Biomarkers, the Control Panel and Personalized COPD Medicine

bs_bs_banner

INVITED REVIEW SERIES: UNRAVELLING THE MANY FACES OF COPD TO OPTIMIZE ITS CARE AND OUTCOMES SERIES EDITORS: GREGORY G KING AND DON SIN

Biomarkers, the control panel and personalized COPD medicine

ALVAR AGUSTI,1,2,3 JOAQUIM GEA2,4 AND ROSA FANER2,3

1Thorax Institute, Hospital Clinic, University of Barcelona, 2Ciber Enfermedades Respiratorias (CIBERES), 3Thorax Institute, IDIBAPS, 4Respiratory Department, Hospital del Mar-IMIM. DCEXS, University Pompeu Fabra, Barcelona, Spain

ABSTRACT Key words: chronic bronchitis, emphysema, inflammation, network analysis, smoking. Chronic obstructive pulmonary disease (COPD) is a complex and heterogeneous disease, so its clinical Abbreviations: 6MWD, 6-min walk distance; ADAM19, management needs to be personalized as much as pos- metallopeptidase domain 19; ADO, age, dyspnoea and obstruc- sible. ‘Complex’ means that COPD has several compo- tion; AGER, advanced glycosylation end product-specific recep- nents with non-linear dynamic interactions, whereas tor; BICD1, bicaudal D homolog 1 (Drosophila); BMI, body mass ‘heterogeneous’ indicates that not all these compo- index; BODE, body mass index, airflow obstruction, dyspnoea nents are present in all patients or, in a given patient, at and exercise capacity; BMI, body mass index; CAT, Chronic all time points. This complexity and heterogeneity Obstructive Pulmonary Disease Assessment Test; CC16, explains and justifies the need for personalizing the secretoglobin, family 1A, member 1 (uteroglobin); CCL18, assessment and treatment of patients with COPD. We chemokine (C-C motif) ligand 18 (pulmonary and activation regu- propose here that the implementation of a ‘control lated); CHRNA3, cholinergic receptor, nicotinic, alpha 3 panel’ will facilitate the deployment of personalized (neuronal); CHRNA5, cholinergic receptor, nicotinic, alpha 5 COPD medicine in clinical practice. Such a control (neuronal); COPD, chronic obstructive pulmonary disease; CRP, panel should provide the practicing clinician comple- C-reactive protein; CVD, cardiovascular disease; CYP2A6, mentary and relevant information on treatable clinical cytochrome P450, family 2, subfamily A, polypeptide 6; DBH, characteristics in a single patient at a given time point. dopamine beta-hydroxylase (dopamine beta-monooxygenase); EBC, exhaled breath condensate; ECLIPSE, evaluation of COPD For the purpose of this discussion, we consider these longitudinally to identify predictive surrogate end-points; variables to be ‘biomarkers’. Which treatable clinical FAM13A, family with sequence similarity 13, member A; FEV1, characteristics should the COPD control panel include forced expiratory volume in 1 s; FFMI, the fat-free mass index; has not yet been formally validated. The review below FGF7, fibroblast growth factor 7; FTO, fat mass and obesity- suggests and discusses which ones might be considered associated; FVC, forced vital capacity; GOLD, global initiative for in the future and should be viewed as a working chronic obstructive lung disease; GWAS, genome-wide associa- proposal. tion studies; HHIP, hedgehog interacting protein; HLA-C, major histocompatibility complex, class I, C; IC/TLC, inspiratory to total lung capacity ratio; IREB2, iron-responsive element binding protein 2; LABA, long-acting b2 agonists; LAMA, long-acting anti- Correspondence: Alvar Agusti, Thorax Institute, Hospital muscarinic bronchodilators; miDNA, mitochondrial DNA; Clinic, Villarroel 170 (Escala 3, Planta 5), 08036 Barcelona, Spain. miRNA, microRNA; mMRC, modified Medical Research Council Email: [email protected] Dyspnoea Scale; NIH, National Institutes of Health; ncRNA, non- The Authors: Alvar Agusti is a Senior Consultant and Director of the Thorax Institute, Hospital Clinic, University of Barcelona coding RNA; PaO2, arterial oxygen pressure; SCGB1A1, (Spain). He is interested in translational COPD research and has secretoglobin, family 1A, member 1; SFTPD, surfactant protein D; published over 400 peer-reviewed articles. He is a member of the SGRQ, St George’s Respiratory; SNP, single nucleotide board of directors and scientific committee of GOLD. Joaquim polymorphisms; SP-D, surfactant protein D; sRAGE, soluble form Gea is Head of the Respiratory Medicine Department at Hospital of receptor for advanced glycation end products; VOC, volatile del Mar and Professor of Physiology at Universitat Pompeu organic compound; WBC, whole blood cell counts. Fabra, Barcelona (Spain). His main field of interest is muscle and exercise physiology in COPD, and has published around 290 peer-reviewed articles. He is the Principal Investigator of the INTRODUCTION project BIOMEPOC (novel Biomarkers in COPD). Rosa Faner is a Post-Doctoral Researcher at Fundació Privada Clinic, Barcelona (Spain). She holds a PhD in immunology, and her research inter- Chronic obstructive pulmonary disease (COPD) is the est is understanding the immunological basis of COPD. end result of a complex set of interactions between Received 10 April 2015; invited to revise 4 May 2015; revised 7 the environment and the genetic background of the May 2015; accepted 23 May 2015 individual (Fig. 1). As a result, it is often stated that

