You are here

Article Text

Article menu
State of the art review
Sputum microbiome profiling in COPD: beyond singular pathogen detection
  1. Benedikt Ditz1,2,
  2. Stephanie Christenson3,
  3. John Rossen4,
  4. Chris Brightling5,
  5. Huib A M Kerstjens1,2,
  6. Maarten van den Berge1,2,
  7. Alen Faiz1,2,6
  1. 1 Department of Pulmonary Diseases, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
  2. 2 Groningen Research Institute for Asthma and COPD, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
  3. 3 Department of Medicine, Division of Pulmonary, Critical Care, Allergy and Sleep Medicine, University of California, San Francisco, the United States
  4. 4 Department of Medical Microbiology and Infection Prevention, University Medical Center, University of Groningen, Groningen, the Netherlands
  5. 5 Institute of Lung Health, University of Leicester, Leicester, UK
  6. 6 Respiratory Bioinformatics and Molecular Biology, University of Technology Sydney, Sydney, New South Wales, Australia
  1. Correspondence to Benedikt Ditz, Pulmonology, University Medical Center Groningen, Groningen 9713 GZ, The Netherlands; l.b.ditz{at}umcg.nl

Abstract

Culture-independent microbial sequencing techniques have revealed that the respiratory tract harbours a complex microbiome not detectable by conventional culturing methods. The contribution of the microbiome to chronic obstructive pulmonary disease (COPD) pathobiology and the potential for microbiome-based clinical biomarkers in COPD are still in the early phases of investigation. Sputum is an easily obtainable sample and has provided a wealth of information on COPD pathobiology, and thus has been a preferred sample type for microbiome studies. Although the sputum microbiome likely reflects the respiratory microbiome only in part, there is increasing evidence that microbial community structure and diversity are associated with disease severity and clinical outcomes, both in stable COPD and during the exacerbations. Current evidence has been limited to mainly cross-sectional studies using 16S rRNA gene sequencing, attempting to answer the question ‘who is there?’ Longitudinal studies using standardised protocols are needed to answer outstanding questions including differences between sputum sampling techniques. Further, with advancing technologies, microbiome studies are shifting beyond the examination of the 16S rRNA gene, to include whole metagenome and metatranscriptome sequencing, as well as metabolome characterisation. Despite being technically more challenging, whole-genome profiling and metabolomics can address the questions ‘what can they do?’ and ‘what are they doing?’ This review provides an overview of the basic principles of high-throughput microbiome sequencing techniques, current literature on sputum microbiome profiling in COPD, and a discussion of the associated limitations and future perspectives.

  • COPD
  • sputum microbiome
  • high-throughput sequencing techniques
https://creativecommons.org/licenses/by/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

https://doi.org/10.1136/thoraxjnl-2019-214168

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Chronic obstructive pulmonary disease (COPD) is a highly prevalent respiratory disease and the leading cause of morbidity and mortality worldwide.1 2 It is heterogeneous, but many patients experience progressive airflow obstruction, exacerbations and/or persistent symptoms of dyspnoea, cough and sputum production.2 Exacerbations are heterogeneous and vaguely defined as acute worsening of respiratory symptoms outside of normal day-to-day variation, usually associated with a healthcare utilisation event (eg, prescription of oral steroids and/or antibiotics, emergency room visit). They are associated with accelerated disease progression, increased mortality and major healthcare costs.2–4 The heterogeneity of COPD overall, and exacerbations, in particular, complicates the clinical assessment of disease severity and outcome. Determining the fundamental biology underlying COPD heterogeneity is vitally important, as is the identification of biomarkers to guide personalisation of therapeutic strategies.5–7

    Elucidating the role of respiratory microbiota with respect to COPD pathobiology has potential implications for improving diagnostics and predicting outcomes.8–12 The respiratory microbiome, which may be assayed using culture-independent microbial sequencing, is characterised by the micro-organisms and their genomes present in the respiratory tract under specific environmental conditions (eg, exposures or disease states).13 Metrics assessing respiratory microbiome composition differ between healthy controls and COPD, and alterations in bacterial abundance and microbial diversity (ie, microbes present in differing proportions) correlate with altered airway inflammatory processes.14–17 Current knowledge of the respiratory microbiome still relies mainly on the 16S rRNA gene sequencing approach. Other high-throughput techniques, such as metagenomics, metatranscriptomics and metabolomics, will aid in both a broader investigation of microbial genomes and the examination of associated functional aspects. In this review, we provide an overview of the basic principles of high-throughput techniques for microbiome assessment and analytical approaches. Further, we discuss the current evidence, limitations and future perspectives of sputum microbiome profiling in COPD.

    Basic principles of high-throughput sequencing techniques: 16S rRNA gene sequencing, shotgun metagenomics, metatranscriptomics, metabolomics

    Culture-independent high-throughput sequencing techniques have revealed that the human body harbours a microbiome with a complexity far beyond the scope that is detectable by conventional culturing methods.18 19 Importantly, cultivation efforts can be extended in order to increase the detection rate of microbes, which represents a significant benchmark for evaluating new culture-independent sequencing technologies.20 The choice of technique for microbiome assessment and characterisation should be based on the research question, with consideration of the strengths and weaknesses of each technique. The respiratory tract, certainly in stable COPD, is a low-biomass environment in which the microbial density decreases in a gradient from the upper (URT) to lower respiratory tract (LRT), where microaspiration from the URT is considered to be the dominant source.21 Therefore, microbial signals must be discerned from possible contamination during the sample collection and processing prior to analysis.22 23 Furthermore, appropriate sample collection and processing following standardised protocols, with particular attention to DNA and RNA extraction methods, should be followed as multiple steps in this process can influence microbe yield and diversity.24 A comprehensive discussion of best practices for microbiome analysis overall and with respect to the respiratory tract have been well described previously.22 23 25 26

    16S rRNA gene sequencing was used in the pioneering work in respiratory microbiome research, and current knowledge still relies mainly on this approach.8 27 16S rRNA sequencing is based on PCR amplification using primers that target the 16S ribosomal gene variable regions of bacterial genomes, which can be used for taxonomic classification.28 This approach is rapid, well tested and relatively low cost. By focusing on a specific region of the bacterial genome, it requires only a limited sequencing depth. 16 s rRNA sequencing is used to survey microbial communities, answering the question ‘who is there?’

