Inﬂammatory Biomarkers Improve Clinical Prediction of Mortality in Chronic Obstructive Pulmonary Disease
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 185 2012
Bartolome R. Celli1, Nicholas Locantore2, Julie Yates2, Ruth Tal-Singer3, Bruce E. Miller3, Per Bakke4, Peter Calverley5, Harvey Coxson6, Courtney Crim2, Lisa D. Edwards2, David A. Lomas7, Annelyse Duvoix7, William MacNee8, Stephen Rennard9, Edwin Silverman1, Jørgen Vestbo10,11, Emiel Wouters12, and Alvar Agustı´13,14, for the ECLIPSE Investigators
1Pulmonary and Critical Care Division, Brigham and Women’s Hospital, Boston, Massachusetts; 2GlaxoSmithKline, Research Triangle Park, North Carolina; 3GlaxoSmithKline, King of Prussia, Pennsylvania; 4University of Bergen, Bergen, Norway;5Department of Respiratory Medicine, University Hospital Aintree, Liverpool, United Kingdom; 6Department of Radiology, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia, Canada; 7Department of Medicine, University of Cambridge, Cambridge Institute for Medical Research, Cambridge, United Kingdom; 8University of Edinburgh and Royal Inﬁrmary, Edinburgh, United Kingdom; 9University of Nebraska Medical Center, Omaha, Nebraska; 10Department of Cardiology and Respiratory Medicine, Hvidovre Hospital/University of Copenhagen, Copenhagen, Denmark; 11University of Manchester, Manchester Academic Health Science Centre, Manchester, United Kingdom; 12Department of Respiratory Medicine, Maastricht University Medical Centre, Maastricht, The Netherlands; 13Thorax Institute, Hospital Clinic, IDIBAPS, University of Barcelona, Barcelona, Spain; and 14FISIB, CIBER Enfermedades Respiratorias, Mallorca, Spain
AT A GLANCE COMMENTARY
Scientiﬁc Knowledge on the Subject
Several clinical variables are known to be associated with mortality in chronic obstructive pulmonary disease (COPD), and in some cases these have been combined into multidimensional matrices that provide better prediction than the sum ofthe individual components. The knowledge base on the importance of biomarkers in COPD is expanding, but the role of biomarker levels in predicting mortality is unknown.
What This Study Adds to the Field
This study investigated associations between mortality and levels of a broad range of biomarkers collected in a large cohort of subjects with COPD studied over 3 years. The addition of a panel of selected biomarkers to established clinicalmeasures increases the capacity to predict mortality.
Rationale: Accurate prediction of mortality helps select patients for interventions aimed at improving outcome. Objectives:Because chronic obstructive pulmonarydisease is characterizedby low-grade systemic inﬂammation, we hypothesized that addition ofinﬂammatorybiomarkers to establishedpredictivefactors will improve accuracy. Methods:Atotal of 1,843 patients enrolled in the Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints study were followed for3years. Kaplan-Meier curves, log-rank analysis, and Cox proportional hazards analyses determined the predictive value for mortality of clinical variables, whileC statistics assessed the added discriminative power offered by addition of biomarkers. Measurements and Main Results:At recruitment we measured anthro-pometrics,spirometry,6-minutewalkdistance,dyspnea,BODEindex, historyofhospitalization,comorbidities,and computed tomography scan emphysema. White blood cell and neutrophil counts, serum or plasmalevelsofﬁbrinogen,chemokineligand18,surfactantproteinD, C-reactive protein, Clara cell secretory protein-16, IL-6 and -8, and tumor necrosis factor-a were determined at recruitment and subsequentvisits.A total of 168 of the1,843patients(9.1%) died. Nonsurvivorswereolderandhadmoresevereairﬂowlimitation,increased dyspnea,higherBODE score,more emphysema,andhigherratesof comorbidities and history of hospitalizations. The best predictive modelfor mortality usingclinical variablesincluded age,BODE,and hospitalizationhistory(C statistic of0.686; P , 0.001). One single biomarker(IL-6) signiﬁcantlyimproved theC statistic to0.708,but thiswasfurtherimproved to0.726(P ¼ 0.003)bytheadditionof all biomarkers.
