) is related to airway remodeling and loss of elastic recoil due to emphysema.

Aims: We performed a case-control study to test the hypothesis that for a given GOLD stage, airway distensibility is lower in patients with an emphysema-predominant as opposed to an airway-predominant phenotype. We also determined the number of patients who had a paradoxical decrease in airway diameter with lung inflation (airway constrictors, ACs). Finally, we compared the clinical, functional, and CT features of ACs with non-ACs.

Methods: Our cohort included 132 smokers from the COPDGene study. All subjects underwent CT scans at full inspiration (TLC) and at expiration (FRC). Whole lung emphysema was defined as the percentage of low attenuation area (< -950 HU). The 3^rd and 4^th airway generations (AG) of the right upper lobe apical segment and right lower lobe posterior segment were matched and their dimensions measured in both CT scans using Slicer software. Airway distensibility was expressed as the ratio of absolute change in airway inner diameter to the cubic root of absolute change in lung volume from FRC to TLC. ACs had a decrease or lack of increase in airway diameter from FRC to TLC at one or more AGs. COPD phenotypes were defined as airway-predominant (AP) or emphysema-predominant (EP) based on the percentage of emphysema (<13% vs. ≥25%, respectively). Patients were matched by gender and age and divided into 5 groups: controls (n=40), GOLD 2-AP (n=23), GOLD 2-EP (n=23), GOLD 4-AP (n=23), and GOLD 4-EP (n=23).

Results: Compared to controls, airway distensibility was significantly reduced at the 4^th AG in all COPD groups. For a given GOLD stage, there were no significant differences in airway distensibility between AP and EP phenotypes (Fig 1). Thirty-one out of 132 patients (23%) were ACs. Compared to non-ACs, ACs were significantly more likely to be African-American and current smokers. ACs also had a significantly higher dyspnea score, a lower FEV₁% predicted, a shorter six minute walk distance, and a higher percentage of emphysema (Table 1).

Conclusions: In this cohort, an emphysema predominant phenotype was not associated with reduced airway distensibility of medium-to-large airways. Perhaps regionalized assessments will demonstrate a significant effect of emphysema on airway distensibility. A paradoxical decrease in airway diameter was frequent and was associated with worse lung function and exercise capacity.

 

CT detected pulmonary hypertension is associated with severe COPD exacerbation in the COPDGene population

Mark T. Dransfield, Naseer Abbas, Elizabeth Westfall, MeiLan K. Han, George R. Washko, James Mamary, Russell P. Bowler, Hrudaya Nath for the COPDGene Investigators

RATIONALE: COPD exacerbations are critical events that are associated with loss of lung function, poor quality of life and both short and long term mortality.    The identification of patients at risk for exacerbations is important for clinical practice and the design of clinical trials.  Prior data suggests that pulmonary hypertension is associated with exacerbation frequency and exacerbation-related death but no study has examined the utility of CT markers of pulmonary hypertension in predicting these outcomes.  We hypothesized that CT detected pulmonary hypertension would be associated with a history of severe COPD exacerbation in the COPDGene population.

METHODS: Demographic, clinical and CT data from 790 COPDGene participants with stage II - IV COPD as defined by the Global Initiative for Obstructive Lung Disease were analyzed.  Patients reported whether they had suffered a severe exacerbation of COPD (defined as those requiring an emergency room visit or hospitalization) in the 12 months prior to enrolment.  A single CT interpreter measured the diameter of the main pulmonary artery (PA) at the level of the bifurcation and using the same image measured the diameter of the aorta (A) in its maximum dimension.  We performed univariate logistic regression to examine the associations between patient characteristics and CT metrics of pulmonary hypertension (main PA diameter > 28 mm; PA/A ratio>1) and a history of severe exacerbation.  Multivariate analyses were then performed including variables that were associated with severe exacerbation (p<0.10).

RESULTS: 790 patients were included in the study and 145 reported a severe COPD exacerbation in the previous year.  The population was 50% male, 81% Caucasian, had a mean age of 64 (range 45-81) and a mean FEV1% predicted of 56% (range 30-79).  We found that younger age, African American race, lower FEV1%, shorter 6 minute walk distance, absence of O2 therapy, reported asthma, increasing MMRC and greater bronchial wall thickness each exhibited statistically significant associations with severe exacerbations.  We also found associations between severe exacerbations and the presence of PA>28 [1.85 (1.26-2.70); p=0.002] and PA/A>1 [4.05 (2.78-5.88); p<0.001}.  On multivariate testing including all the variables listed above, PA/A>1 remained a significant predictor of severe exacerbation [2.92 (1.92-4.44); p<0.001].  In a separate multivariate model, PA>28 was not independently associated with severe exacerbation (p=0.08).

CONCLUSIONS: CT detected pulmonary hypertension (PA/A>1) is associated with a history of severe COPD exacerbation.  This has important implications for clinical practice and trial design.

Supported by grant numbers U01 HL089897 and U01 HL089856.

 

 

Airway Extraction in Inspiratory Volumetric CT using Scale-Space Particles

Raúl San José Estépar¹, Ph.D., James C. Ross², Gordon L Kildmann, Ph.D.³, Alejandro A. Diaz², M.D., M.S., Edwin K. Silverman⁴, M.D., Ph.D., George R Washko², M.D. Other COPDGene contributors.

¹Surgical Planning Laboratory, Department of Radiology, Brigham and Women's Hospital, Boston MA ; ²Pulmonary and Critical Care Division, Brigham and Women's Hospital, Harvard Medical School, Boston MA; ³ Division of Pulmonary and Critical Care, University of Chicago, Chicago IL; ⁴Channing Laboratory, Brigham and Women's Hospital, Boston MA

Background: A major limitation to large genetic and epidemiologic imaging based studies of COPD is the ability to objectively assess CT measures of airway disease in an efficient fashion. New methods of automated analysis are needed for dependable airway segmentation and analysis.

Methods: We have previously developed a method based on scale-space particles which is an approach that optimally searches for image features using unique characteristics of its second-order derivative. Airways are unique in that the present two large positive second-order derivative when moving from lumen to airway wall and a small derivatives along the longitudinal axis. After initially placing the particles in given scale locations within low attenuation structures adjacent to the pulmonary vasculature, their placement is refined in two steps. First, a constraint step moves the particles to locations of minimal local intensity along the direction of the largest variation of the second order derivative (centering the particle within the airway lumen). Second, particles interact with each other in both the spatial and scale dimensions to minimize a global energy. This latter step ensures homogenous sampling along the length of the airway. A final filtering step removes outlier particles that were placed in airway-looking locations in the parenchymal field. The CT is then reformatted orthogonal to the airway axis at each remaining particle location and the wall thickness was measured using phase congruency for 30 rays cast from the particle point. One thousand six hundred CT scans from the COPDGene study were analyzed in such an automated fashion. For each subject, Pi10 was computed as well as the mode of the WA% histogram. Univariate and multivariate linear regression was used to show the association with FEV1%.