    There are many techniques for analysing microbial communities present within and across samples, but alpha and beta diversity are two of the most common metrics. Alpha diversity captures the number of operational taxonomic units (OTUs) present in a sample (‘richness’) and how they are distributed (‘evenness’), reducing the complexity of these relationships into a single metric. Commonly used alpha diversity metrics include the Shannon index or Simpson index.26 29 Beta diversity compares the dissimilarity between samples and provides a matrix of distance. Alpha and beta diversity do not, however, capture the full complexity and dimensionality of microbial community relationships. Thus, more recent work has begun to explore the microbiome in the context of microbial community ecology.30 Community ecology theory provides a framework and set of analytical tools to consider how microbes interact with each other and their human host over time and space and in response to exposures or insults (eg, disease states). Newer sequencing approaches that capture microbial function and presence are poised to contribute greatly to our understanding of these ecological dynamics.

    16S rRNA gene sequencing exhibits other important shortcomings that may be overcome by newer technologies. For example, it cannot distinguish between live and dead microbes, which may be relevant for several reasons, such as assessing the impact of antibiotic therapy in clinical samples.31 Also, 16S rRNA gene sequences are relatively short. Thus, closely related bacteria may share similar sequences, making a separation between different species challenging.26 Finally, 16S sequencing analyses are restricted to the detection of bacteria and archaea since viruses or fungi do not carry 16S rRNA genes.

    Shotgun metagenomic sequencing is based on unrestricted DNA sequencing of genetic content in a sample. This approach can capture DNA-based viruses and eukaryotic DNA, including fungal microorganisms, in addition to bacteria.26 Shotgun metagenomics allows for an unbiased and deeper taxonomical analysis of the microbiome than 16S rRNA gene method, since dissimilar regions of microbial genomes are sequenced. The assembly of short sequencing reads into larger fragments can increase the discriminatory power between microbes, enhancing taxonomic resolution up to the species or strain level.25 32 Feigelman et al demonstrated how unbiased DNA sequencing in sputum from patients with cystic fibrosis (CF) can enhance in-depth microbiome profiling.33 Furthermore, shotgun metagenomic analyses add insight into the question ‘what can these microbes do?’ The potential metabolic capacity of the microbiome can be evaluated by assigning microbial genes to pathway profiles and investigating potential cross-talk between the microbiome and human host and between microbes.34 35

    Metagenomic sequencing has several weaknesses. As with 16S rRNA gene sequencing, it is unable to distinguish between live and dead microbes. The high sequencing depths necessary for taxonomic classification may be prohibitively expensive, although these costs are diminishing.25 26 Most respiratory viruses are RNA-based and therefore not detectable by metagenomic approaches. Additionally, pathway profiling of microbial genes will only allow for the evaluation of potential metabolic capacities, and can thus only conjecture at functional roles.

    Metatranscriptomic sequencing examines RNA-level genetic content. RNA sequencing (RNA-seq) provides a survey of microbiota that include RNA-based respiratory viruses.36–38 Unlike DNA-based methods, metatranscriptomics focuses on living microbiota. By assessing gene expression (transcriptome) of the microbiota, it can enable a better assessment of microbial functional activity. Further, this approach allows for a simultaneous assessment of the host transcriptome. Thus, metatranscriptomics may provide broader insight into the question ‘what are they doing?’ than metagenomics by allowing for an assessment of functional interactions between different microbiota and between host and microbes.

    There are drawbacks to metatranscriptomics as well.39 RNA is vulnerable with a short half-life and requires careful collection and storage associated with high costs. Further, results can be biased towards microbial genes with higher transcription rates. Consequently, it is preferable that RNA-based approaches are complemented by DNA-based sequencing so that RNA/DNA ratios and microbial transcriptional activity may be examined. Finally, the application of RNA-seq to microbiome research is still relatively new, especially with respect to the respiratory tract. Therefore, best practices for sample processing and analysis are still evolving.

    Metabolomic data generation is based on chromatography techniques and detection methods, such as mass spectrometry or nuclear magnetic resonance, instead of sequencing.40 Application of these technologies to microbiome research is still in its infancy and has been better studied in the gut than the respiratory tract. However, the techniques provide an opportunity to more accurately answer the question ‘what are the microbes doing?’ when used in combination with microbial and host sequencing that identify relevant alterations in microbial communities and host response. Microbial-derived metabolites include those synthesised directly by microbes and those produced by the host and biochemically modified by microbes.41 They are known to affect host physiology (host–microbiota interactions) as well as dynamics within the microbiome (microbe–microbe interactions). They have been shown to play important roles in health and disease. These include crucial roles in the maintenance of the epithelium and immune cell function, processes dysregulated in COPD pathogenesis.41 42 Thus, metabolomics-based analyses provide a promising opportunity for studying microbial-host dynamics within the respiratory tract.

    Overall, evolving high-throughput methods are revolutionising microbiome research. Although 16S rRNA gene sequencing is still the most common technology used in respiratory microbiome research, shotgun metagenomics, metatranscriptomics and metabolomics are increasingly applied to microbiome research in other compartments (eg, gastrointestinal tract). A crucial challenge remains the depletion of host DNA, which represents the vast majority of DNA present in low-biomass samples.22 Samples can be pretreated to reduce the host genome background, however, its effect on an unbiased microbiome detection remains unknown.22 43 Further, standardisation of methods, including sample collection, processing and analytical pipelines, is required in order to further elucidate their utility in respiratory microbiome research.

    Sputum microbiome profiling in COPD: the current body of evidence

    The respiratory tract is an anatomically and physiologically complex organ. Physiological parameters, such as pH level, relative humidity or temperature, change along the respiratory tract, creating niche-specific growth conditions for microbes.44 Consequently, it is likely that the microbiome varies along the respiratory tract. Furthermore, airway microbiome profiles depend on sampling method (eg, nasal brushing, induced or expectorate sputum, bronchoalveolar lavage (BAL), epithelial brushing, bronchial biopsies) consistent with this niche specificity.17 45–47 To exploit respiratory microbiome characteristics for COPD diagnostics, phenotyping and outcome prediction, it is necessary to identify microbiome biomarkers that are easily obtainable in a patient-friendly manner. Sputum closely aligns with these characteristics, making it a preferred sampling technique. Sputum microbiome characteristics, such as microbial diversity or relative abundance, are associated with disease severity and outcomes in stable COPD and exacerbations (see online supplementary table 1).10 11 45 48 49

