Millares et al. BMC Pulmonary Medicine (2019) 19:112 https://doi.org/10.1186/s12890-019-0867-x RESEARCH ARTICLE Open Access Relationship between the respiratory microbiome and the severity of airflow limitation, history of exacerbations and circulating eosinophils in COPD patients Laura Millares1,2,3, Sergi Pascual4, Concepción Montón3,5,6,7, Marian García-Núñez1,2,3, Cristina Lalmolda1,2, Rosa Faner2,7, Carme Casadevall2,4,8, Laia Setó1,2, Silvia Capilla9, Amàlia Moreno5, Ady Angélica Castro-Acosta2,10, Carlos José Alvarez-Martinez2,10, Oriol Sibila7,11, Germán Peces-Barba2,12, Borja G. Cosio2,13, Alvar Agustí2,7,14, Joaquim Gea2,4,7,8 and Eduard Monsó2,5,7,15* Abstract Background: The respiratory microbiome is altered in COPD patients but its relationship with core components of the disease, such as the severity of airflow limitation, the frequency of exacerbations or the circulating levels of eosinophils, is unclear. Methods: Cross-sectional study comprising 72 clinically stable COPD patients (mean age 68 [SD 7.9] years; FEV1 48.7 [SD 20.1]% of reference) who provided spontaneous sputum samples for 16S rRNA gene amplification and sequencing. The microbiome composition was analysed with QIIME. Results: We observed that: (1) more severe airflow limitation was associated with reduced relative abundance (RA) of Treponema and an increase in Pseudomonas; (2) patients with ≥2 exacerbations the previous year showed a significantly different bacterial community with respect to non-exacerbators (p = 0.014), with changes in 13 genera, including an increase of Pseudomonas, and finally, (3) peripheral eosinophils levels ≥2% were associated with more diverse microbiome [Chao1 224.51 (74.88) vs 277.39 (78.92) p = 0.006; Shannon 3.94 (1.05) vs 4.54 (1.06) p = 0.020], and a significant increase in the RAs of 20 genera. Conclusion: The respiratory microbiome in clinically stable COPD patients varies significantly according to the severity of airflow limitation, previous history of exacerbations and circulating eosinophils levels. Keywords: Bacterial community, Diversity, Eosinophils, Exacerbations, Sputum, Stable COPD Summary at a glance Background Core components of COPD such as airflow limi- Chronic obstructive pulmonary disease (COPD) is a het- tation, history of previous exacerbations and level of erogeneous disease [1–3]. The study of the respiratory circulating eosinophils have an impact in the microbiome in COPD has revealed a specific bacterial bronchial respiratory microbiome of clinically stable community composition in these patients [4, 5]. COPD patients. However, the relationship between this microbiome and core components of the disease, such as, the severity of the airflow limitation and the type of treatment received * Correspondence: emonso@tauli.cat remains unclear. In addition, changes in the microbiome 2 Centro de Investigación Biomédica en Red de Enfermedades Respiratorias have been described in COPD exacerbations [6, 7], but it (CIBERES), Instituto de Salud Carlos III (ISCIII), Madrid, Spain 5 is not known if differences between patients who suffer Department of Respiratory Medicine, Hospital Universitari Parc Taulí, Parc Taulí 1, 08208, Sabadell, Barcelona, Spain two or more exacerbations per year, who are considered Full list of author information is available at the end of the article frequent exacerbators [8, 9], and non-exacerbators can © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 2 of 9 be detected during clinical stability. Likewise, levels of Sample collection, DNA extraction, PCR amplification and circulating eosinophils ≥2% in clinically stable patients 16S sequencing identifies a subgroup of COPD patients who are prone Spontaneous sputum samples were collected and proc- to recurrent exacerbations and are more responsive to essed within 60 min on the day of the visit. Sputum treatment [10–12], but it is not clear if this is associated quality was assessed according to Murray-Washington with a different respiratory microbiome. This work criteria [17] and only samples with > 25 leucocytes per sought to investigate