Conclusions:The additionofapanelof selected biomarkers improves theability of established clinicalvariablestopredict mortalityin chronic obstructive pulmonary disease. Clinical trial registeredwith www.clinicaltrials.gov (NCT00292552).
Keywords: pulmonarydisease,chronicobstructive;prognosis;mortality; biologic markers
Chronic obstructive pulmonary disease (COPD) is currently the fourth highest cause of death in the world, and it is predicted to be the third by the year 2020 (1). Accurate prediction of mortality is important because it helps identify patients inwhom the implementation of speciﬁc therapeutic measures can improve outcome. Several variables that predict mortality have been identiﬁed in COPD, including the severity of airﬂow limitation as measured by FEV1 (2), the presence of arterialhypoxemia or hypercapnia (3), exercise performance (4, 5), degree of breathlessness (6), and a low body mass index (BMI) (7). Their integration into multidimensional indices, such as the BODE (BMI, FEV1, dyspnea, and 6-minute walk distance[6MWD]) (8) and the ADO (age, dyspnea, and FEV1) (9) for mortality or the DOSE (dyspnea, FEV1, smoking status, and frequency of exacerbations) (10) for exacerbations, has been shown to predict outcome better than any of the individualvariables by themselves.
This likely reﬂects the fact that COPD is a complex, heterogeneous disease with pulmonary and extrapulmonary manifestations (11, 12) that are not captured by a single variable. So far all predictive variables in COPD, either alone or incombination, are clinical in nature. Because COPD is also a complex and heterogeneous disease at the genetic, cellular, and molecular level, it is likely that the predictive accuracy of clinical measures can be extended with the use of biomarkers thatreﬂect pathobiologic pathways that may be altered in this disease, as has been shown in cardiovascular diseases (13).
It is now recognized that COPD is characterized by low-grade chronic systemic inﬂammation (14). Several biomarkers, including C-reactive protein (CRP) in some (15) but not all (16) studies, chemokine (C-C motif) ligand 18/pulmonary andactivation-regulated chemokine (CCL-18/PARC) (17), IL-6 (18), and surfactant protein-D (SP-D) (19) have been thought to be associated with increased risk of death in patients with respiratory disease. However, all of these biomarkers havebeen studied singly and no study has evaluated their value compared with accepted clinical predictors of death in patients with COPD. We hypothesized that the addition of a panel of biomarkers to clinical variables known to predict mortality inCOPD,suchasage,FEV1, BODE, or hospitalizations because of exacerbations of the disease, will improve the accuracy for predicting the risk of death in patients with COPD. Here, we tested this hypothesis using data prospectively collected inthe Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) study, a 3-year observational study aimed at identifying predictive surrogate endpoints in COPD.
The study design of ECLIPSE (Clinicaltrials.gov identiﬁer NCT00292552; GSK study code SCO104960) has been published previously (20). Brieﬂy, ECLIPSE is an observational, longitudinal, and controlled study where, after the baseline visit, participantswere evaluated at 3 months, 6 months, and then every 6 months for 3 years. In this report we present the longitudinal analysis of mortality using the clinical and biomarker data obtained at baseline. Death was determined up to Day 1,060 of the study. All-cause mortality was used as the outcome; no attempts were made to determine cause of death. ECLIPSE complies with the Declaration of Helsinki and Good Clinical Practice Guidelines, and has been approved by the ethics committees of the participatingcenters. All participants provided written informed consent before the performance of all study-related assessments.
ECLIPSE studied 2,164 patients with COPD (Global Initiative for Chronic Obstructive Pulmonary Disease stages II–IV); 337 smoking control subjects; and 245 nonsmoking control subjects. The current analysis includes only the patients with COPD whohad full biomarker data. Inclusion criteria were as follows: male and female subjects aged 40–75 years, baseline post-bronchodilator FEV1 less than 80% of the reference value and FEV1/FVC of less than or equal to 0.7, and current or ex-smokers with asmoking history of greater than or equal to 10 pack-years. Key exclusion criteria were the presence of a respiratory disorder other than COPD, other signiﬁcant inﬂammatory diseases, or a reported COPD exacerbation within 4 weeks of enrollment.Patients with COPD were recruited from the outpatient clinics of the participating centers.