    Supplemental material

    Sputum microbiome profiles in stable COPD

    In stable COPD, reduced sputum bacterial microbiome diversity has been shown to be associated with more severe airflow obstruction.17 45 49 50 Galiana et al found that participants with severe airflow obstruction exhibited lower alpha diversity and a higher bacterial load in sputum samples, compared with participants with mild or moderate airflow obstruction.49 This finding has since been reproduced in both sputum and BAL sampling.17 50 51 These studies also reported a shift in the microbial flora with decreasing lung function towards potentially pathogenic bacterial genera, particularly those belonging to the gammaproteobacteria class (eg, Pseudomonas or Haemophilus spp). Thus, the observed decrease in microbial diversity appears to be associated with their outgrowth. Although reduced microbial diversity has been identified at other mucosal sites (eg, gastrointestinal tract) in association with chronic inflammatory diseases (eg, inflammatory bowel disease), it remains of interest whether these associations are influenced by therapeutic interventions.52 Further, identifying causal relationships between chronic inflammation and microbial community changes is an active area of research.

    The association between bacterial load and features of stable COPD appears to be complex and affected by chosen therapies.49 53 Garcha et al showed that higher bacterial load of potential pathogens, such as Haemophilus influenzae, Streptococcus pneumoniae, Moraxella catarrhalis, correlated with more severe airflow limitation in stable COPD.53 Further, they reported a positive correlation between bacterial load and inhaled corticosteroids (ICS) dosage. In a randomised controlled trial, Brill et al reported that chronic antibiotic therapy, often used as an exacerbation preventative therapy, was associated with a ≥3 fold increase in antibiotic resistance among airway bacteria without decreasing the total bacterial load in sputum.54 These associations between therapies and microbial load or function suggest that therapies regularly used in COPD management may have unintended consequences on the microbiome, although the significance of these consequences is unclear.

    There is preliminary evidence that sputum microbial communities from COPD patients with more severe airflow obstruction may have shifted such that they lack potentially commensal bacteria, favouring more pathogenic microbes.49 Several studies identified potential commensal bacteria in the URT, including Dolosigranulum spp and Corynebacterium spp, which appear to be associated with respiratory health and pathogen clearance.55–57 Galiana e t al found that the potentially commensal bacterial genus Actinomyces was less frequently recovered in sputum from patients with COPD with severe compared with moderate airflow obstruction.49 Iwase et al showed that commensal respiratory bacteria can directly suppress the outgrowth of potential pathogens belonging to the same genus or family, highlighting the importance of microbe–microbe interactions in maintaining homeostasis.58 Potential new therapeutic approaches might not only aim at reducing colonisation of pathogenic microbes but also at increasing colonisation of commensal bacteria in the respiratory tract.

    Sputum microbiome profiles in COPD exacerbations

    GOLD guidelines recommend performing sputum cultures in patients with frequent exacerbations, severe airflow limitation and/or those requiring mechanical ventilation.2 While the goal of sputum culture in this setting is to determine the presence of bacterial pathogens and to guide antibiotic treatment, culture is very limited in this role.59–61

    Evaluating the sputum microbiome with modern high-throughput sequencing technologies might contribute to enhanced exacerbation precipitant diagnosis. Sputum microbiome profiles have been shown to exhibit dynamic changes between stable COPD and exacerbations in at least a patient subset, involving the outgrowth of pathogens, decreased microbial diversity and increased abundance of pathogenic bacterial communities.11 62 There is also evidence that bacterial outgrowth, as occurs in overt infection, may not be necessary for bacteria to trigger an exacerbation. Strain shifts towards more pathogenic bacterial strains, without an overall change in bacterial abundance, have been identified as potential precipitants for some exacerbations.63

    Host biomarkers and microbial profiling in sputum samples suggest that at least three exacerbation subtypes exist, defined by both the airway inflammatory state and the suspected exacerbation precipitant: bacterial, viral and eosinophilic.10 Ghebre et al found that sputum samples clustered by exacerbation subtype exhibited differing microbial community characteristics.10 Exacerbations categorised as bacteria associated had the greatest Proteobacteria abundance and a high Proteobacteria/Firmicutes ratio in their sputum, while exacerbations categorised as eosinophilic exhibited greater sputum microbial alpha diversity.10 Mayhew et al found that exacerbations classified as bacterial and eosinophilic were more like to be repeated within a subject over time than viral exacerbations.51 Thus, exacerbations exhibit substantial biological heterogeneity. Considering exacerbations within the subtyping framework may add considerably to our understanding of underlying mechanisms and provide insight into more appropriate precision-guided management.

    Although changes in bacterial community structure appear to occur in only an exacerbation subset, dysbiosis is indeed associated with particular exacerbation features and worse outcomes. Wang et al found that the sputum microbial community composition changes significantly from stable to exacerbation state in only ~40% of their cohort.48 Participants with microbial dysbiosis had worse health status and lung function, particularly if they also had evidence of eosinophilic inflammation. Furthermore, the abundance of Moraxella spp increased in this study, corroborating other reports that find specific pathogenic phyla and genera are enriched in the exacerbation microbiome.11 62 Leitao Filho et al found that reduced sputum microbial alpha diversity and altered beta diversity during hospitalised exacerbations were associated with increased 1-year mortality.64 Interestingly, associations between limited diversity and disease progression has also been described in other lung diseases, such as CF.65 Together, these results indicate that the heterogeneous contribution of microbes to exacerbations may be clinically relevant.

    Several reports have speculated that the contribution of typically non-pathogenic bacteria to alterations in microbial community ecology may be crucial to exacerbation pathogenesis. Wang et al identified an outgrowth of non-pathogenic Proteobacteria, in addition to pathogenic microbiota, during exacerbations. They speculate that this state of dysbiosis, with differing microbe–microbe and microbe–host interactions, leads to a dysregulated host inflammatory response.11 66 Huang et al reported a positive relationship between the abundance of COPD-associated pathogens during an exacerbation, (eg, H. influenzae, Pseudomonas aeruginosa or Moraxella catarrhalis) and phylogenetically related non-pathogenic bacteria.62 Leitao Filho et al speculate that the absence of potentially commensal bacterial genera in the sputum microbiome is associated with clinical outcomes in exacerbations.64 They found that both the presence of Staphylococcus spp (potentially pathogenic) and the absence of Veillonella spp (potentially commensal) in sputum were strongly associated with the increased risk of mortality after 1 year. This work highlights the need for a better understanding of the role of both pathogenic and commensal microbes within the respiratory microbiome during exacerbations.