physiological or imaging measurements, according to the definition endorsed by the National Institutes of Health (NIH) in the USA, a biomarker is ‘a character- istic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to therapeutic interventions’, 10 so we use the term biomarker here in this broader sense (albeit we discuss in greater detail the more traditional ‘biological’ biomarkers). Like- wise, we organize the discussion below using the Figure 1 Relationships between the ‘exposome’ (a term that three ‘modules’ proposed originally in the description describes the ‘totality of human environmental exposures, from of the ‘COPD control panel’: severity, activity and conception onwards’,1 the genetic background of the individual impact (Fig. 2).9 (Genome), the Endotype (biological networks that enable and restrict reactions2) and the final clinical expression of the disease (Clinical phenotype). Large arrows indicate different therapeutic SEVERITY MODULE strategies. Reproduced with permission from Woodruff et al. (2015).3 For further explanation, see text. The severity of any disease (including COPD) relates to the ‘extent of functional impairment of the target organ(s)’.11,12 The ‘severity’ of COPD has been tradi- COPD is a complex and heterogeneous disease.4 Yet, tionally determined by the degree of airflow limita- the meaning of the terms ‘complex’ and ‘hetero- tion (assessed by measuring forced expiratory volume 13 geneous’ in this setting might not be immediately in the 1 s (FEV1)). FEV1 is a good estimate of the obvious. In our view, ‘complex’ means that COPD has pulmonary functional impairment, although other several components with non-linear dynamic inter- physiological measurements such as the inspiratory actions, whereas ‘heterogeneous’ indicates that not capacity (IC) to total lung capacity (TLC) ratio, arterial all these components are present in all patients or, in blood gases and exercise capacity provide comple- a given patient, at all time points.5 This complexity mentary information that also reflects the severity of and heterogeneity explains and justifies the need for COPD.14,15 Importantly, these are ‘treatable traits’ personalizing the assessment and treatment of because they can be improved by the use of one or patients with COPD.6 two long-acting bronchodilators, long-term oxygen It has been proposed that a first step towards per- therapy, chronic non-invasive ventilation and/or sonalized COPD medicine is the identification of rehabilitation, respectively.14 homogeneous groups of patients (i.e. clinical pheno- More recently, to account for the complexity of the types) with similar characteristics and outcomes.7 disease, a number of composite indexes have been This approach (the so-called stratified medicine) is proposed, including among others the BODE (body probably correct in a research setting, where the mass index (BMI), airflow obstruction, dyspnoea and investigation of mechanisms of disease or response to exercise capacity)16 and ADO (age, dyspnoea and therapy necessarily requires the study of homo- obstruction).17 They are also measurements of disease 6 geneous groups of patients. However, it may be diffi- severity and are probably better than the FEV1 in cult to implement in the clinic because these clinical order to establish the prognosis of a given patient. phenotypes may overlap in the same patient, and the However, in order to guide therapy in clinical practice, presence of one does not necessarily exclude the pres- they necessarily need to be deconstructed into their ence of a second (or a third) one.8 Likewise, the same original, treatable components. clinical phenotype could result from different biologi- Finally, although the previously mentioned defini- cal mechanisms (i.e. endotypes) (Fig. 1).3 tion of severity refers to the main target organ (the An alternative approach for the implementation of lungs in the case of COPD) it is also important to personalized COPD medicine in clinical practice is consider that comorbidities are extremely frequent in the use of a ‘control panel’ (Fig. 2).9 This approach these patients (cardiovascular and metabolic diseases proposes to use a number of biological and clinical in particular) that they may have a profound impact variables (organized in ‘modules’) that can provide in their health status and prognosis18 and that they the practicing clinician complementary and relevant seem to share molecular pathways with COPD.19 information for the proper management of the indi- Hence, although not strictly indicators of the severity vidual patient, either because of its prognostic impli- of the ‘pulmonary component’ of COPD, we propose cations and/or requirement for specific therapeutic that their presence should be noted and treated intervention. In other words, the control panel pro- appropriately, as the GOLD document recommends vides information on treatable clinical characteristics specifically.14 in a single patient at a given time point.8 These treatable clinical characteristics have not yet been formally validated. The review below discusses ACTIVITY MODULE which ones might be considered in the future. For the purpose of this discussion, we consider these vari- The ‘activity of a disease’ (please note that we are not ables to be ‘biomarkers’. Although biomarkers are referring here to the ‘activities of daily living’ of the generally considered to be substances rather than patient) relates to the ‘level of activation of the

Figure 2 Chronic obstructive pulmonary disease (COPD) control panel. Reproduced with permission from Agusti and MacNee (2013).9 For further explanations, see text. 6MWD, 6-min walk distance; CAT, Chronic Obstructive Pulmonary Disease Assess- ment Test; FEV1, forced expiratory volume in 1 s; IC/TLC, inspiratory to total lung capacity ratio; mMRC, modiﬁed Medical Research Council Dyspnoea Scale; PaO2, arterial oxygen pressure.

biological processes that drive disease progression’. 20 marker of disease activity may be the frequency of Although this concept is well established in other dis- exacerbations because they can occur at any level of eases (e.g. tuberculosis or rheumatoid arthritis), so far airflow limitation.28 Likewise, unintentional weight it has not been well defined in COPD.21 Yet, it is likely loss is associated with poor prognosis in COPD29 and that the treatment of a patient with a very ‘active’ could therefore be also considered a clinical marker of disease (as discussed below) should be very different disease activity. On the other hand, the identification from that of a patient with a ‘stable’ disease. For and validation of ‘biological’ biomarkers in COPD has instance, consider the case of tuberculosis. If your generated a great deal of interest over the past decade patient suffers ‘active’ tuberculosis, pharmacological or so.30 Table 1 presents a number of them, as treatment for several months would be indicated, reviewed recently.32 Main findings can be summa- whereas if, after appropriate diagnostic assessment, rized as follows. you conclude that he/she has ‘latent’ disease,22 treatment would not be indicated. Importantly, activity and severity do not necessarily Cellular biomarkers run in parallel. For instance, consider now the case of Despite that sputum neutrophilia is a relatively stable rheumatoid arthritis. In the early stages of the disease, biomarker in COPD, it does not seem to be a major activity may be high but severity mild. Under these surrogate of clinical or pathophysiological abnormal- circumstances, the rheumatologist will likely try to ities in COPD, hence limiting its potential clinical use the most appropriate treatment to reduce the application.32 High circulating leukocyte number is activity of the disease and prevent joint deterioration. weakly associated with frequent exacerbations and By contrast, in advanced disease there may be severe mortality both in COPD28,35 and in the general popu- joint deformities but low disease activity due to the lation.57 Yet, numeric differences between COPD spontaneous and/or therapeutically induced down- patients and controls are small and often within the regulation of the biological mechanisms that caused range of normal laboratory values. Other cellular it. Under these alternative circumstances, the man- biomarkers, such as the percentage of circulating agement of the patient is likely to be different. To date, eosinophils, merit further investigation.58 we are not considering these circumstances in the management of our COPD patients. This is, in part, because we have not really thought Protein biomarkers carefully enough about this and, in part, because Plasma fibrinogen is relatively stable over time and it what is/are the most appropriate markers of disease is significantly associated with symptoms, exercise activity in COPD are not resolved.20 Yet, several ‘clini- capacity, exacerbation rate, the BODE index and mor- cal’ candidate biomarkers of disease activity in COPD tality.35,36 It is currently being considered a potential can be conceived. For instance, the rate of FEV1 candidate for regulatory qualification as a prognostic decline or the rate of progression of emphysema are biomarker.37 Club-cell 16 (CC16)38 is weakly associ- obvious candidates, because recent research has ated with lung function decline,23 emphysema26 and 23–25 39 shown FEV1 decline rate (and emphysema pro- depression in COPD, and it has been recently gression26) varies greatly among patients with COPD. reported to have a protective role in COPD.59 Likewise, given that smoking is the major COPD risk Surfactant protein D (SP-D) was also weakly related factor, current smoking may also be considered a with exacerbations28,40 and, interestingly, it seems marker of disease activity.27 Another potential clinical to respond to oral and inhaled corticosteroid