Results: Pi10 measured by scale-space particles is associated to FEV1% (r=0.36, p<0.0001). In a multivariate model, Pi10 and percentage of emphysema remains as independent predictors of lung function (r2=0.5).

Conclusions: Scale-space particles are a feasible approach to efficiently sample airway locations in volumetric CT scans. The method is totally automatic and it can be run without supervision. Although, scale space particles do not provide topological information about the airway tree, further processing can be applied to link nearby particles and extract this information.

 

 

Alpha-1 Antitrypsin PI MZ Heterozygosity is Associated with COPD and CT Emphysema in the COPDGene Study

¹Marilyn G. Foreman, MD, MS; ²Dawn L. DeMeo, MD, MPH; ² Craig P. Hersh, MD, MPH; ²Barbara J. Klanderman, PhD; ²John Ziniti; ³James Murphy, PhD; ⁴John E. Hokanson, MPH, PhD; ⁵Terri Beaty, PhD;³James D. Crapo, MD; ²Edwin K. Silverman, MD, PhD

¹Morehouse School of Medicine, Atlanta, GA;² Channing Laboratory, Brigham and Women's Hospital, and Harvard Medical School, Boston, MA;³ National Jewish Health, Denver, CO; ⁴University of Colorado, Denver, CO; ⁵John Hopkins Bloomberg School of Public Health, Baltimore, MD

Objectives: The association between PI MZ alpha-1 antitrypsin genotype and COPD has been controversial for decades.  The Genetic Epidemiology of COPD (COPDGene) Study is a multi-center epidemiologic and genetic study of COPD that will enroll 10,000 non-Hispanic White (NHW) and African American (AA) subjects with > 10 pack-years of smoking.  We hypothesized that PI MZ would be associated with COPD affection status and with percent emphysema determined by computed tomography (CT emphysema) in these racial groups.

Methods: Among 6919 subjects enrolled to date, genotyping for the Z (rs289294740) and S (rs17580) alleles was performed with 5^′ to 3′ exonuclease assays implemented in TaqMan.  All PI ZZ and PI MZ calls were confirmed with repeat genotyping.  All non-S and non-Z subjects were considered PI MM.  Regression models were constructed to analyze COPD case status (GOLD stages 2/3/4 vs. normal spirometry) and CT emphysema as dependent variables.  CT percent emphysema was determined by Slicer software.

Results: There were 6 PI ZZ subjects; none were African American.  There were 189 NHW and 16 (8%) AA PI MZ subjects.  Compared to PI MM subjects, PI MZ were older, 63 vs. 59 years (p<0.0001) but did not differ in pack-years of smoking (p > 0.05).  Stratified by race, univariate analysis of NHW subjects revealed that PI MZ  subjects had a lower FEV₁ percent predicted, 67 + 26 vs. 73 + 36% (p < 0.001); lower FEV₁/FVC, 60 + 18 vs. 64 + 17 (p = 0.004); greater CT emphysema, 12 + 13 vs. 9 + 11% (p = 0.01); and greater risk for COPD (p = 0.058); among African Americans,  PI MZ subjects had lower FEV₁ percent predicted, 65 + 34 vs. 83 + 23% (p = 0.06); lower FEV₁/FVC, .60 + .20 vs. .72 + .13 (p = 0.03); greater CT emphysema, 13 + 15 vs. 5 + 8% (p = 0.2); and greater risk for COPD (p = 0.02). In multivariable models adjusted for age, gender, and smoking status, PI MZ alpha-1 antitrypsin type was significantly associated with COPD case status and CT percent emphysema in both White and African American subjects (p < 0.05).

Conclusion: In the COPDGene Study, we found that PI MZ heterozygosity is associated with COPD affection status and CT emphysema in both non-Hispanic White and African American subjects.  The consistency of the findings across both ethnicities suggests that PI MZ is a consistent risk factor for COPD.

 

Heterogeneity of Blood Monocyte Subsets in Chronic Obstructive Pulmonary Disease; a sub-study of COPDGene®.

C.M. Freeman^1,2, P. Attawala², J.C. Todt¹, C.H. Martinez¹, F.J. Martinez¹, M.K. Han¹, L. McCloskey¹, D.L. Thompson¹, J.L. Curtis^1,2

¹University of Michigan, Ann Arbor, ²Veterans Affairs Medical Center, Ann Arbor

Rationale: Monocytes could play an important role in Chronic Obstructive Pulmonary Disease (COPD), both by releasing inflammatory mediators and by migrating into the lungs where they differentiate into the population of lung macrophages that is expanded in all disease stages.  The existence of monocyte subsets with distinct phenotypes and functions is well established; but monocyte heterogeneity has not been studied in COPD.  In a site-specific sub-study of COPDGene® (NCT00608764), we correlated spirometric data with the frequency in peripheral blood of two major monocyte subsets, CD14++ CD15+ (classical) and CD14+ CD15- (proinflammatory).

Methods: We performed spirometry (EasyOne^TM; ndd Inc., Zurich) pre/post administration of albuterol. and collected peripheral blood from 120 consented subjects (COPD, n = 69; without COPD, n = 51) .  We made the following measurements: (a) surface expression of multiple monocyte receptors by 12-color flow cytometry; (b) serum levels of matrix metalloproteinase-9 (MMP-9), platelet derived growth factor-BB (PDGF-BB), macrophage colony-stimulating factor (M-CSF), adiponectin, soluble RANK ligand, and CCL2 by multiplex bead analysis; & (c) mRNA transcripts for chemokine and chemokine receptor in isolated monocytes (which comprise both subsets) by real-time RT-PCR.

Results: Expressed as a percentage of all CD45+ leukocytes, CD14+ CD15- blood monocytes significantly decreased as FEV₁ % predicted decreased, whereas no change was seen in the percentage of CD14++ CD15+ monocytes.  Protein analysis revealed higher serum levels in more advanced COPD of PDGF-BB and MCP-1, both of which are chemotactic and activating factors for human blood monocytes; these mediators also  correlated with lower percentages of CD14+ CD15- cells.  With increasing COPD severity, there were increases in monocyte mRNA transcripts for CCR2 and CCL2.  The latter correlation was strengthened when current smokers were analyzed by themselves and was not evident in the population of ex-smokers.