    Importantly, two studies did not find changes in diversity metrics between disease stability and exacerbations.51 67 Millares et al additionally failed to identify changes in bacterial abundance but did report changes in functional metabolic pathways of the microbiome during exacerbations. Mayhew et al did not identify a change in diversity metrics but did find an increase in the relative abundance of Moraxella spp during exacerbations.51 67 These reports provide further evidence of the overall phenotypic heterogeneity in exacerbations, but also emphasise the potential importance of strain shifts without changes in species or class-level microbial community structure in precipitating bacterial exacerbations.

    Standardised follow-up studies are required to validate current findings. Eventually, causation and mechanisms between sputum microbial community structure and host inflammatory response require further investigation, which demands causal models and rigorous methods.68 69 A promising approach was recently published by Sanna et al, using a bidirectional Mendelian randomisation analysis to assess causality between the gut microbiome and metabolic diseases.70

    Sputum microbiome profiles after therapeutic interventions of exacerbations

    Current knowledge about the impact of empirical approaches to the treatment of exacerbations on sputum microbiome profiles is still scarce. Wang et al reported that steroid treatment alone was associated with reduced alpha diversity, increased abundance of Proteobacteria (including Haemophilus and Moraxella spp) and an increased Proteobacteria:Firmicutes ratio, whereas opposite trends were found in exacerbations treated with antibiotics (with or without steroids).11 Huang et al also reported an enrichment of Proteobacteria, in addition to other bacterial phyla with steroid treatment, indicating that this treatment alone might increase the burden of specific microbiota.62 In the same study, antibiotic treatment had suppressive effects on this bacterial burden. In contrast to Wang et al findings, Huang et al observed a trend towards increased microbial diversity after steroid treatment and decreased diversity following antibiotic treatment.

    Further research is needed to better understand treatment effects on respiratory microbial dynamics over time as well as microbial–host interactions. Segal et al took a promising approach to study the antimicrobial effects of azithromycin in COPD.71 Using a combination of 16S sequencing and metabolomics they found decreased alpha diversity in azithromycin-treated BAL samples and increased bacteria-produced metabolites known to influence the host inflammatory response. More mult-‘omic studies integrating microbial and host data are needed to elucidate and characterise the potentially deleterious impact of chronic therapies in COPD to better guide therapy decisions.

    In summary, the current body of evidence indicates that the severity of COPD is inversely correlated with sputum microbial diversity and is associated with an outgrowth of potentially pathogenic bacteria. Exacerbation subtypes exhibit altered sputum microbial community characteristics, providing potential targets for precision-guided management. Further studies, integrating microbial and host information, are needed to understand the impact of therapeutic interventions on the microbiome and interactions with the host. Eventually, potential new therapeutic approaches might not only aim at reducing the colonisation of pathogenic microbes but also to either increase the colonisation of commensal bacteria or modulate the associated host response.

    Sputum microbiome profiling in COPD: limitations, challenges and future perspectives

    There are several limitations of sputum microbiome profiling as well as challenges that should be addressed in future studies to better assess the utility of the technique in COPD microbiome research (figure 1): 

    Figure 1
    Figure 1

    Sputum microbiome profiling—beyond singular pathogen detection. COPD, chronic obstructive pulmonary disease.

    Limitations and challenges: the biogeography of sputum

    LRT sampling methods, such as BAL and bronchial biopsies or brushings, are invasive approaches and thus poorly suited for longitudinal assessment. However, COPD is generally a disease of the LRT. Thus, understanding the similarities and differences between these LRT sampling methods and more easily obtainable sputum sampling is essential for interpretation.47 72 In a small study (n=20), Sulaiman et al found that induced sputum microbiota were more similar to those obtained by oral wash than those obtained from the lower airway (bronchial biopsies) in participants with and without non-tuberculous mycobacterial infections.72 Durack et al also reported an enrichment of oral bacteria in induced sputum samples compared with bronchial biopsies.47 However, they found that the abundance of genera identified in bronchial biopsies correlated quite well with those detected in both induced sputum and oral wash. Thus, despite the overall difference in microbial composition between URT and LRT sampling, sputum does have utility in assessing LRT microbiome composition, in particular, relative microbial abundance. These correlations between lower and upper airway microbe abundance underscore the theory that microaspiration of microbes from the URT is the primary source of bronchial microbiota.21 46 73 While characterising the airway microbiome using sputum should be done considering the limitations, the considerable strengths in comparison to LRT sampling should be considered as well. Sputum is easily obtainable and alterations in the sputum microbiome correlate with clinical outcomes. Thus, sputum microbiome biomarkers may be useful even if they do not accurately reflect the LRT microbiome.

    Limitations and challenges: induced versus spontaneously expectorated sputum

    Differences in microbiome profiles obtained from spontaneous versus induced sputum samples are particularly understudied. Although patients without spontaneous sputum production can safely undergo sputum induction even during exacerbations, preliminary results suggest that microbial composition may be affected by sampling technique.74–76 Furthermore, Gershman et al found that the duration of sputum induction affects the host cellular and biochemical composition in samples and suggest that the large airways are sampled at the beginning, whereas peripheral airways are sampled at later time periods.77 Tangedal et al found that microbiota compositions differed between pairwise spontaneous and induced sputum samples in participants with COPD.76 Further studies are needed in which protocols, and in particular sputum volume obtained, sample quality and procedure duration, are standardised across expectorated and induced sputum sampling.

    Limitations and challenges: reproducibility of findings

    COPD sputum microbiome studies to date have mainly used a cross-sectional design. Within-individual longitudinal microbial profiling reports are lagging, but data are starting to emerge. Sinha et al collected repeated induced sputum samples from four stable patients ith COPD over 9 months. They reported the stability of bacterial composition and diversity over 2 days, with increased variability over 9 months.78 In a larger longitudinal cohort (n=101), Mayhew et al found relative stability in microbial composition in a participant subset,51 while those with more variability were more likely to have higher exacerbation frequency.