Table 1 Summary of different COPD31 ‘biological’ biomarker studies published using the ECLIPSE database. Repro- duced with permission from Faner et al. (2014)32

Type of biomarker Main ﬁndings

Cellular biomarkers Sputum neutrophils 3.5% variability at 1-year follow up33

Sputum neutrophils Weak/absent association with FEV1%, SGRQ, emphysema, systemic inflammatory markers, exacerbation frequency or lung function decline33 Circulating WBC Associated with persistent systemic inflammation,34 frequent exacerbations28 and mortality35 Blood protein biomarkers Fibrinogen Significantly associated with symptoms, exercise capacity, exacerbation rate and the BODE index.35,36 Currently undergoing a regulatory qualification process37 CC16 Weakly associated with lung function decline, emphysema and depression23,26,38,39 SP-D Weak association with COPD exacerbations;28,40 sensitive to treatment with oral and inhaled corticosteroids40 CCL18 (PARC) Increased risk of cardiovascular hospitalization or mortality41 sRAGE Lower circulating sRAGE levels are associated with emphysema severity, and genetic polymorphisms in the AGER locus are associated with circulating sRAGE levels42 Inflammome Patients with persistent systemic inflammation (16%) had higher mortality and exacerbation rate than non-inflamed patients (30%).34 Systemic inflammation was also associated with heart disease, hypertension and diabetes43 Adipokines Leptin and adiponectin levels were (+) and (−) related to CRP, respectively; BMI and gender were the strongest determinants of both adipokines44 Vitamin D Low levels of vitamin D were related to emphysema, 6MWD, airways reactivity and CC-16 levels45 Gene studies Smoking history Suggestive associations identified for age at smoking initiation (chromosomes 2q21 and 6p21), lifetime mean number of cigarettes per day (CHRNA3/CHRNA5 and CYP2A6), current number of cigarettes smoked per day (CYP2A6) and smoking cessation (DBH)46 COPD susceptibility Several genomic regions associated with COPD susceptibility (FAM13A, HHIP, CHRNA3/CHRNA5/IREB2 and a region on chromosome 19). Other (ADAM19, FGF7, and SP-D) need replication in other populations COPD subtypes CHRNA3/5 significantly associated with pack-years, emphysema and airflow limitation;47 HHIP 47 not associated with pack-years, but related FEV1/FVC, lean body mass and exacerbations Emphysema Borderline genome-wide significant association with BICD148 49 49 Cachexia Suggestive association of BMI and FFMI with FTO gene. The latter also related to FEV1 Blood biomarkers Genome-wide significant associations identified only for CC16 (chromosome 11) and SP-D (SFTPD and SNP on chromosomes 6 and 16)50 Sputum transcriptomics Airflow limitation 277 genes associated with the severity of airflow limitation (GOLD grade)51 Emphysema 198 genes associated with the presence of emphysema51 Blood biomarkers SNP affecting circulating CC16 protein levels were significantly associated with sputum mRNA expression of SCGB1A1, the CC16 coding gene on chromosome 1150 COPD susceptibility An integrative genomics approach located potential functional variants in two genes located within a COPD GWAS locus on chromosome 15 (CHRNA5 and IREB2) and one locus in the HLA-C region on chromosome 652 Serum metabolomics Serum profile Results indicate that:53,54 (i) there is increased protein turnover in all COPD patients; (ii) increased protein degradation in individuals with emphysema and cachexia Exhaled breath condensate

pH Lower pH in COPD and smoking controls, but not related to FEV1 or sputum leukocyte counts, and not responsive to steroid treatment55 56 Adenosine/purines Increased concentrations in COPD patients, and adenosine levels correlated with FEV1

6MWD, 6-min walk distance; ADAM19, ADAM metallopeptidase domain 19; AGER, advanced glycosylation end product-specific receptor; BICD1, bicaudal D homolog 1 (Drosophila); BMI, body mass index; BODE, body mass index, airflow obstruction, dyspnoea and exercise capacity; CC16, Club-cell 16; CCL18, chemokine (C-C motif) ligand 18 (pulmonary and activation regulated); CHRNA3, cholinergic receptor, nicotinic, alpha 3 (neuronal); CHRNA5, cholinergic receptor, nicotinic, alpha 5 (neuronal); COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; CYP2A6, cytochrome P450, family 2, subfamily A, polypeptide 6; DBH, dopamine beta-hydroxylase (dopamine beta-monooxygenase); ECLIPSE, evaluation of COPD longitudinally to identify predictive surrogate end-points; FAM13A, family with sequence similarity 13, member A; FEV1, forced expiratory volume in 1 s; FFMI, the fat-free mass index; FGF7, fibroblast growth factor 7; FTO, fat mass and obesity-associated; FVC, forced vital capacity; GOLD, global initiative for chronic obstructive lung disease; GWAS, genome-wide association studies; HHIP, hedgehog interacting protein; HLA-C, major histocompatibility complex, class I, C; IREB2, iron-responsive element binding protein 2; SCGB1A1, secretoglobin, family 1A, member 1; SFTPD, surfactant protein D; SGRQ, St George’s Respiratory; SNP,single nucleotide polymorphisms; SP-D, surfactant protein D; sRAGE, soluble form of receptor for advanced glycation end products; WBC, whole blood counts.