Conclusions: As COPD increases in spirometrically-defined severity, the balance between the two best-defined subsets of peripheral blood monocytes populations changes and there are increases in several factors favoring monocyte mobilization and possibly recruitment to the lungs.  The heterogeneous nature of blood monocytes must be considered when investigating the complex mechanisms involved in COPD progression.

Funding Source: Career Development Award, Biomedical Research & Development Service, Department of Veterans Affairs; R01HL082480, N01HR046162, K23HL093351, K24HL04212, U01HL089897 and U01Hl089856 from NHLBI.

 

 

Radiologic variability in airway dimensions of the normal lung

MeiLan K. Han, Phillip Westgate, Jordan Zach, Susan Murray, Fernando Martinez, George Washko, Ella Kazerooni, David Lynch

Background: The radiologic assessment of the human lung with quantitative CT is becoming increasingly common.  However, in order to understand what constitutes "disease", careful radiologic characterization of the normal human lung is first required.

Methods: 100 normal subjects from the COPDGene cohort with no smoking history and an FEV₁/FVC ratio of > 0.7 were analyzed.  All subjects underwent volumetric lung CT scanning with emphysema measurements using a -950 HU threshold, and airway morphology evaluation of the segmental airways (wall thickness, wall area percent and lumen area).  The airway measurements in each of 6 lobes (right and left upper, middle, lingual, and right and left lower lobes) were averaged to create "upper", "middle" and "lower" lobe variables.  Multivariate linear regression was used to study associations between sex, age and BMI with the measures of airway morphology.

Results: Results of the multivariate analysis are presented in Table 1 for representative men and women with low, medium and high BMI.  Variation in airway measures was seen by BMI and gender, but not age, which was therefore dropped from the model.  Wall thickness was greater for men than women, and for subjects with higher BMI.  For both men and women, segmental walls tended to be thinnest in the middle lobe/lingula.  For all men, and high BMI women, the upper lobe was thickest.  Lumen area was greater for men than women.  Lumen areas were smallest in the middle lobe/lingula and greatest in the lower lobes regardless of sex or BMI.  Wall area percent was greater for women than men, but less sensitive to BMI differences.  Lower lobes demonstrated the smallest wall area percent in all subjects.

Conclusions: We demonstrate notable radiologic variation of both airway thickness and lumen measures in normal individuals by gender, BMI and location in both airway thickness and lumen measures.  In particular, the smallest lumens and thinnest walls are found in the middle lobes, the clinical significance of which is yet to be determined.

 

 

Genetics of Sputum Gene Expression Supports Two Distinct COPD Susceptibility Genes on Chromosome 15q25

C.P. Hersh¹, W. Qiu², M.H. Cho², J. Riley³, W. Anderson⁴, D. Singh⁵, P. Bakke⁶, A. Gulsvik⁶, A.A., Litonjua⁷, D.A. Lomas⁸, J. Crapo⁹, T.H. Beaty¹⁰, B. Celli¹¹, S.I. Rennard¹², R. Tal-Singer¹³, S. Fox³, E.K. Silverman², &. ECLIPSE Investigators¹⁴, &. COPDGene Investigators¹⁵

¹Brigham and Women's Hospital / - Boston, MA/US, ²Brigham and Women's Hospital - Boston, MA/US, ³GlaxoSmithKline - Uxbridge/UK, ⁴Glaxo SmithKline - Research Triangle Park, NC/US, ⁵University of Manchester - Manchester/UK, ⁶Haukeland University Hospital - Bergen/NO, ⁷Channing Lab / Brigham and Women's Hospital - Boston, MA/US, ⁸Cambridge Institute for Medical Research, University of Cambridge - Cambridge/UK, ⁹National Jewish Health - Denver, CO/US, ¹⁰Johns Hopkins Bloomberg School of Public Health - Baltimore, MD/US, ¹¹Brigham and Women's Hospital - Boston/US, ¹²University of Nebraska - Omaha, NE/US, ¹³GlaxoSmithKline - King Of Prussia, PA/US, ¹⁴ECLIPSE Study - Various/US, ¹⁵COPDGene Study - Various/US

Rationale: Genomewide association studies (GWAS) have identified a region on chromosome 15q25 associated with risk of COPD. However, several potential COPD susceptibility genes are located within this region.

Methods: The Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE) Study is a multi-center 3-year observational study. Genome-wide SNP genotyping was performed on all ECLIPSE subjects using the Illumina HumanHap550 BeadChip. Gene expression profiling on induced sputum samples from 131 COPD cases was performed using the Affymetrix Human U133 Plus2 array. We integrated the GWAS and gene expression datasets to identify regulatory single nucleotide polymorphisms (rSNPs) associated with expression levels.

Results: Examining all SNPs within 50kb of a transcript (562,787 tests), we found 4315 SNP-probe associations to be significant at FDR-adjusted p<0.05. These 4315 rSNPs were tested for association with COPD in a combined GWAS which included subjects from the ECLIPSE, Norway, and NETT-NAS studies (2940 cases, 1380 controls). After adjusting for 4315 tests (p<1.16x10-5), 2 SNPs were significantly associated with COPD. These

SNPs were located in 2 separate genes on chromosome 15: CHRNA5 and IREB2. Detailed analysis of chromosome 15 demonstrated additional rSNPs for IREB2 located within that gene. rSNPs for CHRNA5 mapped to CHRNA5, IREB2, and sequences between the 2 genes. The rSNPs for IREB2 and CHRNA5 were not in linkage disequilibrium. An additional 70 rSNPs which showed evidence of association with COPD at p<0.01 were genotyped in the International COPD Genetics Network and in the COPDGene Study. We found replication for a SNP in PSORS1C1, in the HLA-C region of chromosome 6 (combined p=2.2x10-4).

Conclusions: Integrative analysis of SNPS from a GWAS with sputum gene expression data supports CHRNA5 and IREB2 as distinct COPD susceptibility genes on chromosome 15q25 and has identified a potential novel COPD susceptibility gene on chromosome 6.