    Overall, both within-individual and between-cohort reproducibility represent crucial milestones in the development of a biomarker and require further scientific attention and evaluation with respect to sputum microbiome profiles in COPD.79 80

    Future perspectives: integrate the non-bacterial microbiome

    16S rRNA sequencing approaches in sputum samples have provided important insights into the dynamics of the bacterial microbiome in stable COPD and exacerbations. However, the respiratory microbiome includes viruses and fungi as well as bacteria. The prevalence and clinical significance of fungi (mycobiome) in the respiratory tract in the context of COPD are particularly understudied.81 82 Bacterial-fungal and bacterial-viral interactions can affect microbial dynamics and microbe–host interactions and require further attention.83 84 Furthermore, the outgrowth of potentially pathogenic bacteria can be influenced by prior non-bacterial infections. Postviral bacterial respiratory infections are common overall. However, Molyneaux et al suggest that postviral bacterial dysbiosis is particularly an issue in COPD. They found that rhinovirus infection was followed by an outgrowth of H. influenzae in COPD sputum samples and associated with an increased neutrophilic inflammatory response.85 They did not observe this outgrowth in healthy controls.

    Overall, mechanisms of postviral secondary bacterial infections are known to be complex, mediated by interactions between viruses and bacteria (intermicrobial interactions) as well as the host immune system.86 It will be crucial to evaluate not only bacterial microbiome dynamics but viral and fungal dynamics as well, to evaluate the influence of inter-microbial interactions on both homoeostasis and pathogenesis within the respiratory tract. Shotgun metatranscriptomic approaches may be particularly well suited to these types of analyses as they can survey both bacterial and non-bacterial microbes and may provide an assessment of functional activity.

    Future perspectives: integrate the host response

    Although technically challenging, high-throughput sequencing techniques, such as metatranscriptomics, offer the possibility to simultaneously sequence the host and microbial genomes.36 In principle, host genome and microbiome profiles could be analysed in ‘one run’ in airway samples from COPD patients, offering the possibility to directly analyse the interactions between microbiota and the host immune profiles on a personal genome level. Langelier et al recently demonstrated how the integration of host response and microbiome profiling can contribute to diagnosis and characterisation of LRT infections (LRTI) in critically ill patients with acute respiratory failure, based on metagenomic and metatranscriptomic sequencing in tracheal aspirate samples.87 Combining pathogen, microbiome and host gene expression metrics enhanced LRTI diagnosis in cases of unknown aetiology, showcasing the potential of this model to optimise diagnostics for airway diseases.

    Conclusion

    In conclusion, high-throughput sequencing techniques allow for an increasingly detailed assessment of the COPD airway microbiome. Sputum likely only partially represents the respiratory microbiome in COPD, given the enrichment for oral flora. However, the increasing evidence of sputum microbiome associations with characteristics of stable COPD and exacerbations, treatment response, and outcomes suggest that sputum microbiome profiling may provide important biomarkers to better understand COPD biology and guide therapies. Many outstanding questions remain. Answering these questions will take exacting studies implementing standardised protocols and employing the evolving technologies and statistical approaches to study microbial community structure in relation to the COPD-relevant host response.