these questions. field (M-W ≥ 3) were considered for the study. Sputum samples were frozen until processing, which was carried out in a certified BSL2 hood with appropriate laminar Methods flow.16S rRNA gene was amplified following the 16S Methods are detailed in the Additional file 1 and Metagenomic Sequencing Library Preparation Illumina summarized below. protocol (Part # 15044223 Rev. A, Illumina, CA, USA). Details are provided in the Additional file 1. Study design and ethics This is a cross-sectional, prospective, uncontrolled, mul- Sequence analysis ticentre, observational study. The study protocol was ap- The Quantitative Insights Into Microbial Ecology proved by the Ethics Committees of the participating (QIIME) pipeline 1.9.0 [18] was used for sequence hospitals (IMIM-Hospital del Mar, Hospital Universitari processing to obtain taxonomic information. Further Parc Taulí, Hospital Clinic, Hospital 12 Octubre, Funda- technical details are provided in the Additional file 1. ción Jimenez Díaz and Hospital Son Espases), and all pa- tients included signed their informed consent. Statistical analyses Details are provided in the Additional file 1. In brief, cat- Population egorical variables are expressed as absolute and relative Current or former smokers (≥ 10 pack-year) with stable frequencies, and continuous variables as means and COPD, attending the outpatients’ clinics of five Spanish standard deviations (SD) when the distribution was hospitals between 2014 and 2016 were included in this normal, or as medians and interquartile range (IQR) study. The diagnosis and severity staging of COPD was otherwise. Linear discriminant analysis Effect Size established in accordance with GOLD criteria [8]. Exclu- (LEfSe) was used to identify the differentially abundant sion criteria were: age less than 40 years; a lifetime diag- taxa that explained the differences between the groups nosis of asthma, cystic fibrosis, bronchiectasis or cancer; of participants. The threshold value of the logarithmic patients receiving long-term treatment with oral cortico- LDA score for discriminative features was 2.0. Bacterial steroids or immunosuppressants; any comorbidity limit- α-diversity was assessed through the Chao1 estimator ing cognitive capabilities and ≥ 3 admissions or 1 episode [19] and the Shannon index [20], calculating both severe enough to require more than 30 days in hospital indexes after subsampling with QIIME so as to avoid the previous year. Patients who had been treated with sequencing effort bias. Principal Coordinates Analysis short-term antibiotics and/or corticosteroids at any time (PCoA) with Bray-Curtis dissimilarity index [21] was during the previous three months were considered used to study community composition, assessing the unstable and not considered for the study. statistical significance of the differences in sample groupings through Adonis testing. Interaction between Variables and measurements independent variables was assessed through stratification Sociodemographic data were recorded by specific and multivariate analyses with α-diversity as dependent questionnaires. Lung function values during stability variable. Statistical tests used in the study were two- were obtained from the most recently available forced sided, and a p value of 0.05 or less was reported as sta- spirometry with reversibility testing performed according tistically significant. Statistical analyses were performed to standard techniques the previous year [13]. Peripheral using the SPSS statistical software package version 18 blood cell counts were obtained at enrolment and used (SPSS Inc., Chicago, IL, USA). to identify patients with ≥2% circulating eosinophils [14]. Episodes of increased dyspnoea, sputum production Results and/or purulence during the previous year were identi- Patient characteristics fied and considered as exacerbations when treated with Table 1 summarizes the main demographic and clinical antibiotics and/or corticosteroids [15, 16]. Participants characteristics of the 72 patients included. They were were considered as frequent exacerbators (FE) when they mostly men (88.9%), with a mean age of 68 (SD 7.9) reported ≥2 exacerbations the previous year. years and FEV1 of 48.7 (SD 20.1)% of reference. Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 3 of 9 Table 1 Demographic and clinical characteristics of the patients N All patients Exacerbations previous year p 0 1 ≥2 72 31 18 23 Age, mean (SD) 68 (7.9) 66 (9) 69 (7) 68 (7) 0.387 Sex (male), n (%) 64 (88.9) 27 (87.1) 16 (88.9) 21 (91.3) 0.888 Cumulative smoking (pack-year), median (IQR) 60 (45–80) 60 (44–76) 50 (41–85) 60 (49–97) 0.383 Postbronchodilator FEV1%, median (IQR) 44 (33–60) 52 (42–70) 35 (32–52) 35 (28–49) 0.001 BMI, median (IQR) 27 (24–30) 28 (25–29) 28 (23–30) 26 (23–30) 0.669 Blood eosinophils (× 109/L), median (IQR) 200 (100–270) 200 (130–3009 185 (92–255) 200 (100–300) 0.414 Blood eosinophils (%), median (IQR) 2.4 (1.4–3.4) 2.8 (1.7–3.6) 2 (1.3–2.8) 1.8 (1.1–3.4) 0.156 9 Blood leucocytes (×10 /L), median (IQR) 7845 (6635–9180) 7210 (6520–8940) 7915 (6505–8510) 8110 (7030–10,170) 0.481 Exacerbations last year, median (IQR) 1 (0–2) 0 1 3 (2–4) Airflow limitation severity (GOLD), n (%) GOLD 1 6 (8.3) 5 (16.1) 0 (0) 1 (4.3) 0.013 GOLD 2 19 (26.4) 11 (35.5) 4 (22.2) 4 (17.4) GOLD 3 36 (50) 14 (45.2) 12 (66.7) 10 (11.1) GOLD 4 11 (15.3) 1 (3.2) 2 (18.2) 8 (34.8) 16S rRNA analysis patients studied (p = 0.963), when compared with patients At phylum level, 13 different phyla were identified, six of not receiving this treatment. Similarly, the continuous use them with median relative abundance (RA) above 0.1% of LAB as monotherapy was not associated with signifi- (Additional file 1: Table S1). At genus level, 190 different cant changes in the respiratory microbiome (p = 0.854). genera were identified and, after removing the genera present in only one sample, 171 remained for subsequent Exacerbation frequency analyses, 26 of them with RA above 0.1% (Table 2). In the previous year, 31 patients (43.1%) did not report any acute episodes, 18 (25%) referred only one and 23 Age suffered two or more (31.9%), and were considered FE. Alpha-diversity parameters showed a negative relationship Demographic and clinical characteristics of these three with age (R2 = 0.075 p = 0.020 and R2 = 0.074 p = 0.020 re- groups only showed statistically significant differences in spectively), but β-diversity analysis did not show signifi- lung function, with lower values in COPD patients cant differences in relation with this variable (p = 0.389). reporting one or more exacerbations the previous year (Table 1) Comparisons between their respiratory Airflow limitation microbiomes were made in pairs using patients without We found a significant progressive increase in the RA of exacerbations as the reference. Patients with one exacer- Pseudomonas genus and a decrease in the RA of bation had significantly lower RA of the phylum TM7 Treponema in patients with more severe airflow limita- (Additional file 1: Figure S1) and lower RAs of 13 differ- tion (Fig. 1). Regarding bacterial diversity, neither α- ent genera (Additional file 1: Figure S2). However, α- diversity parameters nor β-diversity analysis showed diversity parameters did not show significant differences significant differences between GOLD grades of airflow between the groups, and β-diversity analysis did not limitation. Of note, airflow limitation severity was not demonstrate bacterial communities with a different related to age (p = 0.245). composition (p = 0.081). FE showed a significant decrease in the RA of TM7 and Spirochaetes at phylum Pharmacological treatment level (Additional file 1: Figure S3). At genus level, the Forty-nine COPD patients had not modified their inhaled RAs of Pseudomonas, Selenomonas and Anaerococcus maintenance treatment during the previous year; thirty- increased, while 10 different genera decreased (Fig. 2). six of them (73.5%) used a combination of LAB/ICS, 9 Alpha-diversity analysis did not show significant (18.4%) were treated with LAB as monotherapy and 4 differences between groups, but β-diversity analysis (8.2%) were not receiving COPD treatment. LAB/ICS demonstrated that the bacterial communities of treatment did not have any effect on either α-diversity COPD patients with frequent exacerbations differed (p = 0.365) or bacterial community composition in the significantly (p = 0.014). Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 4 of 9 Table 2 Relative abundance of the genera detected. Only Circulating eosinophils genera appearing in more than one sample and with median Forty-two of the participants (58.3%) had ≥2% blood relative abundances over 0.1% are shown eosinophils. There were no significant differences in age Genera Relative abundance, median (IQR) (p = 0.368), sex (p = 1.00) and number of exacerbations Rothia 18.65 (9.37–30.33) the previous year (p = 0.080) between patients with ≥2% Gemellaceae_g 7.32 (2.24–13.51) circulating eosinophils or less. The bacterial community in Prevotella 6.87 (2.33–15.05) the former had significantly higher RAs of the phyla Bac- teroidetes and Spirochaetes (Additional file 1: Figure S4). Granulicatella 4.43 (2.16–6.58) At genus level, 20 genera showed significantly higher RA Fusobacterium 2.23 (0.26–3.93) and one genus, Peptostreptococcus, had lower RA in these Porphyromonas 1.97 (0.13–8.22) patients (Fig. 3). Alpha-diversity was significantly higher in Actinomyces 1.80 (0.58–4.41) patients with ≥2% circulating eosinophils [Chao1 index: Streptococcus 1.92 (1.23–3.44) 224.51 (74.88) vs 277.39 (78.92), p = 0.006; and Shannon Pseudomonas 1.39 (0.40–6.36) index: 3.94 (1.05) vs 4.54 (1.06), p = 0.020] (Fig. 4). Pear- son's correlation coefficients were r = 0.282 (p = 0.016) for Veillonella 1.00 (0.50–1.44) Chao1 and r = 0.231 (p = 0.051) for Shannon. β-diversity Atopobium 0.69 (0.30–1.56) analysis showed a trend towards different bacterial com- Oribacterium 0.62 (0.15–1.26) munities (p = 0.072). Leptotrichia 0.51 (0.09–1.90) Multivariate analyses were performed with α-diversity Lachnospiraceae_g 0.50 (0.08–1.09) as dependent variable and eosinophils levels as predict- [Prevotella] 0.44 (0.04–1.87) ive factor, including age and lung function as covariates. Eosinophils in blood, expressed as percentage, kept a Moryella 0.35 (0.06–0.94) statistically significant relationship with Chao1 in this Campylobacter 0.35 (0.10–0.77) analysis (p = 0.026) and a borderline significance for Capnocytophaga 0.29 (0.01–0.99) Shannon (p = 0.051), a finding confirming that the TM7-3_o_f_g 0.28 (0.04–1.45) bronchial microbiome was related to blood eosinophils Megasphaera 0.25 (0.02–1.17) independently of the functional limitations suffered by Bulleidia 0.20 (0.05–0.85) the patient. To explore potential interactions between the previous Haemophilus 0.19 (0.05–1.32) history of exacerbations and eosinophils levels, we Selenomonas 0.18 (0.04–0.62) compared the microbiome in COPD patients with and Parvimonas 0.11 (0.01–0.54) without circulating eosinophils ≥2% stratified by the Lactobacillus 0.12 (0.01–1.07) frequency of exacerbations. We found that the signifi- Lactobacillales_Other_Other 0.11 (0.04–0.24) cant differences in the microbial composition related to patients with eosinophils ≥2% were maintained in the subsample of patients with no exacerbations or only one episode (p = 0.033), but this effect disappeared in FE (p = 0.995). Fig. 1 Genera showing significant differences in their relative abundance according to GOLD severity level, with higher figures for Treponema in GOLD 1 (b) and for Pseudomonas in GOLD 4 (a) (dotted line = median) Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 5 of 9 Fig. 2 The RA of three genera significantly increased (a) and 10 genera decreased (b) in patients with ≥2 exacerbations the previous year compared with non-exacerbators (dotted line = median) Discussion gut microbiome [22]. Besides, in patients with severe The main findings of this study were that the diversity asthma, an inverse correlation between α-diversity and and composition of the respiratory microbiome in clinic- age has been also reported [23]. A previous work has ally stable COPD patients change in relation to age, the shown less microbial diversity of the respiratory micro- severity of airflow limitation, exacerbation frequency and biome in younger COPD patients using bronchoalveolar eosinophils in peripheral blood. lavage [24], but this sample targets the peripheral airway In our study, older age was significantly associated of the lung and it is not representative of the bronchial with a loss of diversity, which has been also found in the tree mainly sampled by sputum [25]. Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 6 of 9 Fig. 3 Genera with significantly higher (n = 20) (b) and lower (a) RAs (n = 1) in patients with circulating eosinophils ≥2% (dotted line = median) Fig. 4 α-diversity parameters, Chao1 (a) and Shannon (b), in patients classified according to circulating eosinophils ≥2% Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 7 of 9 We found that patients with more severe airflow bacteria such as Haemophilus influenzae and Streptococ- limitation had a significant decrease in the RA of cus pneumonia which has been reported to be overrep- Treponema and a progressive increase in the RA of resented in patients with eosinophils counts below 2% Pseudomonas. These results suggest that severity-related treated with ICS [12]. Yet, when we stratified the COPD changes in the respiratory microbiome are based on a patients included according to both the level of circulat- decrease in specific genera, which are partially ing eosinophils and the frequency of exacerbations, we substituted by Pseudomonas. This change may be partly observed that the differences related to blood eosino- related to recurrent antibiotic exposure in previous phils disappeared in FE, likely highlighting a higher years, considering the antibiotic sensitivity of the micro- impact of frequent exacerbations on the respiratory organisms part of Treponema genus. Previous cross- microbiome in these patients. sectional studies evaluating the relation between bacter- This study has some potential limitations. First, we do ial diversity and more severe airway limitation have not have a wide representation of the respiratory micro- mostly showed a decline in advanced stages [26–28], biome, which has been shown to be heterogeneous associated with changes in the RAs of specific genera throughout the airway, because only sputum samples such as Haemophilus [28, 29]. These partly discordant were analysed. Second, although the patients included results may be due to patient selection, considering that had not taken antibiotics three months before their most of the previous studies have focused on a inclusion, we lack information on previous antibiotic restricted number of patients with moderate or severe treatments, which may have had an effect on their mi- disease [26, 27] or an overrepresentation of patients crobial communities. Finally, we analysed only bacterial with moderate disease [28] whereas we studied a communities, fungi and virus may also have an effect on wider range of disease severity (GOLD 1–4). Our these patients, either directly or through interactions results, therefore, support a significant role for with other microorganisms and the host. Pseudomonas as the severity of the disease increases to higher lung function impairment. Conclusions We also found that the respiratory microbiome was This study shows that the respiratory microbiome in significantly different in FE. Previous studies have inves- clinically stable COPD patients changes in relation to tigated the characteristics of the respiratory microbiome age, severity of airflow limitation, history of previous during exacerbations [7, 30], and recently, like we do in exacerbations and level of circulating eosinophils. These this study, Mayhew and cols. [28] reported specific factors need to be considered when interpreting respira- characteristics in the bronchial microbiome recovered tory microbiome changes in patients with COPD. from FE patients during clinical stability. Both studies show that FE have a different respiratory microbiome during clinical stability, suggesting that the microbial Additional file changes during exacerbations in FE may be a mixture