 

 

Menthol Cigarette Usage among African Americans Associated with Persistent Tobacco Use

J.A. Kanyuch¹, P.A. Bercz², J. Chang³, F.J. Martinez⁴, G.J. Criner², M.J. Lawrence¹, P.A. Tomko⁵, M.G. Foreman⁶, E. Silverman⁷, J.D. Crapo⁸, F.C. Sciurba⁹, &. COPDGene Investigators¹⁰

¹University of Pittsburgh - Pittsburgh, PA/US, ²Temple University Hospital - Philadelphia, PA/US, ³University Of Pittsburgh - Pittsburgh/US, ⁴3916 Taubman Center / University of Michigan - Ann Arbor, MI/US, ⁵University of Pittsburgh - Pittsburgh/US, ⁶Morehouse School of Medicine - Atlanta, GA/US, ⁷Channing Laboratory - Boston, MA/US, ⁸National Jewish Health - Denver, CO/US, ⁹University of Pittsburgh Medical Center - Pittsburgh, PA/US, ¹⁰COPDGene Study - Various/US

Rationale: Patterns of smoking and presentation of COPD have been shown to differ between Americans of African (AA) compared to those of European (EA) descent. AA are younger at presentation of COPD despite starting smoking later in life. When compared to EA, AA have a lower overall pack-year smoking history for a given severity of pulmonary function. AA smokers have been reported to demonstrate a preference for menthol vs. non-menthol (NM) cigarettes when compared to EA. Given that menthol smokers tend to have higher nicotine levels than NM smokers, it is plausible that smoking persistence or impact is greater in menthol compared to NM smokers. Menthol usage impacts smoking patterns and COPD presentation in AA smokers. We therefore hypothesized that a preference for menthol cigarettes was associated with persistence of smoking within the tobacco exposed COPDGene^TM cohort and that this effect was greater in AA subjects.

Methods: A retrospective analysis of the COPDGene^TM cohort was performed. Participants completed detailed questionnaires, spirometry, and were asked if they now smoke or did smoke menthol cigarettes. Logistic regression analysis, adjusted for gender, was used to assess the relationship between current smoking status and menthol usage in the overall group and racial subgroups.

Results: Table 1. demonstrates the characteristics of the AA and EA subjects. Notably 37.4% of AA but only 25.4% of EA had a history of menthol cigarette usage (p<0.0001). There was a trend in the overall cohort in the relationship between current smoking status and menthol use status, OR 1.92, 95% CI 0.91-4.05, p=0.08. However, among AA, there was a significantly greater likelihood of persistent smoking among menthol users when compared to NM, OR 2.84, 95% CI = 1.13-7.12, p value = 0.03. This association is not found in EA, OR = 1.14, 95% CI = 0.40-3.25, p = 0.81.

Conclusions: The use of Menthol cigarettes is associated with a significantly greater persistence of cigarette smoking in African vs. European descendent subjects. This association may contribute to the difference in racial smoking and disease patterns. The sociologic or genetic factors resulting in this association will require future investigation.

U.S. Dept of HHS. Tobacco Use Among U.S. Racial/Ethnic Minority Groups -1998.
Chatila et. al. Chest. 2006 ;130:108-18.

 

 

Combined HRCT quantitative measurements of airways and lung parenchyma can better detect COPD compared to individual CT parameters

Yu-Il Kim¹,², David Lynch¹, Joyce Schroeder¹, John Newell¹, James R. Murphy¹, Adam Friedlander ¹, Barry Make¹, James Crapo¹, Edwin Silverman³, Russell P. Bowler¹ and COPDGene® Investigators

¹National Jewish Health, Denver, ²Chonnam National University Hospital, Gwangju, Korea, ³Brigham and Women's Hospital, Boston

Introduction: Quantitative CT analysis is a promising method to differentiate phenotypes of chronic obstructive pulmonary disease (COPD). Each parameter from the airway wall and lung parenchyma has been individually shown to be a predictor of lung physiology. We hypothesized that combined airway and lung density measurements would better detect airflow obstruction and COPD compared to individual CT parameters.

Methods: We used quantitative high resolution CT (HRCT) and spirometry data from the first 2270 subjects [995 smoking controls without evidence of airflow obstruction, 1275 subjects with COPD (216, 529, 341 and 189 with GOLD Stage 1, 2, 3 and 4 respectively)] in the COPDGene® (the Genetic Epidemiology of COPD) cohort.  HRCT measurements included square root of the wall area at an airway internal perimeter of 10mm (SQRTWA@pi10), average wall area % (WA%) of segmental airways, emphysema (LAA%<-950 HU on inspiratory CT) and gas trapping [%lung attenuation area (LAA%) < -856 HU on expiratory CT]. Physiologic measurements included: % predicted forced expiratory volume in 1 second (FEV1%) and % predicted forced vital capacity (FVC%) and FEV1/FVC ratio after bronchodilator. The area under the curve (AUC), sensitivity and specificity were evaluated to detect COPD using receiver operating characteristic (ROC) curves. The cutoff points corresponding to the sensitivity and specificity values that yielded the minimal value for (1-sensitivity)2+(1-specificity)2 were determined as optimal cutoffs. ROC curves for individual CT parameters were compared with those for combined CT parameters (the sum of two individual CT values).

Results: The AUCs of SQRTWA@pi10, WA%, emphysema and gas trapping were 0.605, 0.737, 0.832 and 0.889 respectively. The AUCs of combined CT metrics for SQRTWA@pi10+ gas trapping, SQRTWA@pi10+ emphysema, WA%+emphysema, WA%+gas trapping were 0.835(95%CI: 0.818-0.852), 0.890(95%CI: 0.879-0.903), 0.897(95%CI: 0.883-0.910)   and 0.910(95%CI: 0.898-0.922), respectively. AUCs for WA%+emphysema and WA%+ gas trapping were significantly higher than those of single CT parameters using ROC curves (all P ≤ 0.01). Combined CT parameters had higher sensitivity and specificity at each optimal cut-off value compared to single parameters. Among them, the sum of WA% and gas trapping yielded the highest diagnostic accuracy at a cut-off value of 80.8% (sensitivity=81.2%, specificity=85.8%). The optimal combination for detection of GOLD 1 was also the sum of WA% and gas trapping.

Conclusions: Combined CT airway and parenchymal parameters could detect airflow obstruction more accurately than individual CT parameters. Among the combined parameters, the combination of WA% and gas trapping had the highest accuracy for detection of COPD.

Funding Support: This study was supported by National Heart, Lung and Blood Institute (NHLBI RO1HL 095432, U01 HL089856, U01 HL089897); and UL1 RR025780 from NCRR/HIH.