    References

    1. ↵
      2. Vos T ,
      3. Allen C ,
      4. Arora M , et al
      . Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990–2015: a systematic analysis for the global burden of disease study 2015. The Lancet 2016;388:1545–602.doi:10.1016/S0140-6736(16)31678-6
      OpenUrl
    2. ↵
      GOLD : Global Strategy for the Diagnosis, Management and Prevention of COPD, Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2018.
    3. ↵
      2. Seemungal TAR ,
      3. Donaldson GC ,
      4. Bhowmik A , et al
      . Time course and recovery of exacerbations in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2000;161:1608–13.doi:10.1164/ajrccm.161.5.9908022
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Donaldson GC ,
      3. Seemungal TA ,
      4. Bhowmik A
      . Relationship between exacerbation frequency and lung function decline in chronic obstructive pulmonary disease. Thorax 2002;57:847–52.doi:10.1136/thorax.57.10.847
      OpenUrlAbstract/FREE Full Text
    5. ↵
      2. Agusti A
      . The path to personalised medicine in COPD. Thorax 2014;69:857–64.doi:10.1136/thoraxjnl-2014-205507
      OpenUrlAbstract/FREE Full Text
    6. ↵
      2. Han MK ,
      3. Muellerova H ,
      4. Curran-Everett D , et al
      . Implications of the gold 2011 disease severity classification in the COPDGene cohort. Lancet Respir Med 2013;1:43–50.
      OpenUrl
    7. ↵
      2. Lopez-Campos JL ,
      3. Agustí A
      . Heterogeneity of chronic obstructive pulmonary disease exacerbations: a two-axes classification proposal. Lancet Respir Med 2015;3:729–34.doi:10.1016/S2213-2600(15)00242-8
      OpenUrl
    8. ↵
      2. Huang YJ ,
      3. Erb-Downward JR ,
      4. Dickson RP , et al
      . Understanding the role of the microbiome in chronic obstructive pulmonary disease: principles, challenges, and future directions. Transl Res 2017;179:71–83.doi:10.1016/j.trsl.2016.06.007
      OpenUrlCrossRefPubMed
    9. ↵
      2. Sze MA ,
      3. Dimitriu PA ,
      4. Suzuki M , et al
      . Host response to the lung microbiome in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2015;192:438–45.doi:10.1164/rccm.201502-0223OC
      OpenUrlCrossRefPubMed
    10. ↵
      2. Ghebre MA ,
      3. Pang PH ,
      4. Diver S , et al
      . Biological exacerbation clusters demonstrate asthma and chronic obstructive pulmonary disease overlap with distinct mediator and microbiome profiles. J Allergy Clin Immunol 2018;141:2027–36.doi:10.1016/j.jaci.2018.04.013
      OpenUrl
    11. ↵
      2. Wang Z ,
      3. Bafadhel M ,
      4. Haldar K , et al
      . Lung microbiome dynamics in COPD exacerbations. Eur Respir J 2016;47:1082–92.doi:10.1183/13993003.01406-2015
      OpenUrlAbstract/FREE Full Text
    12. ↵
      2. van der Does AM ,
      3. Amatngalim GD ,
      4. Keijser B , et al
      . Contribution of host defence proteins and peptides to Host-Microbiota interactions in chronic inflammatory lung diseases. Vaccines 2018;6:49.
      OpenUrl
    13. ↵
      2. Marchesi JR ,
      3. Ravel J
      . The vocabulary of microbiome research: a proposal. Microbiome 2015;3:1–3.doi:10.1186/s40168-015-0094-5
      OpenUrlCrossRefPubMed
    14. ↵
      2. Barker BL ,
      3. Haldar K ,
      4. Patel H , et al
      . Association between pathogens detected using quantitative polymerase chain reaction with airway inflammation in COPD at stable state and exacerbations. Chest 2015;147:46–55.doi:10.1378/chest.14-0764
      OpenUrlCrossRefPubMed
    15. ↵
      2. Einarsson GG ,
      3. Comer DM ,
      4. McIlreavey L , et al
      . Community dynamics and the lower airway microbiota in stable chronic obstructive pulmonary disease, smokers and healthy non-smokers. Thorax 2016;71:795–803.doi:10.1136/thoraxjnl-2015-207235
      OpenUrlAbstract/FREE Full Text
    16. ↵
      2. Yadava K ,
      3. Pattaroni C ,
      4. Sichelstiel AK , et al
      . Microbiota promotes chronic pulmonary inflammation by enhancing IL-17A and autoantibodies. Am J Respir Crit Care Med 2016;193:975–87.doi:10.1164/rccm.201504-0779OC
      OpenUrlCrossRefPubMed
    17. ↵
      2. Erb-Downward JR ,
      3. Thompson DL ,
      4. Han MK , et al
      . Analysis of the lung microbiome in the “healthy” smoker and in COPD. PLoS One 2011;6:e16384.doi:10.1371/journal.pone.0016384
    18. ↵
      2. Suau A ,
      3. Bonnet R ,
      4. Sutren M , et al
      . Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl Environ Microbiol 1999;65:4799–807.doi:10.1128/AEM.65.11.4799-4807.1999
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Human Microbiome Project Consortium
      . Structure, function and diversity of the healthy human microbiome. Nature 2013;486:207–14.
      OpenUrlWeb of Science
    20. ↵
      2. Sibley CD ,
      3. Grinwis ME ,
      4. Field TR , et al
      . Culture enriched molecular profiling of the cystic fibrosis airway microbiome. PLoS One 2011;6:e22702.doi:10.1371/journal.pone.0022702
    21. ↵
      2. Charlson ES ,
      3. Bittinger K ,
      4. Haas AR , et al
      . Topographical continuity of bacterial populations in the healthy human respiratory tract. Am J Respir Crit Care Med 2011;184:957–63.doi:10.1164/rccm.201104-0655OC
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      2. Eisenhofer R ,
      3. Minich JJ ,
      4. Marotz C , et al
      . Contamination in low microbial biomass microbiome studies: issues and recommendations. Trends Microbiol 2019;27:105–17.doi:10.1016/j.tim.2018.11.003
      OpenUrlCrossRef
    23. ↵
      2. Watson RL ,
      3. de Koff EM ,
      4. Bogaert D
      . Characterising the respiratory microbiome. Eur Respir J 2019;53:1801711.doi:10.1183/13993003.01711-2018
      OpenUrlFREE Full Text
    24. ↵
      2. Costea PI ,
      3. Zeller G ,
      4. Sunagawa S , et al
      . Towards standards for human fecal sample processing in metagenomic studies. Nat Biotechnol 2017;35:1069–76.doi:10.1038/nbt.3960
      OpenUrlCrossRef
    25. ↵
      2. Jovel J ,
      3. Patterson J ,
      4. Wang W , et al
      . Characterization of the gut microbiome using 16S or shotgun metagenomics. Front Microbiol 2016;7:459.doi:10.3389/fmicb.2016.00459
    26. ↵
      2. Knight R ,
      3. Vrbanac A ,
      4. Taylor BC , et al
      . Best practices for analysing microbiomes. Nat Rev Microbiol 2018;16:410–22.doi:10.1038/s41579-018-0029-9