Respirology (2015) © 2015 Asian Pacific Society of Respirology Personalized COPD medicine 5 treatment.40 Serum CCL18 was associated with A, polypeptide 6), current number of cigarettes increased risk of cardiovascular hospitalization and smoked per day (CYP2A6) and smoking cessation mortality.32 A combined panel (‘inflammome’) of four (SNP rs3025343 in DBH locus),46 strongly supporting a ‘biological’ biomarkers (circulating leukocytes, genetic basis for smoking initiation, maintenance and C-reactive protein, fibrinogen and interleukin-6) cessation.32 identified a group of COPD patients in the evaluation of COPD longitudinally to identify predictive surro- Genes associated with susceptibility to develop gate end-points (ECLIPSE) study with persistent COPD in smokers inflammation (16% of the cohort studied) who had a six times higher mortality during 3-year follow up as Several genomic regions are strongly associated with compared with those who were persistently non- the risk of developing COPD among smokers. These inflamed (30% of the cohort studied).34 The consid- include family with sequence similarity 13, member eration of these biomarkers improved the ability of A, hedgehog interacting protein, cholinergic receptor, well-established clinical variables to predict mortality nicotinic, alpha 3 (neuronal)/cholinergic receptor, in COPD.35 These results are in keeping with those of nicotinic, alpha 5 (neuronal), iron-responsive element binding protein 2 (FAM13A, HHIP,CHRNA3/ other studies that have also reported an association 32 between systemic inflammation and heart disease, CHRNA5/IREB2) and a region on chromosome 19. 43 Other genomic regions, including ADAM19, FGF7 and hypertension and diabetes. The two main 32 adipokines (leptin and adiponectin) appear to be SP-D, require replication in other studies. abnormally regulated in COPD and related to systemic inflammation, BMI and gender.44 In ECLIPSE, Genes associated with different COPD subtypes low levels of vitamin D in blood were related to the The potential genetic contributions to the phenotypic presence of emphysema, exercise capacity (as deter- heterogeneity of COPD are unclear, but recent studies mined by the 6-min walk distance), airways reactivity are beginning to illuminate them although, admit- and blood CC-16 levels.45 Finally, lower circulating tedly, many reported associations require replication levels of the soluble receptor for advanced glycation in additional cohorts. With this caveat in mind, end products (sRAGE) were associated with emphy- CHRNA3/5 has been associated with cumulative sema severity and a genetic polymorphisms in the smoking exposure (pack-years), presence of emphy- advanced glycosylation end product-specific receptor sema and severity of airflow limitation.47 HHIP is not (AGER) locus (the gene coding for RAGE).42 In associated with pack-years, but it is related to the summary, several systemic ‘biological’ biomarkers, FEV1/forced vital capacity ratio, with body composi- alone or in combination, have potential relevance for tion and COPD exacerbations in ECLIPSE.47 There is a the enrichment of clinical trials aimed at validating borderline genome-wide significant association future therapeutic interventions.32 between the presence of bicaudal D homolog 1 (Drosophila) (BICD1) SNP and emphysema.48 Given that variants in BICD1 are associated with telomere Genetic biomarkers length,48 this suggests that accelerated ageing can be a As Figure 1 illustrates, COPD is the end result of potential pathogenic mechanism of emphysema.61 A complex gene–environment interactions. It is not sur- SNP (rs8050136) located in the first intron of the fat prising, therefore, that over the last several years there mass and obesity-associated (FTO) gene is associated had been a great interest in studying and understand- 6,60 with the BMI and the fat-free mass index (FFMI) and, ing the genetics of COPD (Table 1). These studies interestingly, with the severity of airflow limitation.49 have actually investigated the genetic basis of a Likewise, several genetic determinants of the circulat- number of different but relevant aspects of COPD, ing levels of the protein biomarkers discussed above such as its relationship with smoking habits (the main have also been identified.50 Significant genome-wide risk of COPD), the susceptibility to develop the significant associations were identified for the blood- disease among smokers (only a proportion of them stream levels of CC16 (SCGB1A1 on chromosome 11, develop the disease) and/or the phenotypic heteroge- and another one located more than 20 Mb away on neity of COPD. Main results (Table 1) are discussed the same chromosome) and surfactant protein D below. (SP-D) (several SNP near its coding gene (SFTPD)).50 In addition, SNP on chromosomes 6 and 16 also dem- Genes associated with smoking onstrated genome-wide significant associations with Genome-wide association studies (GWAS) in four SP-D serum levels.50 Finally, very recently, it has been independent cohorts including more than 3000 ever- shown that COPD patients with different b2- smoking COPD patients failed to definitely identify adrenergic receptor (ADRB2) polymorphisms may genomic regions (as assessed by single nucleotide respond differently (at least in terms of reduction of polymorphisms (SNP) analysis) associated with exacerbations) to treatment with long-acting b2 smoking intensity or behaviour.46 However, suggestive agonists (LABA) or long-acting anti-muscarinic association were found for several loci associated bronchodilators (LAMA).62 If confirmed prospectively, with age at smoking initiation (chromosome 2q21 and this may have implications for the personalized treat- chromosome 6p21), lifetime mean number of ciga- ment of COPD. Likewise, resistance to steroids has rettes per day (cholinergic receptor, nicotinic, alpha 3 been shown to be associated with activation of a (neuronal)/cholinergic receptor, nicotinic, alpha 5 number of different kinases, which in turn may be the (neuronal) and cytochrome P450, family 2, subfamily consequence of phosphatase gene down-regulation.63

Epigenetic biomarkers breath prints can identify the presence of airway bac- Epigenetic biomarkers, including histone methy- terial colonization in clinically stable patients with lation and miRNAs, have also been investigated in COPD.70 COPD. It has been reported that gene promoters are hypo-methylated in COPD patients as compared with non-smokers or smokers with normal spirometry.64,65 IMPACT MODULE Likewise, both in lung tissue and plasma of COPD patients, expression changes in specific miRNAs The impact of any disease depends on how the acting on inflammatory mediators have been also patient perceives both the severity and the activity of described,66,67 albeit clearly more work is required the disease (as defined above) and how the patient in this area to validate and contextualize these modifies the ‘activities of daily living’ (please note that observations. we are not talking here about the ‘biological activity’ of the disease, discussed above). This perception Sputum and systemic transcriptomic varies substantially in patients with asthma31 and is biomarkers therefore likely to do the same in patients with COPD, Integrative genomics approach of sputum trans- although this has not been formally demonstrated to criptomics has the potential to identify potential our knowledge. On the contrary, somewhat naïvely, COPD susceptibility loci.52 As shown in Table 1, we have traditionally assumed that mild COPD (as sputum transcriptomics studies in the ECLIPSE assessed by the severity of airflow limitation) has a cohort identified: (i) 277 genes associated with airflow minor impact on the patient, whereas the impact is limitation and 198 with emphysema;51 and, (ii) an much greater in severe disease. Yet, the systematic use association of mRNA expression of CC16 and several of instruments that measure such an impact, like the SNP influencing its circulating levels.50 Similarly, a St George’s Respiratory Questionnaire, has clearly gene expression signature in frequent exacerbators shown that this assumption was wrong, because the has been reported in circulating leukocytes, which relationship between FEV1 and health status is poor 4 may be useful to help in understanding the molecular and individual variability is enormous (Fig. 3). In pathogenesis of COPD exacerbations.68 fact, it is precisely because of this variability that it is recommended in the GOLD 2011 document to determine the level of symptoms as a key component of the Serum metabolic biomarkers current assessment of COPD patients.14 Similar argu- Several ECLIPSE studies investigated the serum ments can be applied to other relevant domains of metabolomic profile of patients with COPD (Table 1). COPD, such as exacerbations because a number of patients suffer frequent exacerbations,28 other appar- Main results indicate that: (i) there is increased 71 protein turnover in all COPD patients, particularly in ently have ‘unreported exacerbations’ and others have severely impaired lung function and, neverthe- those with emphysema and cachexia; and, (ii) 72 whereas there are some promising metabolomic less, do not apparently have exacerbations. All in all, these observations suggest that as described in signals detected in the serum of certain subtypes of 31 COPD patients, replication in other cohorts is asthma, it is likely that the perception of symptoms required.32 vary in different COPD patients. This hypothesis deserves specific research because of its potential therapeutic implications, particularly in ‘frequent 28 Exhaled breath condensate and electronic exacerbators’. It is also worth mentioning that activities of daily living are impaired in some, but not all, nose biomarkers patients with COPD,73 and that objectively measured Despite initial enthusiasm for the exhaled breath con- physical activity is a strong predictor of all-cause mor- densate (EBC) as a potential source of valid COPD 74 32 tality in patients with COPD. Considering all of the biomarkers, results have been mostly disappointing. above arguments, we propose the need of an impact Except for the pH value of EBC, which was signifi- module in the control panel (Fig. 2), albeit what are cantly lower in patients with COPD and smoking con- the optimal questionnaires and activity monitor to trols than in non-smokers (albeit not related to use requires validation. airflow limitation or sputum neutrophils,55 using con- ventional methodology, investigators in the ECLIPSE study were not able to reliably measure protein biomarkers in EBC. They found, however, that the TOWARDS PERSONALIZED COPD relative concentrations of adenosine and adenosine MEDICINE: THE CONTROL PANEL AS A monophosphate were elevated in COPD patients, WAY TO MOVE FORWARD the former being correlated with airflow limitation severity.56 A recent symposium held in Barcelona reviewed On the other hand, exhaled breath contains a the critical steps required to move forward towards complex mixture of volatile organic compounds personalized respiratory medicine.6 To address the (VOC), some of which could potentially represent biological complexity of COPD (Fig. 1), these include, biomarkers for lung diseases.69 In fact, very recent among others, the use of network analysis75 and studies suggest that an electronic nose, a novel non- an extremely close collaboration between basic, clini- invasive technology capable of distinguishing VOC cal and computer scientists.76 The potential for