 

History of Occupational Exposure to Dust and Fumes is Associated with a History of Cardiovascular Disease in COPD Subjects

Gregory L Kinney¹, Nathaniel Marchetti², Jennifer Black-Shinn¹, Adam L Friedlander³, Barry Make³, Elizabeth Regan³, R. Graham Barr⁴, James Murphy³, Edwin K Silverman⁵, James Crapo³, John E Hokanson¹, and the COPDGene Investigators⁶

University of Colorado, Denver¹, Temple University, Philadelphia², National Jewish Health, Denver³, Columbia University⁴, Brigham and Women's Hospital, Boston⁵, COPDGene Study, Various US⁶

RATIONALE: Cardiovascular disease (CVD) is among the most important comorbid conditions of Chronic Obstructive Pulmonary Disease (COPD) and is a leading cause of death. Whether this COPD-CVD relationship is due to a biological link or shared risk factors is not fully understood.  Occupational exposure (OE) to gas, smoke, chemical vapors, fumes or dust is a potential risk factor for COPD and exposure to airborne particulates has been associated with CVD. This study explores the association between OE and a history of clinical CVD in a smoking population.

METHODS: This study reports on the initial 2500 smokers (aged 45 to 80 years) recruited by COPDGene® of which 1845 had complete data for this analysis. CVD (Transient Ischemic Attack, Angioplasty, Heart Attack, Congestive Heart Failure, Coronary Artery Disease, Peripheral Vascular Disease, or Coronary Artery Bypass Surgery) was reported as yes/no at the time of enrolment and OE was reported using standard questions. Other CVD risk factors (high blood pressure (HBP), high cholesterol and diabetes) were recorded by an interviewer, Body Mass Index (BMI) was calculated using height and weight measured at the study visit and COPD was assessed using GOLD criteria. Logistic Regression was performed using Proc SurveyLogistic (SAS 9.2) to account for the multi-center nature of the study.

RESULTS: 1108 (60%) reported OE and these participants were more likely to be male (62% vs 38%, p<0.0001), younger (60.2 vs 62.4, p<0.0001), have more pack years of smoking (48.7 vs 43.0, p<0.0001) and have COPD (54% vs 46%, p=0.0005). They did not differ by BMI (28.5 vs 28.3, p=0.37), or subject report off physician diagnosis of diabetes (13% vs 11%. p= 0.3), HBP (42% vs 40%, p=0.6) or high cholesterol (38% vs 40%, p=0.4). OE was significantly associated with a history of clinical CVD (OR 1.8, 95% C.I. 1.2-2.6, p=0.003) after controlling for age, pack years, COPD, sex, BMI, diabetes, cholesterol and HBP. A domain analysis was conducted to examine factors associated with CVD separately in those with and without COPD and OE was found to be significantly associated with CVD only in those with COPD (OR 2.3, 95% C.I. 1.5-3.8, p=0.0004 in COPD  vs OR 1.1 95% C.I. 0.6-2.1, p=0.8 in those without COPD, interaction p=0.087). Other CVD risk factors (age, HBP and high cholesterol) were associated with CVD in those with and without COPD while smoking and gender were only important in those without COPD.

CONCLUSIONS: Occupational exposure was independently associated with a history of clinical CVD in the COPDGene cohort of smokers, primarily due to a strong, positive association in those with COPD.

Supported by NHLBI U01 HL089856 and U01 HL089897

 

 

Chronic Obstructive Pulmonary Disease and Gastroesophageal Reflux Disease in COPDGene

C.H. Martinez¹, F.J. Martinez¹, Y. Okajima², S. Murray¹, E. Regan³, P. Hanna³, R. Balkissoon³, E.A. Kazerooni¹, J.L. Curtis¹, B.J. Make³, G.R. Washko², M.K. Han¹

¹The University of Michigan Health System - Ann Arbor, MI/US, ²Brigham & Womens Hosp Div of Pul and Critical Care Medicine - Boston, MA/US, ³National Jewish Health - Denver, CO/US

Background: A high prevalence of gastroesophageal reflux disease (GERD) has been reported in subjects with COPD. Others have reported a relationship between the presence of GERD and exacerbation frequency. Here we extend that analysis to further characterize the relationship between not only GERD and exacerbations but also COPD symptoms, lung function and radiologic characteristics.

Methods: 1,282 subjects meeting GOLD criteria for COPD from the COPDGene cohort were analyzed. All subjects had spirometry and volumetric lung CT scanning with emphysema measurement utilizing a -950 HU threshold; bronchial wall thickness was averaged over six segmental airways (one from each lobe). We examined the relationship between subject reported physician diagnosis of GERD and respiratory symptoms, functional status, and CT measures of emphysema and bronchial wall thickness. Univariate comparisons were performed with chi-square or t-student's test where appropriate; multivariate analysis was used to evaluate the independent effect of GERD on measures of dyspnea and functional status.

Results: COPD patients with (n=360) and without GERD (n=903) were of similar age (64.3 vs. 64.0 years, p=0.52) but more were women (53.1% vs. 44%, p=0.005). GERD patients had a greater history of smoking (55 vs. 50.7 pack-years, p=0.008) but with less frequency were current smokers (27.6% vs. 37.9%, p=0.0003), reported more frequent chronic bronchitis symptoms (30.6% vs. 23.5%, p=0.01), and higher frequency of exacerbations in the year prior to enrollment (0.9 vs 0.6 exacerbations, p<0.001). Severity of airflow obstruction was similar (55.2% vs. 56.4% FEV1% predicted, p=0.41) as was emphysema percent between groups (14.0% vs. 14.6%, p=0.48), segmental bronchial wall thickness (1.59 mm for both groups, p=0.98), and segmental wall area percent (62.3% vs. 62.5%, p=0.54). Health status was worse (St. George's Respiratory Questionnaire 41.6 vs 34.1, p<0.001) and dyspnea was worse (MMRC 2.3 vs. 1.8, p<0.001) in those with GERD. BODE index was also higher for subjects with GERD (2.9 vs. 2.6, p=0.02). In multivariate analysis additionally adjusted for age, BMI, gender, lung function, education, and smoking history, GERD had a significant independent association with higher SGRQ score (4.4 points, p<0.001), MMRC (0.5 points, p<0.001) and exacerbation frequency (1.20 fold greater, p=0.03).

Conclusions: In COPD patients, a history of GERD is associated with more frequent exacerbations, greater dyspnea and worse quality of life. More research is needed to determine whether more aggressive treatment of GERD would result in improvements in COPD outcomes.