Clinical characterization. All methods have been described in the baseline and protocol (20, 21) ECLIPSE manuscripts. In summary, the American Thoracic Society respiratory questionnaire, the modiﬁed Medical Research Questionnaire, and the COPD-speciﬁc version of the St. George’s Respiratory Questionnaire were used to record clinical data. Exacerbations requiring treatment with antibiotics, oral corticosteroids, or hospitalization in the year before the study were also recorded. Comorbidities wereself-reported and registered using the American Thoracic Society–Division of Lung Diseases-78 questionnaire. Nutritional status was assessed by the BMI.
Functional measurements. Spirometry and the 6MWD test were performed according to international guidelines (22). Spirometric reference values were those of the European Community for Coal and Steel (23). The BODE index was calculated aspreviously reported (8).
Quantiﬁcation of emphysema by computed tomography scan. All subjects underwent a low-dose computed tomography (CT) scan of the chest using multidetector-row CT scanners (GE Healthcare [Milwaukee, WI] or Siemens Healthcare [Erlangen,Germany]) as described elsewhere (20). All scans were evaluated centrally at the University of British Columbia, Vancouver, Canada. Emphysema was quantiﬁed as the percentage of lung CT voxels below a threshold of 2950 Hounsﬁeld units using thesoftware Pulmonary Workstation 2.0 (VIDA Diagnostics, Iowa City, IA).
Inﬂammatory biomarkers. Whole blood was collected by venipuncture into vacutainer tubes. Serum was prepared by allowing the blood to clot for 30 minutes at room temperature followed by centrifugation at 1,500 3 g for 10–15 minutes. Plasma(ethylenediaminetetraacetic acid anticoagulant) was obtained by centrifugation of vacutainer tubes at 2,000 3 g for 10–15 minutes. Serum and plasma were stored at 2808C until analyzed. CCL-18/PARC, SP-D, IL-8, Clara cell secretory protein 16 (CC-16),and tumor necrosis factor (TNF)-a were measured in serum samples. Fibrinogen and CRP (high-sensitivity method) were measured in plasma samples. All protein biomarkers were measured by validated immunoassays. Total white blood cells (WBC) andneutrophils were counted by automated method. The biomarker performance information is presented in the online supplement (see Table E1).
Demographic characteristics have been summarized as mean and SD or percentage, as applicable. Blood biomarkers (excluding WBC and neutrophil counts) have been summarized using median and interquartile range, and have been log transformed andstandardized before modeling to conform to the normality assumptions of the underlying models. Survivor and nonsurvivor characteristics were compared using an analysis of variance test for continuous variables and Fisher’s exact test for categoricalvalues. Correlations between variables of interest were explored using Spearman Rho.
Kaplan-Meier curves of the individual clinical risk factors and biomarkers analyzed in this cohort are presented. To establish the relationships of biomarkers to death after adjusting for the clinical variables analyzed here (age, BODE, number of previousexacerbations) we used Cox proportional hazards regression. The added discriminative power offered by the addition of biomarkers to clinical variables was analyzed using C statistics according to the method described in Pencina and D’Agostino (24).Differences in C statistics between any two models were estimated using the jackknife estimation method described in Antolini and coworkers (25). In addition, we divided the cohort into two groups. Patients were matched on age and BODE and the Cstatistic analysis on the subgroups was re-run and the results are shown in the online supplement (see Table E2). All tests performed (SAS Version 9.1.3, SAS, Cary, NC) were two-sided tests at the 0.05 level of signiﬁcance. All P values are nominal, as noadjustment was made for multiple comparisons.
Role of the Funding Source
The study was sponsored by GlaxoSmithKline. A Steering Committee and a Scientiﬁc Committee comprising 11 academics and 5 representatives of the sponsor developed the original study design and concept, the plan for the current analyses, approvedthe statistical plan, had full access to the data, and were responsible for decisions with regard to publication. The study sponsor did not place any restrictions with regard to statements made in the ﬁnal paper.