of the dysbiosis found in stability and specific Additional file 1: Table S1. Relative abundances of the phyla detected. Figure S1. The TM7 phylum had significantly lower relative abundance exacerbation-related perturbations of the lung bacteria in patients with one exacerbation than patients without exacerbations community composition [28, 31]. the previous year (dotted line = median). Figure S2. Thirteen genera with Circulating eosinophils ≥2% were associated with significantly lower relative abundances in COPD patients with one exacerbation the previous year compared to non-exacerbators. Figure S3. higher microbial diversity in the population studied. A significant reduction in the RA of phyla TM7 and Spirochaetes in Patients with ≥2% blood eosinophils have been reported patients with ≥2 exacerbations the previous year, using patients to have more frequent exacerbations and a better re- without exacerbations as the reference (dotted line = median). Figure S4. Phyla with significantly higher relative abundances in sponse to ICS preventive therapy [32] Previous studies COPD patients showing circulating eosinophils ≥2%. (DOCX 725 kb) have demonstrated a different bronchial microbiome in eosinophilic COPD exacerbations [7, 31], which seems Abbreviations related to Th2 inflammation in both COPD and asthma COPD: Chronic obstructive pulmonary disease; DNA: Deoxyribonucleic acid; [33]. In our study we observed that in patients with ≥2% FE: Frequent exacerbators; FEV1: Forced expiratory volume the first second; blood eosinophils higher microbial diversity is already GOLD: Global Initiative for Chronic Obstructive Lung Disease; ICS: Inhaled corticosteroids; IQR: Interquartile range; LAB: long-acting bronchodilator; present in stability, with an increase in the RA of 20 LDA: Linear discriminant analysis; LEfSe: Linear discriminant analysis Effect genera. Similar results have been reported in stable Size; PCoA: Principal Coordinates Analysis; PCR: Polymerase chain reaction; asthmatic patients, who showed a good correlation QIIME: Quantitative Insights Into Microbial Ecology; RA: Relative abundance; rRNA: Ribosomal ribonucleic acid; SD: Standard deviations; vs: Versus between the percentage of eosinophils in bronchoalveo- lar lavage and bacterial diversity [34]. Higher bacterial Acknowledgements diversity may have a protective role in patients with ≥2% Authors thank participants for their willingness to contribute to medical blood eosinophils avoiding the presence of pathogenic research and Michael Maudsley for providing an outline for this manuscript. Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 8 of 9 Authors’ contributions 5. Faner R, Sibila O, Agustí A, Bernasconi E, Chalmers JD, Huffnagle GB, et al. CM, CC, AM, AC, CA, OS, GPB, BGC, AA, JG and EM recruited the patients and The microbiome in respiratory medicine: current challenges and future obtained the samples. EM, JG, BGC, OS and AA obtained funding for the perspectives. Eur Respir J. 2017;49(4):1602086. project. LM, SP, MGN, CL, LS, RF and SC analyzed the samples and obtained 6. Millares L, Ferrari R, Gallego M, Garcia-Nuñez M, Pérez-Brocal V, Espasa M, et the clinical information. LM, SP and EM interpreted the data. LM and EM al. Bronchial microbiome of severe COPD patients colonised by wrote the paper. AA and JG contributed in writing the manuscript. All Pseudomonas aeruginosa. Eur J Clin Microbiol Infect Dis Off Publ Eur Soc authors read and approved the final manuscript. Clin Microbiol. 2014;33:1101–11. 7. Wang Z, Bafadhel M, Haldar K, Spivak A, Mayhew D, Miller BE, et al. Funding Lung microbiome dynamics in COPD exacerbations. Eur Respir J. This study was funded by FIS PI15/00167 and PI15/02042 from Instituto de 2016;47:1082–92. Salud Carlos III (ISCIII), SAF 2014–54371 (FEDER) from Ministerio de Economía 8. From the Global Strategy for the Diagnosis, Management and Prevention of y Competitividad, Sociedad Española de Neumología y Cirugía Torácica COPD, Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2017. (SEPAR), Laboratorios Menarini S. A, Fundació Pla Armengol and Centro de Available from: https://goldcopd.org. Investigación Biomédica en Red de Enfermedades Respiratorias (CIBERES). 