 

 

Establishing a Method for Longitudinal Follow-up in a Large Multicenter Cohort of Smokers

Elizabeth A Regan¹, Sarah Moyle¹, Douglas Stinson¹, Carla Wilson¹, Ron Tanner², Gerard Criner³, Barry Make¹, Ed Silverman⁴, James Crapo¹

National Jewish Health, Denver¹, Inventive Labs, Denver², Temple University, Philadelphia³, Brigham and Women's Hospital, Boston⁴

Rationale: Maintaining contact with research subjects is important to understand natural history and disease progression. In large multicenter trials this is especially true but the process of obtaining longitudinal data from subjects by study coordinators is expensive and time intensive.  Contacting subjects by using asynchronous bi-directional forms of communication (automated phone calls or internet communication) has the advantage of routinely contacting patients. However, the impersonal nature of the contact and the need to use advanced technology by elderly patient subjects may limit its effectiveness. We devised a semi-automated process that includes web based data entry and automated phone calls (telephony) to obtain regular information about exacerbations, new medical problems and changes in medications.  We hypothesized that subjects would accept an automated system and participate in regular follow-up reports.

Methods: COPDGene was initiated as a cross-sectional study enrolling 10,000 smokers with and without COPD.  A longitudinal follow-up process was created to supplement the initial project.  An automated calling system using a pre-recorded outgoing message that permitted both voice responses and key pad responses was selected (Inventive Labs Corp).  A questionnaire was designed with branching questions to identify exacerbations and treatments (doctor visit, emergency room visit, hospitalization, antibiotic treatment and steroid treatment).  Subjects are given the option to respond by website or by telephony.  Failures by automated system are contacted by personal coordinator call.

Results: The process was IRB approved and operationalized at 21 clinical centers.  During the roll-out phase of the project approximately 72% of enrolled subjects were contacted at least once.  Subjects are contacted every three or six months for follow-up.  De-identified follow-up data has been aggregated and sent from the web/telephony server to the study data coordinating center where it is merged with cross-sectional data on the subjects. Contact success over two rounds of surveys was 69% for Controls and 73% for COPD cases and increased with age ( 45-55 = 66%, 56-65= 73%, > 65= 76%).The number of subjects requiring a coordinator call declined from 30% in the first cycle to 19% in the second cycle and 14% early in the third cycle.

Conclusion: Regular contact of research subjects for longitudinal data collection can be done cost effectively using a combination of internet and automated phone calls.  Research subjects are able and willing to provide information by the system and with repeated use may tend to shift to an automated system rather than relying on a clinical coordinator.

 

 

Genome-wide association study of smoking behaviors in COPD patients

M. Siedlinski¹, M.H. Cho¹, P. Bakke², A. Gulsvik², D.A. Lomas³, W. Anderson⁴, X. Kong⁵, S.I. Rennard⁶, T.H. Beaty⁷, J.E. Hokanson⁸, J.D. Crapo⁹, E.K. Silverman¹, &. COPDGene, and ECLIPSE Investigators¹⁰

¹Brigham & Women's Hospital and Harvard Medical School - Boston, MA/US, ²University of Bergen - Bergen/NO, ³University of Cambridge - Cambridge/UK, ⁴GlaxoSmithKline Research and Development - Research Triangle Park, NC/US, ⁵GlaxoSmithKline Research and Development - King Of Prussia, PA/US, ⁶University of Nebraska Medical Center - Omaha, NE/US, ⁷Johns Hopkins School of Public Health - Baltimore, MD/US, ⁸University of Colorado - Aurora, CO/US, ⁹National Jewish Medical and Research Center - Denver, CO/US, ¹⁰Various - Various/US

Rationale: Cigarette smoking is a major risk factor for COPD and COPD severity. Smoking behaviors, such as age at smoking initiation (ASI), smoking cessation, and number of cigarettes smoked per day (CPD), have substantial heritability. We aimed to identify single nucleotide polymorphisms (SNPs) associated with these phenotypes among COPD subjects. Additionally we sought to verify the associations for smoking cessation and CPD with candidate SNPs, i.e. those previously associated at genome-wide significance level in other populations, among COPD patients.

Methods: Genome wide association studies (GWAS) were conducted in 4 independent cohorts (National Emphysema Treatment Trial (NETT), COPDGene, the GenKOLS cohort from Bergen, Norway, and Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE)) including 3,429 ever-smokers of European ancestry with COPD (GOLD stage II or higher; table 1). Regression analyses of ASI, smoking cessation, current and lifetime CPD were adjusted for potential confounders (e.g. sex, age, % of predicted FEV₁ and principal components for genetic ancestry). SNP imputation resulted in approximately 2.5 million SNPs analyzed per phenotype. A fixed effect, meta-analytic p<5x10^-8 was considered genome-wide significant (5x10^-8≤p<5x10^-7 was interpreted as a suggestive association while p<0.05 as a suggestive association for candidate SNPs).

Results: SNP rs7617754 in the MDS1 and EVI1 complex locus (MECOM) was the most significantly associated SNP in the meta-analysis on lifetime average CPD (p=4.1x10^-7; figure 1). Nominally significant associations with candidate SNPs within alpha-nicotinic acetylcholine receptors 3/5 (CHRNA3/CHRNA5) and Cytochrome P450 2A6 (CYP2A6) loci were observed for lifetime average CPD (figure 1), however only CYP2A6 showed evidence of significant association with current CPD. Several SNPs near the Human Leukocyte Antigen region on chromosome 6p21 and on chromosome 2q21 showed evidence of association with ASI (p=2x10^-7 for the most significant rs7743060 and rs13408379 respectively). No SNPs were associated with current CPD or smoking cessation at the p<10^-6 level. A candidate SNP (rs3025343) in the Dopamine Beta-Hydroxylase (DBH) was significantly (p=0.016) associated with smoking cessation.

Conclusions: We identified two regions associated with ASI and the MECOM locus associated with lifetime average CPD at suggestive genome wide-significance levels among COPD subjects. Interestingly, the 2^nd most significant SNP in the MECOM for analysis of lifetime average CPD (rs6774494; p=2.2x10^-6) was previously associated with nasopharyngeal carcinoma at genome-wide significance level (Bei JX et al.,2010 Nat Genet;42:599-603). Associations between SNPs in CHRNA3/CHRNA5 and CYP2A6 loci with lifetime average CPD and DBH with smoking cessation are also evident in these COPD patients.