The consort diagram of the patients with COPD included in this study is shown in Figure 1. Of the 2,164 patients enrolled in ECLIPSE, complete clinical and biomarker data were available in 1,843 of them (85.2%), 168 of whom (9.1%) diedduring follow-up. The clinical characteristics of the patients excluded from the analysis because of incomplete biomarker data were similar in anthropometrics, lung function, walking distance, and degree of CT emphysema but had slightly worseBODE index, St. George’s Respiratory Questionnaire scores, and oxygen saturation.
Table 1 compares the baseline clinical and physiologic characteristics of survivors and nonsurvivors. The latter were older, had more severe airﬂow limitation, reported more dyspnea, had a lower 6MWD, had more emphysema by CT scan,had a higher BODE score, and had more comorbidities. By contrast, sex, smoking status, and BMI were not different between these two groups. Kaplan-Meier analysis showed differences in survival for different age groups, BODE index groups,and the incidence of hospitalization caused by exacerbations of COPD in the year before the study (Figure 2). These three variables were used as the baseline clinical model because the addition of any or all of the other clinical variables failed toimprove the model. Age and BODE were entered into the Cox regression model as continuous covariates.
The levels of the biomarkers determined in the study were higher in nonsurvivors (Table 1). This was not the case for TNF-a, the levels of which were not signiﬁcantly different between groups (results not shown). However, most patients hadundetectable low levels of TNF-a.
Kaplan-Meier survival analysis conﬁrmed that patients with values that were higher than the median value obtained in the control subjects in ECLIPSE of IL-6, CCL-18/PARC, ﬁbrinogen, CRP, and SP-D, but not CC-16, were less likely tosurvive at the end of 3 years (P , 0.001 by log-rank test) (Figure 3).
Cox regression analysis (Table 2) showed that, after adjusting for age, previous hospitalizations, and the BODE index, abnormal levels of some (WBC, neutrophils, IL-6, CCL-18/ PARC, CRP, IL-8, ﬁbrinogen, and SP-D), but not all (CC-16)biomarkers were independently and signiﬁcantly associated with mortality.
To evaluate the effect of adding a panel of biomarkers to the baseline clinical model (age, BODE, and previous hospitalizations), only those individual biomarkers that were independently associated with mortality in the Cox regression modeladjusted for the clinical variables (Table 2) were considered, with two exceptions. Because the correlation between WBC and neutrophils was extremely high (Rho ¼ 0.92), neutrophils were excluded. Likewise, because of incomplete data forCCL-18/PARC (n ¼ 1569), this marker was also excluded from the main analysis.
Table 3 shows how the addition of biomarkers improves the predictive value of the baseline clinical model using C statistics. The C statistic value of the clinical model alone (age, BODE, and hospitalizations) was 0.686. The individual additionof the biomarkers discussed previously did improve the predictive value (C statistic) of the combined index, but their contribution was relatively small (Table 3), with only IL-6 signiﬁcantly improving the C statistic on its own. By contrast, whenall these biomarkers were added together as a panel, the improvement in predictive value (C statistic ¼ 0.726) was statistically signiﬁcant. In the subset of 1,579 subjects with all biomarkers including CCL-18/PARC, the results were similar to themain analysis. CCL-18/PARC behaved similarly to most individual biomarkers, and contributed as part of the full biomarker panel to an improve-DISCUSSION ment in predictive value (C statistic ¼ 0.697 in the clinical model
This prospective study in a large and well-characterized cohort and 0.742 in the full model), which was also statistically signifiof patients with moderate to very severe COPD provides two cant. The C statistic values were similar when the cohort was split important findings: the level of several inflammatory biointo two subgroups (see Table E2). markers determined at recruitment was significantly higher in nonsurvivors over the 3 years of the study; and the addition of a selected panel of biomarkers to a model that includes well-established clinical factors improves signiﬁcantly the risk stratiﬁcation for all-cause mortality in these patients.