9. Miravitlles M, Soler-Cataluña JJ, Calle M, Molina J, Almagro P, Quintano JA, CIBERES is an initiative of the ISCIII. et al. Spanish guidelines for Management of Chronic Obstructive Pulmonary This work was also supported by grants from Agència de Gestió d’Ajuts Disease (GesEPOC) 2017. Pharmacological treatment of stable phase. Arch Universitaris i de Recerca—Generalitat de Catalunya (2017 GRC-716). Bronconeumol. 2017;53:324–35. 10. Ho J, He W, Chan MTV, Tse G, Liu T, Wong SH, et al. Eosinophilia and clinical outcome of chronic obstructive pulmonary disease: a meta-analysis. Sci Rep. Availability of data and materials 2017;7:13451. Bacterial 16SrRNA datasets from this study are accessible in the European 11. Vedel-Krogh S, Nielsen SF, Lange P, Vestbo J, Nordestgaard BG. Blood Nucleotide Archive under the study PRJEB26773 with the sample numbers eosinophils and exacerbations in chronic obstructive pulmonary disease. ERS2486515–609. The Copenhagen general population study. Am J Respir Crit Care Med. 2016;193:965–74. Ethics approval and consent to participate 12. Contoli M, Pauletti A, Rossi MR, Spanevello A, Casolari P, Marcellini A, et al. This is a cross-sectional, prospective, uncontrolled, multicentre, observational Long-term effects of inhaled corticosteroids on sputum bacterial and viral study. The study protocol was approved by the Ethics committees of the loads in COPD. Eur Respir J. 2017;50:1700451. participating hospitals (IMIM-Hospital del Mar, Hospital Universitari Parc Taulí, 13. Standardization of Spirometry, 1994 Update. American Thoracic Society. Am Hospital Clinic, Hospital 12 Octubre, Fundación Jimenez Díaz and Hospital J Respir Crit Care Med. 1995;152:1107–36. Son Espases), and all patients included signed their informed consent. 14. Moermans C, Heinen V, Nguyen M, Henket M, Sele J, Manise M, et al. Local and systemic cellular inflammation and cytokine release in chronic obstructive pulmonary disease. Cytokine. 2011;56:298–304. Consent for publication 15. Rodriguez-Roisin R. Toward a consensus definition for COPD exacerbations. Not applicable. Chest. 2000;117(5 Suppl 2):398S–401S. 16. Anthonisen NR, Manfreda J, Warren CP, Hershfield ES, Harding GK, Nelson Competing interests NA. Antibiotic therapy in exacerbations of chronic obstructive pulmonary The authors declare that they have no competing interests. disease. Ann Intern Med. 1987;106:196–204. 17. Murray PR, Washington JA. Microscopic and baceriologic analysis of Author details expectorated sputum. Mayo Clin Proc. 1975;50:339–44. 1 Fundació Parc Taulí- Institut d’ Investigació i Innovació Parc Taulí (I3PT), 18. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello Barcelona, Spain. 2Centro de Investigación Biomédica en Red de EK, et al. QIIME allows analysis of high-throughput community sequencing Enfermedades Respiratorias (CIBERES), Instituto de Salud Carlos III (ISCIII), data. Nat Methods. 2010;7:335–6. Madrid, Spain. 3Universitat Autònoma de Barcelona, Esfera UAB, Barcelona, 19. Chao A. Nonparametric estimation of the number of classes in a population. Spain. 4Respiratory Medicine Department, Hospital del Mar – IMIM, Barcelona, Scand J Stat. 1984;11:265–70. Spain. 5Department of Respiratory Medicine, Hospital Universitari Parc Taulí, 20. Shannon CE. The mathematical theory of communication. 1963. MD Parc Taulí 1, 08208, Sabadell, Barcelona, Spain. 6Health Services Research on Comput Comput Med Pract. 1997;14:306–17. Chronic Diseases Network- REDISSEC, Galdakao, Spain. 7Barcelona Respiratory 21. Bray JR, Curtis JT. An ordination of the upland forest communities of Network, Barcelona, Spain. 8DCEXS, Universitat Pompeu Fabra, Barcelona, southern Wisconsin. Ecol Monogr. 1957;27:325–49. Spain. 9Department of Microbiology, Hospital Universitari Parc Taulí, Sabadell, 22. Maffei VJ, Kim S, Blanchard E, Luo M, Jazwinski SM, Taylor CM, et al. Spain. 10Department of Respiratory Medicine, Hospital Universitario 12 de Biological aging and the human gut microbiota. J Gerontol A Biol Sci Med Octubre, Madrid, Spain. 