Funding: This work was supported by U.S. National Institutes of Health (NIH) grants R01 HL075478, R01 HL084323, P01 HL083069, U01 HL089856 (Silverman), K12HL089990  (Cho), and U01 HL089897 (Crapo). The National Emphysema Treatment Trial was supported by the National Heart, Lung, and Blood Institute contracts (N01HR76101, N01HR76102, N01HR76103, N01HR76104, N01HR76105, N01HR76106, N01HR76107, N01HR76108, N01HR76109, N01HR76110, N01HR76111, N01HR76112, N01HR76113, N01HR76114, N01HR76115, N01HR76116, N01HR76118, N01HR76119), the Centers for Medicare and Medicaid Services, and the Agency for Healthcare Research and Quality. The Norway cohort and the ECLIPSE study (Clinicaltrials.gov identifier NCT00292552; GSK Code SCO104960) are funded by GlaxoSmithKline.

The COPDGene® project described was supported by Award Number U01HL089897 and Award Number U01HL089856 from the National Heart, Lung, And Blood Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, And Blood Institute or the National Institutes of Health.

The COPDGene® project is also supported by the COPD Foundation through contributions made to an Industry Advisory Board comprised of AstraZeneca, Boehringer Ingelheim, Novartis and Sepracor.

M. Siedlinski is a recipient of a postdoctoral fellowship from the Niels Stensen Foundation.

 

 

A Phantom for characterizing trachea air CT numbers in a Multi-center Quantitative CT Lung Study

JP Sieren¹, EA Hoffman¹, J Guo¹, DA Lynch², JD Newell², PF Judy³, and the COPDGene Investigators⁴

University of Iowa, Iowa City¹, National Jewish Health, Denver², Brigham & Womens Hospital, Boston³, COPDGene Study, Various US⁴

Purpose: Multi-center studies using quantitative CT to evaluate emphysema have observed that tracheal air CT numbers are significantly greater than the standardized -1000 Hounsfield Unit (HU) value for air. In order to investigate these differences between scanner models, a phantom was designed to simulate a human tracheal air measurement.

Methods: Human tracheal air was measured on 761 subjects enrolled in the COPDGene study, scanned on seven of the most used scanners in the study.  Three phantoms were used in this study.  The COPDGene Phantom surrounds lung-equivalent foam with a ring of water equivalent material. The new phantom modified the COPDGene phantom by adding a 9 cm thick acrylic cylinder with a central 3 cm air hole. In the modified Catphan 500, an air hole was bored off center in the water equivalent cylindrical phantom.  The performance of these three phantoms was compared across four models of Siemens scanners: Definition Flash (DF), AS+, AS and Definition. Scan protocol was: 0.75mm thick, pitch 1, CTDIvol 15mGy, B35 kernel. CT number of outside air was measured from images within the scan gantry beyond the z-axis limit of the phantom.

Results: In the 761 COPDGene subjects variability in tracheal air was scanner model and manufacturer dependent, ranging from -985 to -955HU.The data for all four scanners using the B35 kernel are summarized in the table. In the new phantom, the inside air showed a positive shift in values that were similar to the modified Catphan500, although it showed increased standard deviation reflecting increased noise. COPDGene Phantom is least effective in regards to simulating in vivo tracheal air measurements.

	Outside Air (HU range; ±SD)	Inside Air (HU range; ±SD)	Water (HU range; ±SD)
COPDGene	-1002 to -1000; ±2	-1004 to -1000; ± 7	-8 to -2; ± 10
New Phantom	-1001 to -1000;  ± 2	-992 to -981; ±20	-10 to -4; ± 12
modCatphan500	-1001 to -1000;  ± 2	-997 to -985; ± 8	10 to 14; ± 10

Conclusion: Measurements of CT numbers in holes surrounded by water like substances show evidence of positive shift. However, these phantoms do not reproduce the magnitude of CT number shift found within the trachea. The full explanation for why tracheal air is more positive than seen in any of these phantoms remains a topic of investigation.

 

 

Genome Wide Association Study of Exacerbation Phenotypes in COPD

E.S. Wan, M.H. Cho, X. Kong, W. Anderson, R. Tal-Singer, D. Lomas, B.J. Klanderman, J.S. Sylvia, J.P. Ziniti, P. Bakke, A. Gulsvik, E. Regan, J. Murphy, B. Make, J. Crapo, J.E. Hokanson, T. Beaty, E.K. Silverman, C.P. Hersh

Introduction: Acute exacerbations of COPD are a significant source of morbidity and mortality.  Recent work supports the existence of distinct groups of COPD subjects with regards to exacerbation frequency.  Although exacerbation frequency generally increases with declining lung function, the "frequent exacerbator" phenotype remains a distinct subgroup in all GOLD stages.  Conversely, a subset of COPD subjects ("non-exacerbators") remain relatively protected from acute exacerbations, often despite significant lung function impairment.  Genetic variation may contribute to each of these distinct phenotypes.

Methods: We performed an analysis of COPD subjects of European descent from four independent cohorts: the Evaluation of COPD Longitudinally to Identify Predictive and Surrogate End-points trial (ECLIPSE, n = 1951), the Norway-Bergen cohort (n=630), a subset of the National Emphysema Treatment Trial (NETT, n=365), and an initial GWAS subset of the COPDGene study (n=495).  "Frequent exacerbators" in each cohort were defined as having ≥2 exacerbations per year.  "Non-exacerbators" had no reported exacerbations during the period of ascertainment.  Genome wide single nucleotide polymorphism (SNP) data from the Illumina platform was analyzed in each cohort for association using logistic regression.  A separate analysis was performed for each of the dichotomized phenotypes: (a) "frequent" (n = 326) versus "infrequent" (n= 3115) and (b) "non-exacerbators" (n= 1184) versus "any exacerbation" (n= 2257) .  Adjustment for population stratification within each cohort was accomplished using principal components analysis.  All models were adjusted for FEV₁ % predicted.  Combined p-values were generated using a fixed-effects meta-analysis implemented in PLINK (v1.07).

Results: Quantile-quantile plots from the combined analyses for "frequent exacerbators" and "non-exacerbators" are shown in Figure 1 (a & b) and reveal minimal evidence of inflation of the test statistic due to population stratification.  In the combined analysis comparing "non-exacerbators" to subjects who reported any history of exacerbations, significant associations (lowest p-value 3.1 x 10^-7) were found on chromosome 1q25 within an intron of the RAB GTPase activating protein 1-like (RABGAP1L) (Table 1).  Significant associations were also found within neighboring genes ring finger and CCCH-type domains 1 (RC3H1) and serpin peptidase inhibitor clade C member 1 (SERPINC1).