Over the past few years there has been a growing interest in the ﬁeld of biomarkers in COPD. Unfortunately, most of the studies have been based on existing databases of patients recruited for pharmacologic trials and/or studies that are cross-sectional in nature. CRP was the ﬁrst biomarker to be investigated in COPD. Most studies have shown that CRP levels are elevated in these patients, compared with nonsmokers and smokers without airﬂow obstruction (15, 16, 26–28), but therelationship between CRP levels and mortality remains controversial. Whereas Dahl and coworkers found an association between CRP levels and hospitalization and death in a population study (15), this was not conﬁrmed by DeTorres andcoworkers (16). A recent report by Sin and coworkers (17) used data from the Lung Health Study and ECLIPSE and demonstrated an association between CCL-18/PARC and increased risk of death, but whether its addition yielded any prognostic value to clinical variables already known to predict outcome was not investigated. Other candidate biomarkers studied in COPD include circulating levels of CC-16 (29) and SP-D (30). The former, a marker of Claracell toxicity, appears reduced in patients with COPD and its levels were associated with rate of decline of FEV1 in the same ECLIPSE cohort (31). The lung-derived protein SP-D, however, is associated with presence of pulmonary inﬂammationand is elevated in smokers (with or without COPD). None of these studies evaluated the relationship between the levels of these biomarkers and survival in stable patients with COPD and whether they add value to accepted predictors ofsurvival. Using a protein microarray platform, Pinto-Plata and coworkers (32) identiﬁed a panel of 24 markers of inﬂammation, tissue destruction, and repair that were signiﬁcantly related to lung function, exercise capacity, the BODE index, andexacerbation frequency. However, because the study was cross-sectional, the proteomic proﬁle could not be related to mortality. By contrast, Man and coworkers (33) analyzed data from the Lung Health Study and reported an associationbetween a high ratio of CRP (inﬂammatory marker) to ﬁbronectin (repair marker) with mortality. Yet, this study used serum collected midway through an interventional study, included patients primarily with mild COPD, and had a very lowmortality rate. To our knowledge, our study is the ﬁrst to investigate if the addition of biomarker levels to well-established clinical predictors of outcome adds relevant prognostic information.
Interpretation of Findings
Our results show that a panel of selected biomarkers (WBC counts, IL-6, ﬁbrinogen, CCL-18/PARC, CRP, IL-8, and SP-D) were not only elevated in nonsurvivors compared with survivors (Table 1), but were also associated with mortality over3 years (Figure 3) after adjusting for clinical variables known to predict death in COPD (Table 2), whereas this was not the case for other biomarkers previously thought to be potential predictors of outcome in COPD, such as TNF-a or CC-16.That WBC and neutrophil counts, IL-6, IL-8, ﬁbrinogen, CCL-18/PARC, CRP, and SP-D do add independent predictive information is further supported by the results of the Cox proportional analyses (Table 2) and the use of C statistic (Table3). Using C statistics, only IL-6 independently added predictive power to the basic clinical model, whereas the other biomarkers individually improved the model only marginally. We determined WBC and neutrophils and the correlation betweenthem is extremely high (Spearman rho ¼ 0.92), these two values are essentially interchangeable, and no value is added by combining the two measures together. However, the addition of all the biomarkers in the panel increased the C statisticsignificantly suggesting that the use of integrative analyses describes better the complexity of COPD. It is impossible to determine the proportion or number of patients whose abnormal biomarker expression would identify increased risk of death above those detected using clinical variables because the magnitudeof increase provided by the C statistic has not been related to precise clinical metrics. However, the magnitude of the additional predictive power for mortality provided by the biomarkers in this study is similar to that described recently inpatients with cardiovascular disease using different biomarkers (13).