Figure 3 The relationship between the severity of airﬂow limitation (forced expiratory volume in 1 s, FEV1) and health status (as assessed by St George Respiratory Questionnaire) is statistically signiﬁcant at the population level but individual variability is enormous, indicating that health status cannot be predicted from the FEV1 value in a single patient. Reproduced with permission from Agusti et al. (2010).4 For further explanations, see text.

translation of all this new knowledge to clinical practice is illustrated in Figure 4 (right column). A better understanding of the genetic level will identify (as discussed above) genetic markers that can help clinicians to better determine the future risk of a given patient (risk of deterioration of the disease and/or development of complications and/or response to a given therapy).21 The biological level will eventually provide relevant information on ‘biological’ markers (also as discussed above) that can help clinicians in their assessment and monitoring of a given patient. A deeper understanding of the relationships between different comorbidities at the clinical level19,77 can 78,79 facilitate better strategies for integrated care, the Figure 4 Diagram illustrating the network relationships and dif- development of so-called Clinical Decision Support ferent levels of complexity of chronic obstructive pulmonary Systems (CDSS),6 a generic name for the COPD control disease (COPD; environment, clinical, biological and genetic), as panel discussed here9 and, eventually, the develop- well as the potential outcomes of clinical relevance (right-hand ment of a network of guidelines that facilitate column) that each of these levels can be envisaged to deliver the assessment and treatment of COPD patients. when properly understood. Each level only shows some of its Finally, the ‘exposome’ (a term that describes potential components in order to illustrate the concept (the diagram is not intended to be comprehensive). Likewise, links the ‘totality of human environmental exposures, 1 between the different elements of the network are drawn for from conception onwards’ ) also forms a complex illustrative purposes also and do not necessarily reﬂect evidence- network that merits further research and, eventually, based relationships. For further explanations, see text. CVD, car- intervention. diovascular disease; GWAS, genome-wide association studies; We propose that the ‘COPD control panel’ provides miDNA, mitochondrial DNA; miRNA, microRNA; ncRNA, non- a way to visualize the complexity of COPD and that coding RNA. Reproduced with permission from Agusti and the combined assessment of the severity, impact and Vestbo (2011).21 For further explanations, see text.