11Respiratory Department, Biomedical Research Sci. 2017;72:1474–82. Institute Sant Pau (IIB-Sant Pau), Hospital de la Santa Creu i Sant Pau, 23. Taylor SL, Leong LEX, Choo JM, Wesselingh S, Yang IA, Upham JW, et al. Barcelona, Spain. 12Department of Respiratory Medicine, Fundación Jiménez Inflammatory phenotypes in patients with severe asthma are associated Díaz, Madrid, Spain. 13Department of Respiratory Medicine, Hospital with distinct airway microbiology. J Allergy Clin Immunol. Universitari Son Espases-IdISBa, Mallorca, Spain. 14Institut Respiratori, Hospital 2018;141:94–103.e15. Clinic, IDIBAPS, Universidad de Barcelona, Barcelona, Spain. 15Department of 24. Pragman AA, Kim HB, Reilly CS, Wendt C, Isaacson RE. The lung microbiome Medicine, Universitat Autònoma de Barcelona (UAB), Barcelona, Spain. in moderate and severe chronic obstructive pulmonary disease. PLoS One. 2012;7:e47305. Received: 14 February 2019 Accepted: 30 May 2019 25. Cabrera-Rubio R, Garcia-Núñez M, Setó L, Antó JM, Moya A, Monsó E, et al. Microbiome diversity in the bronchial tracts of patients with chronic obstructive pulmonary disease. J Clin Microbiol. 2012;50:3562–8. References 26. Garcia-Nuñez M, Millares L, Pomares X, Ferrari R, Pérez-Brocal V, Gallego M, 1. Han MK, Agusti A, Calverley PM, Celli BR, Criner G, Curtis JL, et al. Chronic et al. Severity-related changes of bronchial microbiome in chronic obstructive pulmonary disease phenotypes: the future of COPD. Am J Respir obstructive pulmonary disease. J Clin Microbiol. 2014;52:4217–23. Crit Care Med. 2010;182:598–604. 27. Galiana A, Aguirre E, Rodriguez JC, Mira A, Santibañez M, Candela I, et al. 2. Vestbo J. COPD: definition and phenotypes. Clin Chest Med. 2014;35:1–6. Sputum microbiota in moderate versus severe patients with COPD. 3. Agustí A. Phenotypes and disease characterization in chronic obstructive Eur Respir J. 2014;43:1787–90. pulmonary disease. Toward the extinction of phenotypes? Ann Am Thorac 28. Mayhew D, Devos N, Lambert C, Brown JR, Clarke SC, Kim VL, et al. Soc. 2013;10(Suppl):S125–30. Longitudinal profiling of the lung microbiome in the AERIS study 4. Hilty M, Burke C, Pedro H, Cardenas P, Bush A, Bossley C, et al. Disordered demonstrates repeatability of bacterial and eosinophilic COPD microbial communities in asthmatic airways. PLoS One. 2010;5:e8578. exacerbations. Thorax. 2018;73:422–30. Millares et al. BMC Pulmonary Medicine (2019) 19:112 Page 9 of 9 29. Sze MA, Dimitriu PA, Suzuki M, McDonough JE, Campbell JD, Brothers JF, et al. Host response to the lung microbiome in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2015;192:438–45. 30. Bafadhel M, McKenna S, Terry S, Mistry V, Reid C, Haldar P, et al. Acute exacerbations of chronic obstructive pulmonary disease: identification of biologic clusters and their biomarkers. Am J Respir Crit Care Med. 2011;184:662–71. 31. Wang Z, Singh R, Miller BE, Tal-Singer R, Van Horn S, Tomsho L, et al. Sputum microbiome temporal variability and dysbiosis in chronic obstructive pulmonary disease exacerbations: an analysis of the COPDMAP study. Thorax. 2018;73:331–8. 32. Pascoe S, Locantore N, Dransfield MT, Barnes NC, Pavord ID. Blood eosinophil counts, exacerbations, and response to the addition of inhaled fluticasone furoate to vilanterol in patients with chronic obstructive pulmonary disease: a secondary analysis of data from two parallel randomised controlled trials. Lancet Respir Med. 2015;3:435–42. 33. Ghebre MA, Pang PH, Diver S, Desai D, Bafadhel M, Haldar K, et al. Biological exacerbation clusters demonstrate asthma and chronic obstructive pulmonary disease overlap with distinct mediator and microbiome profiles. J Allergy Clin Immunol. 2018;141:2027–2036.e12. 34. Denner DR, Sangwan N, Becker JB, Hogarth DK, Oldham J, Castillo J, et al. Corticosteroid therapy and airflow obstruction influence the bronchial microbiome, which is distinct from that of bronchoalveolar lavage in asthmatic airways. J Allergy Clin Immunol. 2016;137:1398–1405.e3. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.