Strengths and Limitations
The large sample size and multicenter nature of the cohort studied, its careful clinical, radiologic, functional, and biologic characterization, and its prospective design and long follow-up time are clear strengths of this study. However, severalpotential limitations deserve comment. First, there was no adjudication committee to specify the correct cause of death. The Toward a Revolution in COPD Health (34) and Understanding Potential Long-Term Impacts on Function withTiotropium (35) studies showed that the cause of death can be attributed wrongly if it relies exclusively on the death certiﬁcate. However, for the healthcare provider, it is still important to be able to predict all-cause mortality risk and evaluatethe potentially modiﬁable factors to help guide individual patient management and therapeutic approaches. Second, the panel of biomarkers selected did not include some that have been thought to be important in the pathobiology of COPD, suchas the metalloproteinases and growth factors (36). This does not negate the value of our ﬁndings, and actually provides room to improve accuracy if in due time these other biomarkers also are shown to relate to poor outcome. Third, the samecould be said about the number of comorbidities and their relationship to a poor outcome. This also provides room for future studies aimed at discerning the inﬂuence of comorbidity on biomarker levels and their relationship to the risk of death.Fourth, it could be argued that there was no derivative and validating cohort. However, this is customary for nonvalidated biomarkers, whereas in this study we compared validated clinical and serum biomarkers modeled on studies in thecardiovascular arena. Fifth, although the C statistic is the method most commonly used to assess model discrimination, its use for the evaluation of biomarkers as risk predictors has been questioned. This is because signiﬁcant increases in the Cstatistic require very large independent associations of the marker with the outcome of interest. Thus, the signiﬁcant increments in the C statistic observed in the present study (Table 3) indicate that a multimarker approach represents asubstantial improvement in the performance of the model. The fact that the Cox analysis and the C statistics agreed in the added predictive value of the biomarkers provides strong support to this approach. The panel of clinical predictors studiedhere includes age; the BODE index (that in turn considers BMI, FEV1, dyspnea, and exercise tolerance); and hospitalizations. These variables cover most but not all clinical risk factors thus far identiﬁed. For instance, we did not include in theanalysis arterial blood gases, pulmonary hemodynamics, or heart function, because they were not determined in ECLIPSE. Yet, the clinical variables included in the model are readily available to most practicing physicians.
The addition of WBC counts and the systemic levels of IL-6, CRP, IL-8, ﬁbrinogen, CCL-18/PARC, and SP-D improve significantly the ability of clinical variables to predict mortality in patients with COPD.
Acknowledgment: The authors acknowledge all participants for their willingness to support medical research and all the medical, nursing, and technical staff involved in the Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints study.
- Mannino DM, Buist AS. Global burden of COPD: risk factors, prevalence, and future trends. Lancet 2007;370:765–773.
- Anthonisen NR, Wright EC, Hodgkin JE. Prognosis in chronic obstructive pulmonary disease. Am Rev Respir Dis 1986;133:14–20.
- Nocturnal Oxygen Therapy Trial Group. Continuous or nocturnal oxygen therapy in hypoxemic chronic obstructive lung disease: a clinical trial. Ann Intern Med 1980;93:391–398.
- Cote CG, Casanova C, Marin JM, Lopez MV, Pinto-Plata V, de Oca MM, Dordelly LJ, Nekach H, Celli BR. Validation and comparison of reference equations for the 6-min walk distance test. EurRespirJ 2008;31:571–578.
- Oga T, Nishimura K, Tsukino M, Sato S, Hajiro T. Analysis of the factors related to mortality in chronic obstructive pulmonary disease: role of exercise capacity and health status. Am J Respir Crit Care Med 2003;167:544–549.
- Nishimura K, Izumi T, Tsukino M, Oga T. Dyspnea is a better predictor of 5-year survival than airway obstruction in patients with COPD. Chest 2002;121:1434–1440.
- Schols AM, Broekhuizen R, Weling-Scheepers CA, Wouters EF. Body composition and mortality in chronic obstructive pulmonary disease. Am J Clin Nutr 2005;82:53–59.
- Celli BR, Cote CG, Marin JM, Casanova C, Montes de Oca M, Mendez RA, Pinto Plata V, Cabral HJ. The body-mass index, airﬂow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease. NEnglJMed 2004;350:1005–1012.
- Puhan MA, Garcia-Aymerich J, Frey M, ter Riet G, Anto´ JM, Agustı´ AG, Go´ mez FP, Rodrı´guez-Roisı´n R, Moons KG, Kessels AG, et al. Expansion of the prognostic assessment of patients with chronic obstructive pulmonary disease: the updated BODE index and the ADOindex. Lancet 2009;374:704–711.