© 2015 Asian Pacific Society of Respirology Respirology (2015) 8 A Agusti et al. activity can best inform the physician on the most obstructive pulmonary disease phenotypes: the future of COPD. appropriate management strategies for an individual Am. J. Respir. Crit. Care Med. 2010; 182: 598–604. patient based on the ‘treatable traits’ present in this 8 Agusti A. Phenotypes and disease characterization in chronic particular patient at this specific time point.9 Need- obstructive pulmonary disease. Toward the extinction of pheno- less to say that this proposal needs prospective types? Ann Am Thorac Soc. 2013; 10: S125–30. 9 Agusti A, MacNee W. The COPD control panel: towards person- research and validation because there are a very large alised medicine in COPD. Thorax 2013; 68: 687–90. number of possible factors that can be included into 10 Jones PW, Agusti AGN. Outcomes and markers in the assessment this model of disease characterization with the poten- of chronic obstructive pulmonary disease. Eur.Respir.J. 2006; 27: tial to inform clinical care. Hence, it is important to 822–32. avoid redundancy and duplication so such a compre- 11 Agusti A, Sobradillo P, Celli B. Addressing the complexity of hensive system will eventually be cut down into its chronic obstructive pulmonary disease: from phenotypes and critical components. To this end, systems biology, biomarkers to scale-free networks, systems biology, and P4 network analysis and computer modellers are likely to medicine. Am. J. Respir. Crit. Care Med. 2011; 183: 1129–37. be of great use.75,80 Importantly, the control panel can 12 Agusti A, Celli B. Avoiding confusion in COPD: from risk factors to phenotypes to measures of disease characterisation. Eur. be customized to the needs of the patient and the Respir. J. 2011; 38: 749–51. resources available locally (e.g. rural vs urban health 13 Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, care centres; primary vs specialized care). Fukuchi Y, Jenkins C, Rodriguez-Roisin R, Van Weel C et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive CONCLUSIONS summary. Am. J. Respir. Crit. Care Med. 2007; 176: 532–55. 14 Vestbo J, Hurd SS, Agusti AG, Jones PW, Vogelmeier C, Anzueto A, There is no such thing as ‘the’ COPD biomarker. Dif- Barnes PJ, Fabbri LM, Martinez FJ, Nishimura M et al. Global ferent biomarkers will be needed to assess different strategy for the diagnosis, management and prevention of components of this complex and heterogeneous chronic obstructive pulmonary disease, GOLD executive disease.11 Combining them in a ‘control panel’ has the summary. Am. J. Respir. Crit. Care Med. 2013; 187: 347–65. potential to facilitate their translation to clinical prac- 15 Casanova C, Cote C, de Torres JP, Aguirre-Jaime A, Marin JM, tice.9 By doing so, this strategy has the potential to Pinto-Plata V, Celli BR. Inspiratory-to-total lung capacity move COPD management towards personalized (pre- ratio predicts mortality in patients with chronic obstructive 5,6,81 pulmonary disease. Am. J. Respir. Crit. Care Med. 2005; 171: cision) medicine. 591–7. 16 Celli BR, Cote CG, Marin JM, Casanova C, Montes de Oca M, Mendez RA, Pinto Plata V, Cabral HJ. The body-mass index, Acknowledgements airflow obstruction, dyspnea, and exercise capacity index in The authors thank Dr. W. MacNee (Professor of Respiratory and chronic obstructive pulmonary disease. N. Engl. J. Med. 2004; Environmental Medicine/Honorary Consultant Physician, Univer- 350: 1005–12. sity of Edinburgh, MRC Centre for Inflammation Research) and 17 Puhan MA, Garcia-Aymerich J, Frey M, Ter RG, Anto JM, Agusti the numerous colleagues who, over the past few years, have AG, Gomez FP, Rodriguez-Roisin R, Moons KG, Kessels AG et al. contributed with their ideas and criticisms to the development Expansion of the prognostic assessment of patients with chronic and refinement of the concept of a ‘COPD control panel’. This obstructive pulmonary disease: the updated BODE index and study was supported, in part, by Instituto de Salud Carlos III the ADO index. Lancet 2009; 374: 704–11. (PI12/01117), Recercaixa-2012 (AA084096), SEPAR (PI065/2013, 18 Mannino DM, Thorn D, Swensen A, Holguin F. Prevalence and PI192/2012), FUCAP 2012 and SAF-2011-26908. outcomes of diabetes, hypertension and cardiovascular disease in COPD. Eur. Respir. J. 2008; 32: 962–9. 19 Grosdidier S, Ferrer A, Faner R, Pinero J, Roca J, Cosio B, Agusti A, REFERENCES Gea J, Sanz F, Furlong LI. Network medicine analysis of COPD multimorbidities. Respir. Res. 2014; 15: 111. 1 Vrijheid M. The exposome: a new paradigm to study the impact 20 Vestbo J, Rennard S. Chronic obstructive pulmonary disease of environment on health. Thorax 2014; 69: 876–8. biomarker(s) for disease activity needed–urgently. Am. J. Respir. 2 Lotvall J, Akdis CA, Bacharier LB, Bjermer L, Casale TB, Custovic Crit. Care Med. 2010; 182: 863–4. A, Lemanske RF Jr, Wardlaw AJ, Wenzel SE, Greenberger PA. 21 Agusti A, Vestbo J. Current controversies and future perspectives Asthma endotypes: a new approach to classification of disease in chronic obstructive pulmonary disease. Am. J. Respir. Crit. entities within the asthma syndrome. J. Allergy Clin. Immunol. Care Med. 2011; 184: 507–13. 2011; 127: 355–60. 22 Vilaplana C, Cardona PJ. The lack of a big picture in tuberculosis: 3 Woodruff PG, Agusti A, Roche N, Singh D, Martinez FJ. Current the clinical point of view, the problems of experimental model- concepts in COPD pharmacological treatment: the GOLD (r)evo- ing and immunomodulation. The factors we should consider lution and endotype-driven therapy. Lancet 2015 (in press). when designing novel treatment strategies. Front Microbiol. 4 Agusti A, Calverley P, Celli B, Coxson H, Edwards L, Lomas D, 2014; 5: 55. MacNee W, Miller B, Rennard S, Silverman E et al. 23 Vestbo J, Edwards LD, Scanlon PD, Yates JC, Agusti A, Bakke P, Characterisation of COPD heterogeneity in the ECLIPSE cohort. Calverley PMA, Celli B, Coxson HO, Crim C et al. Changes in Respir. Res. 2010; 11: 122–36. forced expiratory volume in 1 second over time in COPD. N. Engl. 5 Agusti A. The path to personalized medicine in COPD. Thorax J. Med. 2011; 365: 1184–92. 2014; 69: 857–64. 24 Jenkins CR, Jones PW, Calverley PM, Celli B, Anderson JA, 6 Agusti A, Anto JM, Auffray C, Barbe F, Barreiro E, Dorca J, Ferguson GT, Yates JC, Willits LR, Vestbo J. Efficacy of salmeterol/ Escarrabill J, Faner R, Furlong LI, Garcia-Aymerich J et al. Per- fluticasone propionate by GOLD stage of chronic obstructive sonalized respiratory medicine: exploring the horizon, address- pulmonary disease: analysis from the randomised, placebo- ing the issues. Am. J. Respir. Crit. Care Med. 2015; 191: 391–401. controlled TORCH study. Respir. Res. 2009; 10: 59. 7 Han MK, Agusti A, Calverley PM, Celli BR, Criner G, Curtis JL, 25 Decramer M, Celli B, Kesten S, Lystig T, Mehra S, Tashkin DP. Fabbri LM, Goldin JG, Jones PW, MacNee W et al. Chronic Effect of tiotropium on outcomes in patients with moderate