Conclusions: Genetic variants in a region of chromosome 1, including genes RABGAP1L, RC3H1, and SERPINC1, may be associated with protection from acute exacerbations of COPD.

 

 

Clinical and radiologic predictors of GOLD-Unclassified subjects in COPDGene

E.S. Wan, J. Hokanson, E. Regan, J. Murphy, B. Make, D. Lynch, J.D. Crapo, E.K. Silverman

Introduction: A significant proportion of smokers have lung function impairment that cannot be classified under the current GOLD staging system.  These GOLD-Unclassified (GOLD-U) subjects, who exhibit reduced FEV₁ but maintain an FEV₁/FVC ratio >0.7, are a poorly characterized group.  We hypothesized that GOLD-U subjects will have distinct clinical, functional, and radiologic characteristics when compared to smokers with normal spirometry (controls) and COPD subjects (GOLD 2-4).

Methods: COPDGene is an ongoing study which enrolls self-identified non-Hispanic whites or African Americans with ≥10 pack-years of smoking between the ages of 45-80.  All subjects complete pre- and post-bronchodilator spirometry, inspiratory and expiratory volumetric CT scans, and a standardized questionnaire. Data from the first 2500 subjects enrolled in COPDGene were analyzed for this study.  Univariate comparisons of clinical and radiologic variables were made between GOLD-U subjects (n=227) and smoking controls (n=1001) and GOLD 2-4 (n=1059) subjects.   To control for the effects of multiple covariates, a multivariate regression model was constructed using stepwise logistic regression with GOLD-U status as the dependent variable.   All variables with a univariate p-value <0.05 were considered candidates in the multivariate analysis. A separate multivariate regression was performed in the subgroup of subjects with radiologic measurement variables available (n = 217 GOLD-U / 974 controls / 935 COPD subjects).

Results: Results from the univariate comparisons summarized in Table 1 indicate that, compared to smoking controls, GOLD-U subjects have increased body mass index (BMI), decreased six minute walk distance (6MWD), less emphysema, increased airway wall thickness, and reduced total lung capacity (all p<0.05).  Compared to GOLD 2-4 subjects, GOLD-U subjects have increased BMI and reduced TLC (p <0.05). GOLD-U subjects have a higher proportion of subjects with diabetes and who are of non-white race compared to both groups.  The results of the multivariable analyses are summarized in Table 2. Although body mass index was a significant predictor in all of the models, additional significant predictors, including differences in airway wall thickness and resting oxygen saturation, were identified.

Conclusion: GOLD-U subjects have distinct characteristics beyond increased body mass index when compared to smoking controls and COPD subjects.

 

A Comparison of Visual and Quantitative Methods to Identify Interstitial Lung Abnormalities in Cigarette Smokers

Tracy J. Wanner, M.D., ¹ Ivan O. Rosas, M.D.,¹ Isis E. Fernandez, M.D.,¹ Mizuki Nishino, M.D.,^2,3 Yuka Okajima, M.D.,² Tsuneo Yamashiro, M.D.,^{2, 3} James C. Ross, M.S.,^4,5 Raúl San José Estépar, Ph.D.,^3,5 David A. Lynch M.D.,⁶ Alejandro A. Diaz, M.D.,^{1, 7} Katherine D'Aco M.S.,^1, Edwin K. Silverman M.D., Ph.D.,^{1, 4} Hiroto Hatabu M.D., Ph.D.,^{2, 3}, George R. Washko, M.D., M.M.Sc.,¹ Gary M. Hunninghake, M.D., M.P.H.,^{1, 4} and the COPDGene Investigators

¹Pulmonary and Critical Care Division, Brigham and Women's Hospital, Harvard Medical School, Boston MA; ²Department of Radiology, Brigham and Women's Hospital, Boston MA; ³Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Boston MA; ⁴Channing Laboratory, Brigham and Women's Hospital, Boston MA; ⁵Surgical Planning Laboratory, Department of Radiology, Brigham and Women's Hospital, Boston MA; ⁶Department of Radiology, National Jewish Medical and Research Center, Denver CO; ⁷Department of Pulmonary Diseases. Pontificia Universidad Católica de Chile, Santiago, Chile

Introduction: Prior studies have suggested that thresholds of increased lung density identified in cardiac computed tomography (CT) scans can identify interstitial lung abnormalities (ILA), however this has not been validated.

Methods: We measured the percentage of high attenuation areas between -600 and -250 HU (HAA% - an automated quantitative method) of the lung in 2078 inspiratory CT scans from the COPDGene study where ILA had been defined by a sequential reading method (SRM - a method of visual inspection). Linear and logistic regressions were used for multivariate association analyses.

Results: Linear increases in HAA% were associated with ILA (for every 1% increase in the volume of lung occupied by HAAs, there was a 3.3-fold increase in the odds of having ILA, odds ratio [OR] 3.29, 95% confidence interval [CI] 2.60-4.17, P=4x10^-23) after adjustment for relevant covariates including CT metrics of total lung capacity (TLC).  In contrast, the presence of HAAs in >10% of the lung had a positive predictive value of 19% and a negative predictive value of 92% for ILA defined by SRM.  While there was significant correlation between methods (P=0.03), the kappa was low (0.03). Of note, the association between HAA% and ILA defined by SRM was modified (P values <6x10^-5 for the interactions) by relevant covariates (e.g. percent of emphysema and body mass index). We evaluated the performance of alternate HAA% thresholds for identifying ILA by SRM by generating receiver operating characteristic curves.  The area under the curve was maximized (AUC 0.76) for the presence of HAA% > 3.56% of the lung (the 50^th percentile of HAA% in COPDGene).  Finally, after adjusting for covariates, linear increases in HAAs were associated with reductions in TLC in analyses of all subjects, and in analyses limited to subjects with ILA defined by SRM (for every 1% increase in the volume of lung occupied by HAAs in subjects with ILA defined by SRM, there is a 0.278 liter reduction in TLC, 95% CI [-]0.340-[-]0.216 , P<1x10^-30).

Conclusions: Increases in HAA% are associated with the presence of visually-defined ILA, and are also associated with an increasing severity of lung volume reduction in those with visually-defined ILA.  However, our results suggest that the interpretation of increased HAA% is complex and is modified by multiple covariates, such as emphysema and body mass index.  This may explain the relatively low positive predictive value of HAA% for the presence of ILA.

Supported by grants U01HL089897 and U01089856.