10. Jones RC, Donaldson GC, Chavannes NH, Kida K, Dickson-Spillmann M, Harding S, Wedzicha JA, Price D, Hyland ME. Derivation and validation of a composite index of severity in chronic obstructive pulmonary disease: the DOSE Index. Am JRespirCritCareMed2009;180:1189–1195.
11. Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, Fukuchi Y, Jenkins C, Rodriguez-Roisin R, van Weel C, et al.; Global Initiative for Chronic Obstructive Lung Disease. Global strategy for the diagnosis, management, and prevention of chronic obstructivepulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 2007;176:532–555.
12. Celli BR, MacNee W. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 2004;23:932–946.
13. Zethelius B, Berglund L, Sundstrom J, Ingelsson E, Basu S, Larsson A, Venge P, Arnlo¨v J. Use of multiple biomarkers to improve the prediction of death from cardiovascular causes. NEngl J Med 2008;358:2107–2116.
14. Barnes PJ. Small airways in COPD. N Engl J Med 2004;350:2635–2637.
15. Dahl M, Vestbo J, Lange P, Bojesen SE, Tybjaerg-Hansen A, Nordestgaard BG. C-reactive protein as a predictor of prognosis in chronic obstructive pulmonary disease. Am JRespirCritCareMed 2007;175:250–255.
16. de Torres JP, Pinto-Plata V, Casanova C, Mullerova H, Co´rdoba-Lanu´sE, Muros de Fuentes M, Aguirre-Jaime A, Celli BR. C-reactive protein levels and survival in patients with moderate to very severe COPD. Chest 2008;133:1336–1343.
17. Sin DD, Miller B, Duvoix A, Man SF, Zhang X, Silverman EK, Connett JE, Anthonisen NA, Wise RA, Tashkin D, et al. ECLIPSE Investigators. Serum PARC/CCL-18 concentrations and health outcomes in chronic obstructive pulmonary disease. Am J Respir Crit Care Med2011;183:1187–1192.
18. Mehrotra N, Freire AX, Bauer DC, Harris TB, Newman AB, Kritchevsky SB, Meibohm B, Health ABC. Study. Predictors of mortality in elderly subjects with obstructive airway disease: the PILE score. Ann Epidemiol 2010;20:223–232.
19. Calfee CS, Ware LB, Glidden DV, Eisner MD, Parsons PE, Thompson BT, Matthay MA; National Heart, Blood, and Lung Institute Acute Respiratory Distress Syndrome Network. Use of risk reclassiﬁcation with multiple biomarkers improves mortality prediction in acute lung injury. Crit Care Med 2011;39:711–717.
20. Vestbo J, Anderson W, Coxson HO, Crim C, Dawber F, Edwards L, Hagan G, Knobil K, Lomas DA, MacNee W, et al.; ECLIPSE investigators. Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE). Eur Respir J 2008;31:869–873.
21. Agusti A, Calverley PM, Celli B, Coxson HO, Edwards LD, Lomas DA, MacNee W, Miller BE, Rennard S, Silverman EK, et al.; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) investigators. Characterisation of COPD heterogeneity inthe ECLIPSE cohort. Respir Res 2010;11:122.
22. Brusasco V, Crapo R, Viegi G. Coming together: the ATS/ERS consensus on clinical pulmonary function testing. Eur Respir J 2005;26:1–2.
23. Roca J, Burgos F, Sunyer J, Saez M, Chinn S, Anto´ JM, Rodrı´guez-Roisin R, Quanjer PH, Nowak D, Burney P. References values for forced spirometry. Group of the European Community Respiratory Health Survey. Eur Respir J 1998;11:1354–1362.
24. Pencina MJ, D’Agostino RB. Overall C as a measure of discrimination in survival analysis: model speciﬁc population value and conﬁdence interval estimation. Stat Med 2004;23:2109–2123.
25. Antolini L, Boracchi P, Biganzoli E. A time-dependent discrimination index for survival data. Stat Med 2005;24:3927–3944.
26. Broekhuizen R, Wouters EF, Creutzberg EC, Schols AM. Raised CRP levels mark metabolic and functional impairment in advanced COPD. Thorax 2006;61:17–22.