chronic obstructive pulmonary disease (UPLIFT): a prespecified 43 Miller J, Edwards LD, Agusti A, Bakke P, Calverley PM, Celli B, subgroup analysis of a randomised controlled trial. Lancet 2009; Coxson HO, Crim C, Lomas DA, Miller BE et al. Comorbidity, 374: 1171–8. systemic inflammation and outcomes in the ECLIPSE cohort. 26 Coxson HO, Dirksen A, Edwards L, Yates JC, Agusti A, Bakke P, Respir. Med. 2013; 107: 1376–84. Calverley PMA, Celli B, Crim C, Lomas DA et al. The presence 44 Breyer MK, Rutten EP, Locantore NW, Watkins ML, Miller BE, and progression of emphysema in COPD as determined by CT Wouters EF. Dysregulated adipokine metabolism in chronic scanning and biomarker expression: a prospective analysis from obstructive pulmonary disease. Eur. J. Clin. Invest. 2012; 42: 983– the ECLIPSE study. Lancet Respir Med. 2013; 1: 129–36. 91. 27 Anthonisen NR, Connett JE, Murray RP. Smoking and lung func- 45 Berg I, Hanson C, Sayles H, Romberger D, Nelson A, Meza J, tion of Lung Health Study participants after 11 years. Am. J. Miller B, Wouters EF, MacNee W, Rutten EP et al. Vitamin D, Respir. Crit. Care Med. 2002; 166: 675–9. vitamin D binding protein, lung function and structure in COPD. 28 Hurst JR, Vestbo J, Anzueto A, Locantore N, Mullerova H, Respir. Med. 2013; 107: 1578–88. Tal-Singer R, Miller B, Lomas DA, Agusti A, MacNee W et al. Sus- 46 Siedlinski M, Cho MH, Bakke P, Gulsvik A, Lomas DA, Anderson ceptibility to exacerbation in chronic obstructive pulmonary W, Kong X, Rennard SI, Beaty TH, Hokanson JE et al. Genome- disease. N. Engl. J. Med. 2010; 363: 1128–38. wide association study of smoking behaviours in patients with 29 Schols AM, Broekhuizen R, Weling-Scheepers CA, Wouters EF. COPD. Thorax 2011; 66: 894–902. Body composition and mortality in chronic obstructive pulmo- 47 Pillai SG, Kong X, Edwards LD, Cho M, Anderson WH, Coxson nary disease. Am. J. Clin. Nutr. 2005; 82: 53–9. HO, Lomas DA, Silverman EK. Loci identified by genome-wide 30 Agusti A, Sin DD. Biomarkers in COPD. Clin. Chest Med. 2014; 35: association studies influence different disease-related pheno- 131–41. types in COPD. Am. J. Respir. Crit. Care Med. 2010; 182: 31 Davis SQ, Permutt Z, Permutt S, Naureckas ET, Bilderback AL, 1498–505. Rand CS, Stein BD, Krishnan JA. Perception of airflow obstruc- 48 Kong X, Cho MH, Anderson W, Coxson HO, Muller N, Washko G, tion in patients hospitalized for acute asthma. Ann. Allergy Hoffman EA, Bakke P, Gulsvik A, Lomas DA et al. Genome-wide Asthma Immunol. 2009; 102: 455–61. association study identifies BICD1 as a susceptibility gene for 32 Faner R, Tal-Singer R, Riley JH, Celli B, Vestbo J, MacNee W, emphysema. Am. J. Respir. Crit. Care Med. 2010; 183: 43–9. Bakke P, Calverley PMA, Coxson H, Crim C et al. Lessons 49 Wan ES, Cho MH, Boutaoui N, Klanderman BJ, Sylvia JS, Ziniti JP, from ECLIPSE: a review of COPD biomarkers. Thorax 2014; 69: Won S, Lange C, Pillai SG, Anderson WH et al. Genome-wide 666–72. association analysis of body mass in chronic obstructive pulmo- 33 Singh D, Edwards L, Tal-Singer R, Rennard S. Sputum nary disease. Am. J. Respir. Cell Mol. Biol. 2011; 45: 304–10. neutrophils as a biomarker in COPD: findings from the ECLIPSE 50 Kim DK, Cho MH, Hersh CP, Lomas DA, Miller BE, Kong X, Bakke study. Respir. Res. 2010; 11: 77. P, Gulsvik A, Agusti A, Wouters E et al. Genome-wide association 34 Agusti A, Edwards LD, Rennard SI, MacNee W, Tal-Singer R, analysis of blood biomarkers in chronic obstructive pulmonary Miller BE, Vestbo J, Lomas DA, Calverley PMA, Wouters E et al. disease. Am. J. Respir. Crit. Care Med. 2012; 186: 1238–47. persistent systemic inflammation is associated with poor clinical 51 Singh D, Fox SM, Tal-Singer R, Plumb J, Bates S, Broad P, Riley JH, outcomes in COPD: a novel phenotype. PLoS ONE 2012; 7: Celli B. Induced sputum genes associated with spirometric and e37483. radiological disease severity in COPD ex-smokers. Thorax 2011; 35 Celli BR, Locantore N, Yates J, Tal-Singer R, Miller BE, Bakke P, 66: 489–95. Calverley P, Coxson H, Crim C, Edwards LD et al. Inflammatory 52 Qiu W, Cho MH, Riley JH, Anderson WH, Singh D, Bakke P, biomarkers improve clinical prediction of mortality in chronic Gulsvik A, Litonjua AA, Lomas DA, Crapo JD et al. Genetics of obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. sputum gene expression in chronic obstructive pulmonary 2012; 185: 1065–72. disease. PLoS ONE 2011; 6: e24395. 36 Duvoix A, Dickens J, Haq I, Mannino D, Miller B, Tal-Singer R, 53 Ubhi BK, Riley JH, Shaw PA, Lomas DA, Tal-Singer R, MacNee W, Lomas DA. Blood fibrinogen as a biomarker of chronic obstruc- Griffin JL, Connor SC. Metabolic profiling detects biomarkers of tive pulmonary disease. Thorax 2013; 68: 670–6. protein degradation in COPD patients. Eur. Respir. J. 2012; 40: 37 Casaburi R, Celli B, Crapo J, Criner G, Croxton T, Gaw A, 345–55. Kline-Leidy N, Lomas DA, Merrill D, Polkey MI et al. The COPD 54 Ubhi BK, Cheng KK, Dong J, Janowitz T, Jodrell D, Tal-Singer R, Biomarker Qualification Consortium (CBQC). COPD 2013; 10: MacNee W, Lomas DA, Riley JH, Griffin JL et al. Targeted 367–77. metabolomics identifies perturbations in amino acid metabo- 38 Lomas DA, Silverman EK, Edwards LD, Miller BE, Coxson HO, lism that sub-classify patients with COPD. Mol. Biosyst. 2012; 8: Tal-Singer R. Evaluation of serum CC-16 as a biomarker for 3125–33. COPD in the ECLIPSE cohort. Thorax 2008; 63: 1058–63. 55 MacNee W, Rennard SI, Hunt JF, Edwards LD, Miller BE, 39 Hanania NA, Mullerova H, Locantore NW, Vestbo J, Watkins ML, Locantore NW, Tal-Singer R. Evaluation of exhaled breath con- Wouters EF, Rennard SI, Sharafkhaneh A. Determinants of densate pH as a biomarker for COPD. Respir. Med. 2011; 105: depression in the ECLIPSE chronic obstructive pulmonary 1037–45. disease cohort. Am. J. Respir. Crit. Care Med. 2011; 183: 604–11. 56 Esther CR Jr, Lazaar AL, Bordonali E, Qaqish B, Boucher RC. 40 Lomas DA, Silverman EK, Edwards LD, Locantore NW, Miller BE, Elevated airway purines in COPD. Chest 2011; 140: 954–60. Horstman DH, Tal-Singer R, Evaluation of COPD Longitudinally 57 Thomsen M, Ingebrigtsen TS, Marott JL, Dahl M, Lange P, Vestbo to Identify Predictive Surrogate Endpoints study investigators. J, Nordestgaard BG. Inflammatory biomarkers and exacerba- Serum surfactant protein D is steroid sensitive and associated tions in chronic obstructive pulmonary disease. JAMA 2013; 309: with exacerbations of COPD. Eur. Respir. J. 2009; 34: 95–102. 2353–61. 41 Sin DD, Miller BE, Duvoix A, Man SFP, Zhang X, Silverman EK, 58 Singh D, Kolsum U, Brightling CE, Locantore N, Agusti A, Connett JE, Anthonisen NA, Wise RA, Tashkin D et al. Serum Tal-Singer R. Eosinophilic inflammation in COPD: prevalence PARC/CCL-18 concentrations and health outcomes in chronic and clinical characteristics. Eur. Respir. J. 2014; 44: 1697–700. obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. 59 Laucho-Contreras ME, Polverino F, Gupta K, Taylor KL, Kelly E, 2011; 183: 1187–92. Pinto-Plata V, Divo M, Ashfaq N, Petersen H, Stripp B et al.Pro- 42 Cheng DT, Kim DK, Cockayne DA, Belousov A, Bitter H, Cho MH, tective role for club cell secretory protein-16 (CC16) in the devel- Duvoix A, Edwards LD, Lomas DA, Miller BE et al. Systemic opment of chronic obstructive pulmonary disease. Eur. Respir. J. sRAGE is a biomarker of emphysema and associated with AGER 2015; 45: 1544–56. genetic variants in COPD Patients. Am. J. Respir. Crit. Care Med. 60 Lomas DA, Silverman EK. The genetics of chronic obstructive 2013; 188: 948–57. pulmonary disease. Respir. Res. 2001; 2: 20–6.