      OpenUrl
    27. ↵
      2. Hilty M ,
      3. Burke C ,
      4. Pedro H , et al
      . Disordered microbial communities in asthmatic airways. PLoS One 2010;5:e8578.doi:10.1371/journal.pone.0008578
    28. ↵
      2. Janda JM ,
      3. Abbott SL
      . 16S rRNA gene sequencing for bacterial identification in the diagnostic laboratory: pluses, perils, and pitfalls. J Clin Microbiol 2007;45:2761–4.doi:10.1128/JCM.01228-07
      OpenUrlFREE Full Text
    29. ↵
      2. Wagner BD ,
      3. Grunwald GK ,
      4. Zerbe GO , et al
      . On the use of diversity measures in longitudinal sequencing studies of microbial communities. Front Microbiol 2018;9.doi:10.3389/fmicb.2018.01037
    30. ↵
      2. Gilbert JA ,
      3. Lynch SV
      . Community ecology as a framework for human microbiome research. Nat Med 2019;25:884–9.doi:10.1038/s41591-019-0464-9
      OpenUrl
    31. ↵
      2. Rogers GB ,
      3. Marsh P ,
      4. Stressmann AF , et al
      . The exclusion of dead bacterial cells is essential for accurate molecular analysis of clinical samples. Clin Microbiol Infect 2010;16:1656–8.doi:10.1111/j.1469-0691.2010.03189.x
      OpenUrlCrossRefPubMed
    32. ↵
      2. Scholz M ,
      3. Ward DV ,
      4. Pasolli E , et al
      . Strain-level microbial epidemiology and population genomics from shotgun metagenomics. Nat Methods 2016;13:435–8.doi:10.1038/nmeth.3802
      OpenUrlCrossRefPubMed
    33. ↵
      2. Feigelman R ,
      3. Kahlert CR ,
      4. Baty F , et al
      . Sputum DNA sequencing in cystic fibrosis: non-invasive access to the lung microbiome and to pathogen details. Microbiome 2017;5:20.doi:10.1186/s40168-017-0234-1
      OpenUrlCrossRef
    34. ↵
      2. Quince C ,
      3. Walker AW ,
      4. Simpson JT , et al
      . Shotgun metagenomics, from sampling to analysis. Nat Biotechnol 2017;35:833–44.doi:10.1038/nbt.3935
      OpenUrlCrossRef
    35. ↵
      2. Joice R ,
      3. Yasuda K ,
      4. Shafquat A , et al
      . Determining microbial products and identifying molecular targets in the human microbiome. Cell Metab 2014;20:731–41.doi:10.1016/j.cmet.2014.10.003
      OpenUrlCrossRefPubMed
    36. ↵
      2. Westermann AJ ,
      3. Gorski SA ,
      4. Vogel J
      . Dual RNA-seq of pathogen and host. Nat Rev Microbiol 2012;10:618–30.doi:10.1038/nrmicro2852
      OpenUrlCrossRefPubMed
    37. ↵
      2. Wesolowska-Andersen A ,
      3. Everman JL ,
      4. Davidson R , et al
      . Dual RNA-seq reveals viral infections in asthmatic children without respiratory illness which are associated with changes in the airway transcriptome. Genome Biol 2017;18:12.doi:10.1186/s13059-016-1140-8
    38. ↵
      2. Martinez X ,
      3. Pozuelo M ,
      4. Pascal V , et al
      . MetaTrans: an open-source pipeline for metatranscriptomics. Sci Rep 2016;6:26447.doi:10.1038/srep26447
    39. ↵
      2. Shakya M ,
      3. Lo C-C ,
      4. Chain PSG
      . Advances and challenges in Metatranscriptomic analysis. Front Genet 2019;10:904.doi:10.3389/fgene.2019.00904
    40. ↵
      2. Aldridge BB ,
      3. Rhee KY
      . Microbial metabolomics: innovation, application, insight. Curr Opin Microbiol 2014;19:90–6.doi:10.1016/j.mib.2014.06.009
      OpenUrl
    41. ↵
      2. Postler TS ,
      3. Ghosh S
      . Understanding the Holobiont: how microbial metabolites affect human health and shape the immune system. Cell Metab 2017;26:110–30.doi:10.1016/j.cmet.2017.05.008
      OpenUrl
    42. ↵
      2. Sharon G ,
      3. Garg N ,
      4. Debelius J , et al
      . Specialized metabolites from the microbiome in health and disease. Cell Metab 2014;20:719–30.doi:10.1016/j.cmet.2014.10.016
      OpenUrlCrossRefPubMed
    43. ↵
      2. Ren L ,
      3. Zhang R ,
      4. Rao J , et al
      . Transcriptionally active lung microbiome and its association with bacterial biomass and host inflammatory status. mSystems 2018;3:e00199–18.doi:10.1128/mSystems.00199-18
      OpenUrl
    44. ↵
      2. Man WH ,
      3. de Steenhuijsen Piters WAA ,
      4. Bogaert D
      . The microbiota of the respiratory tract: gatekeeper to respiratory health. Nat Rev Microbiol 2017;15:259–70.doi:10.1038/nrmicro.2017.14
      OpenUrlCrossRef
    45. ↵
      2. Cabrera-Rubio R ,
      3. Garcia-Núñez M ,
      4. Setó L , et al
      . Microbiome diversity in the bronchial tracts of patients with chronic obstructive pulmonary disease. J Clin Microbiol 2012;50:3562–8.doi:10.1128/JCM.00767-12
      OpenUrlAbstract/FREE Full Text
    46. ↵
      2. Dickson RP ,
      3. Erb-Downward JR ,
      4. Freeman CM , et al
      . Bacterial topography of the healthy human lower respiratory tract. MBio 2017;8:e02287–16.doi:10.1128/mBio.02287-16
      OpenUrl
    47. ↵
      2. Durack J ,
      3. Huang YJ ,
      4. Nariya S , et al
      . Bacterial biogeography of adult airways in atopic asthma. Microbiome 2018;6:104.doi:10.1186/s40168-018-0487-3
      OpenUrl
    48. ↵
      2. Wang Z ,
      3. Singh R ,
      4. Miller BE , 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.doi:10.1136/thoraxjnl-2017-210741
      OpenUrlAbstract/FREE Full Text
    49. ↵
      2. Galiana A ,
      3. Aguirre E ,
      4. Rodriguez JC , et al
      . Sputum microbiota in moderate versus severe patients with COPD. Eur Respir J 2014;43:1787–90.doi:10.1183/09031936.00191513
      OpenUrlFREE Full Text
    50. ↵
      2. Garcia-Nunez M ,
      3. Millares L ,
      4. Pomares X , et al
      . Severity-Related changes of bronchial microbiome in chronic obstructive pulmonary disease. J Clin Microbiol 2014;52:4217–23.doi:10.1128/JCM.01967-14
      OpenUrlAbstract/FREE Full Text
    51. ↵
      2. Mayhew D ,
      3. Devos N ,
      4. Lambert C , et al
      . Longitudinal profiling of the lung microbiome in the AERIS study demonstrates repeatability of bacterial and eosinophilic COPD exacerbations. Thorax 2018;73:422–30.doi:10.1136/thoraxjnl-2017-210408
      OpenUrlAbstract/FREE Full Text
    52. ↵
      2. Ni J ,
      3. Wu GD ,
      4. Albenberg L , et al
      . Gut microbiota and IBD: causation or correlation? Nat Rev Gastroenterol Hepatol 2017;14:573–84.doi:10.1038/nrgastro.2017.88
      OpenUrl
    53. ↵
      2. Garcha DS ,
      3. Thurston SJ ,
      4. Patel ARC , et al