27. Man SF, Connett JE, Anthonisen NR, Wise RA, Tashkin DP, Sin DD. C-reactive protein and mortality in mild to moderate chronic obstructive pulmonary disease. Thorax 2006;61:849–853.
28. Pinto-Plata VM, Mullerova H, Toso JF, Feudjo-Tepie M, Soriano JB, Vessey RS, Celli BR. C-reactive protein in patients with COPD, control smokers and non-smokers. Thorax 2006;61:23–28.
29. Lomas DA, Silverman EK, Edwards LD, Miller BE, Coxson HO, Tal-Singer R; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) investigators. Evaluation of serum CC-16 as a biomarker for COPD in the ECLIPSE cohort. Thorax2008;63:1058–1063.
30. Lomas DA, Silverman EK, Edwards LD, Locantore NW, Miller BE, Horstman DH, Tal-Singer R; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints study investigators. Serum surfactant protein D is steroid sensitive and associated with exacerbationsof COPD. Eur Respir J 2009;34:95–102.
31. Vestbo J, Edwards LD, Scanlon PD, Yates JC, Agusti A, Bakke P, Calverley PMA, Celli B, Coxson HO, Crim C, et al., for the ECLIPSE Investigators. Changes in forced expiratory volume in 1 second over time in COPD. N Engl J Med 2011;365:1184–1192.
32. Pinto-Plata V, Toso J, Lee K, Park D, Bilello J, Mullerova H, De Souza MM, Vessey R, Celli B. Proﬁling serum biomarkers in patients with COPD: associations with clinical parameters. Thorax 2007;62:595–601.
33. Man SF, Xing L, Connett JE, Anthonisen NR, Wise RA, Tashkin DP, Zhang X,Vessey R,Walker TG, Celli BR, et al. Circulating ﬁbronectin to C-reactive protein ratio and mortality: a biomarker in COPD? Eur Respir J 2008;32:1451–1457.
34. Calverley PM, Anderson JA, Celli B, Ferguson GT, Jenkins C, Jones PW, Yates JC, Vestbo J; TORCH investigators. Salmeterol and ﬂuticasone propionate and survival in chronic obstructive pulmonary disease. N Engl J Med 2007;356:775–789.
35. Tashkin DP, Celli B, Senn S, Burkhart D, Kesten S, Menjoge S, Decramer M; UPLIFT Study Investigators. A 4-year trial of tiotropium in chronic obstructive pulmonary disease. N Engl J Med 2008;359:1543– 1554.
36. Barnes PJ. Emerging pharmacotherapies for COPD. Chest 2008;134: 1278–1286.
(Received in original form October 7, 2011; accepted in ﬁnal form March 5, 2012) SupportedbyGlaxoSmithKline(GSK study codeSCO104960). Editorial supportin
the form of copyediting and assembling ﬁgures was provided by Geoff Weller, Ph.D., of Gardiner-CaldwellCommunicationsand wasfundedbyGlaxoSmithKline. Author Contributions: All authors conceived and designed the study or analyzed
thedata;all authors contributed toandapprovedtheﬁnaldraftofthemanuscript; B.R.C.,P.B.,P.C.,H.C.,D.A.L.,W.M.,S.R.,E.S.,J.V.,E.W., andA.A.collected study data; N.L., J.Y., B.E.M., and L.D.E. conducted statistical analyses.
Correspondence and requests for reprints should be addressed to Bartolome
R. Celli,M.D.,Pulmonary andCritical CareDivision, BrighamandWomen’sHospital,
75 Francis Street, Boston, MA 02115. E-mail: firstname.lastname@example.org This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org
Am J Respir Crit Care Med Vol 185, Iss. 10, pp 1065–1072, May 15, 2012 Copyright ª 2012 by the American Thoracic Society Originally Published in Press as DOI: 10.1164/rccm.201110-1792OC on March 15, 2012 Internet address: www.atsjournals.org
Влезте или се регистрирайте безплатно, за да получите достъп до пълното съдържание и статиите на списанието в PDF формат.