61 Faner R, Rojas M, MacNee W, Agusti A. Abnormal lung aging in 76 Bauer-Mehren A, Furlong LI, Sanz F. Pathway databases and chronic obstructive pulmonary disease and idiopathic pulmo- tools for their exploitation: benefits, current limitations and nary fibrosis. Am. J. Respir. Crit. Care Med. 2012; 186: 306–13. challenges. Mol. Syst. Biol. 2009; 5: 290–6. 62 Rabe KF, Fabbri LM, Israel E, Kogler H, Riemann K, Schmidt H, 77 Divo M, Cote C, de Torres JP, Casanova C, Marin JM, Pinto-Plata Glaab T, Vogelmeier CF. Effect of ADRB2 polymorphisms on the V, Zulueta J, Cabrera C, Zagaceta J, Hunninghake G et al. efficacy of salmeterol and tiotropium in preventing COPD exac- Comorbidities and risk of mortality in patients with chronic erbations: a prespecified substudy of the POET-COPD trial. obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. Lancet Respir Med. 2014; 2: 44–53. 2012; 186: 155–61. 63 Barnes PJ. Corticosteroid resistance in patients with asthma and 78 Bousquet J, Anto J, Sterk P, Adcock I, Chung K, Roca J, Agusti A, chronic obstructive pulmonary disease. J. Allergy Clin. Immunol. Brightling C, Cambon-Thomsen A, Cesario A et al. Systems 2013; 131: 636–45. medicine and integrated care to combat chronic noncommuni- 64 Qiu W, Baccarelli A, Carey VJ, Boutaoui N, Bacherman H, cable diseases. Genome Med. 2011; 3: 43–55. Klanderman B, Rennard S, Agusti A, Anderson W, Lomas DA 79 Bousquet J, Addis A, Adcock I, Agache I, Agusti A, Alonso A, et al. Variable DNA methylation is associated with chronic Annesi-Maesano I, Anto JM, Bachert C, Baena-Cagnani CE, et al. obstructive pulmonary disease and lung function. Am. J. Respir. Integrated care pathways for airway diseases (AIRWAYS-ICPs). Crit. Care Med. 2012; 185: 373–81. Eur. Respir. J. 2014; 44: 304–23. 65 Wan ES, Qiu W, Baccarelli A, Carey VJ, Bacherman H, Rennard SI, 80 Barabasi AL, Gulbahce N, Loscalzo J. Network medicine: a Agusti A, Anderson WH, Lomas DA, DeMeo DL. Systemic steroid network-based approach to human disease. Nat. Rev. Genet. exposure is associated with differential methylation in chronic 2011; 12: 56–68. obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. 81 Mirnezami R, Nicholson J, Darzi A. Preparing for precision medi- 2012; 186: 1248–55. cine. N. Engl. J. Med. 2012; 366: 489–91. 66 Ezzie ME, Crawford M, Cho JH, Orellana R, Zhang S, Gelinas R, Batte K, Yu L, Nuovo G, Galas D et al. Gene expression networks in COPD: microRNA and mRNA regulation. Thorax 2012; 67: 122–31. 67 Wang M, Huang Y, Liang Z, Liu D, Lu Y, Dai Y, Feng G, APPENDIX I: GLOSSARY OF TERMS Wang C. Plasma miRNAs might be promising biomarkers of Term Definition chronic obstructive pulmonary disease. Clin Respir J. 2014; doi: 10.1111/crj.12194. [Epub ahead of print]. Biomarkers Variables providing relevant 68 Morrow JD, Qiu W, Chhabra D, Rennard SI, Belloni P, Belousov A, information on (treatable) clinical Pillai SG, Hersh CP. Identifying a gene expression signature of frequent COPD exacerbations in peripheral blood using network characteristics. methods. BMC Med Genomics. 2015; 8:1. Clinical phenotypes Groups of patients with similar 69 Basanta M, Jarvis RM, Xu Y, Blackburn G, Tal-Singer R, Woodcock characteristics and outcomes A, Singh D, Goodacre R, Thomas CL, Fowler SJ. Non-invasive (stratified medicine). metabolomic analysis of breath using differential mobility spec- Control panel Biological and clinical variables that trometry in patients with chronic obstructive pulmonary disease provide complementary and and healthy smokers. Analyst 2010; 135: 315–20. relevant information for the 70 Sibila O, Garcia-Bellmunt L, Giner J, Merino JL, Suarez-Cuartin management of the individual G, Torrego A, Solanes I, Castillo D, Valera JL, Cosio BG et al. patient. Identification of airway bacterial colonization by an electronic nose in chronic obstructive pulmonary disease. Respir. Med. Modules Group of related variables. 2014; 108: 1608–14. Severity module Treatable traits related to the extent 71 Seemungal TA, Donaldson GC, Bhowmik A, Jeffries DJ, Wedzicha of functional impairment of the JA. Time course and recovery of exacerbations in patients with lung and the presence of chronic obstructive pulmonary disease. Am. J. Respir. Crit. Care comorbidities. Med. 2000; 161: 1608–13. Activity module Treatable traits related to the level of 72 Garcia-Aymerich J, Gomez FP, Benet M, Farrero E, Basagana X, activation of the biological Gayete A, Pare C, Freixa X, Ferrer J, Ferrer A et al. Identification processes that drive disease and prospective validation of clinically relevant chronic obstruc- progression, including clinical and tive pulmonary disease (COPD) subtypes. Thorax 2011; 66: biological biomarkers. 430–7. 73 Watz H, Waschki B, Meyer T, Magnussen H. Physical activity in Impact module Patient’s perception of severity and patients with COPD. Eur. Respir. J. 2009; 33: 262–72. activity modules. 74 Waschki B, Kirsten A, Holz O, Muller KC, Meyer T, Watz H, Exposome Totality of human environmental Magnussen H. Physical activity is the strongest predictor of all- exposures, from conception cause mortality in patients with COPD: a prospective cohort onwards. study. Chest 2011; 140: 331–42. 75 Diez D, Agusti A, Wheelock CE. Network analysis in the investigation of chronic respiratory diseases: from basics to application. Am. J. Respir. Crit. Care Med. 2014; 190: 981–8.