      . Changes in prevalence and load of airway bacteria using quantitative PCR in stable and exacerbated COPD. Thorax 2012;67:1075–80.doi:10.1136/thoraxjnl-2012-201924
      OpenUrlAbstract/FREE Full Text
    54. ↵
      2. Brill SE ,
      3. Law M ,
      4. El-Emir E , et al
      . Effects of different antibiotic classes on airway bacteria in stable COPD using culture and molecular techniques: a randomised controlled trial. Thorax 2015;70:930–8.doi:10.1136/thoraxjnl-2015-207194
      OpenUrlAbstract/FREE Full Text
    55. ↵
      2. Biesbroek G ,
      3. Tsivtsivadze E ,
      4. Sanders EAM , et al
      . Early respiratory microbiota composition determines bacterial succession patterns and respiratory health in children. Am J Respir Crit Care Med 2014;190:1283–92.doi:10.1164/rccm.201407-1240OC
      OpenUrlCrossRefPubMed
    56. ↵
      2. Bomar L ,
      3. Brugger SD ,
      4. Yost BH , et al
      . Corynebacterium accolens Releases Antipneumococcal Free Fatty Acids from Human Nostril and Skin Surface Triacylglycerols. MBio 2016;7:e01725–15.doi:10.1128/mBio.01725-15
      OpenUrlCrossRef
    57. ↵
      2. Pettigrew MM ,
      3. Laufer AS ,
      4. Gent JF , et al
      . Upper respiratory tract microbial communities, acute otitis media pathogens, and antibiotic use in healthy and sick children. Appl Environ Microbiol 2012;78:6262–70.doi:10.1128/AEM.01051-12
      OpenUrlAbstract/FREE Full Text
    58. ↵
      2. Iwase T ,
      3. Uehara Y ,
      4. Shinji H , et al
      . Staphylococcus epidermidis ESP inhibits Staphylococcus aureus biofilm formation and nasal colonization. Nature 2010;465:346–9.doi:10.1038/nature09074
      OpenUrlCrossRefPubMedWeb of Science
    59. ↵
      2. Jain S ,
      3. Self WH ,
      4. Wunderink RG , et al
      . Community-acquired pneumonia requiring hospitalization among U.S. adults. N Engl J Med 2015;373:415–27.doi:10.1056/NEJMoa1500245
      OpenUrlCrossRefPubMed
    60. ↵
      2. Matkovic Z ,
      3. Miravitlles M
      . Chronic bronchial infection in COPD. Is there an infective phenotype? Respir Med 2013;107:10–22.doi:10.1016/j.rmed.2012.10.024
      OpenUrlCrossRefPubMed
    61. ↵
      2. Zaas AK ,
      3. Garner BH ,
      4. Tsalik EL , et al
      . The current epidemiology and clinical decisions surrounding acute respiratory infections. Trends Mol Med 2014;20:579–88.doi:10.1016/j.molmed.2014.08.001
      OpenUrlCrossRefPubMed
    62. ↵
      2. Huang YJ ,
      3. Sethi S ,
      4. Murphy T , et al
      . Airway microbiome dynamics in exacerbations of chronic obstructive pulmonary disease. J Clin Microbiol 2014;52:2813–23.doi:10.1128/JCM.00035-14
      OpenUrlAbstract/FREE Full Text
    63. ↵
      2. Sethi S ,
      3. Evans N ,
      4. Grant BJB , et al
      . New strains of bacteria and exacerbations of chronic obstructive pulmonary disease. N Engl J Med 2002;347:465–71.doi:10.1056/NEJMoa012561
      OpenUrlCrossRefPubMedWeb of Science
    64. ↵
      2. Leitao Filho FS ,
      3. Alotaibi NM ,
      4. Ngan D , et al
      . Sputum microbiome is associated with 1-year mortality after chronic obstructive pulmonary disease hospitalizations. Am J Respir Crit Care Med 2019;199:1205–13.doi:10.1164/rccm.201806-1135OC
      OpenUrl
    65. ↵
      2. Acosta N ,
      3. Heirali A ,
      4. Somayaji R , et al
      . Sputum microbiota is predictive of long-term clinical outcomes in young adults with cystic fibrosis. Thorax 2018;73:1016–25.doi:10.1136/thoraxjnl-2018-211510
      OpenUrlAbstract/FREE Full Text
    66. ↵
      2. Hajishengallis G ,
      3. Darveau RP ,
      4. Curtis MA
      . The keystone-pathogen hypothesis. Nat Rev Microbiol 2012;10:717–25.doi:10.1038/nrmicro2873
      OpenUrlCrossRefPubMed
    67. ↵
      2. Millares L ,
      3. Pérez-Brocal V ,
      4. Ferrari R , et al
      . Functional metagenomics of the bronchial microbiome in COPD. PLoS One 2015;10:e0144448.doi:10.1371/journal.pone.0144448
    68. ↵
      2. Fischbach MA
      . Microbiome: focus on causation and mechanism. Cell 2018;174:785–90.doi:10.1016/j.cell.2018.07.038
      OpenUrlCrossRefPubMed
    69. ↵
      2. Lederer DJ ,
      3. Bell SC ,
      4. Branson RD , et al
      . Control of confounding and reporting of results in causal inference studies. Ann Am Thorac Soc 2019;16:22–8.
      OpenUrl
    70. ↵
      2. Sanna S ,
      3. van Zuydam NR ,
      4. Mahajan A , et al
      . Causal relationships among the gut microbiome, short-chain fatty acids and metabolic diseases. Nat Genet 2019;51:600–5.doi:10.1038/s41588-019-0350-x
      OpenUrl
    71. ↵
      2. Segal LN ,
      3. Clemente JC ,
      4. Wu BG , et al
      . Randomised, double-blind, placebo-controlled trial with azithromycin selects for anti-inflammatory microbial metabolites in the emphysematous lung. Thorax 2017;72:13–22.doi:10.1136/thoraxjnl-2016-208599
      OpenUrlAbstract/FREE Full Text
    72. ↵
      2. Sulaiman I ,
      3. Wu BG ,
      4. Li Y , et al
      . Evaluation of the airway microbiome in nontuberculous mycobacteria disease. Eur Respir J 2018;52:1800810.doi:10.1183/13993003.00810-2018
      OpenUrlAbstract/FREE Full Text
    73. ↵
      2. Cui L ,
      3. Lucht L ,
      4. Tipton L , et al
      . Topographic diversity of the respiratory tract mycobiome and alteration in HIV and lung disease. Am J Respir Crit Care Med 2015;191:932–42.doi:10.1164/rccm.201409-1583OC
      OpenUrlCrossRefPubMed
    74. ↵
      2. Bathoorn E ,
      3. Liesker J ,
      4. Postma D , et al
      . Safety of sputum induction during exacerbations of COPD. Chest 2007;131:432–8.doi:10.1378/chest.06-2216
      OpenUrlCrossRefPubMed
    75. ↵
      2. Brightling CE ,
      3. Monterio W ,
      4. Green RH , et al
      . Induced sputum and other outcome measures in chronic obstructive pulmonary disease: safety and repeatability. Respir Med 2001;95:999–1002.doi:10.1053/rmed.2001.1195
      OpenUrlCrossRefPubMedWeb of Science
    76. ↵
      2. Tangedal S ,
      3. Aanerud M ,
      4. Grønseth R , et al
      . Comparing microbiota profiles in induced and spontaneous sputum samples in COPD patients. Respir Res 2017;18:164.doi:10.1186/s12931-017-0645-3
    77. ↵
      2. Gershman NH ,
      3. Liu H ,
      4. Wong HH , et al
      . Fractional analysis of sequential induced sputum samples during sputum induction: evidence that different lung compartments are sampled at different time points☆☆☆. J Allergy Clin Immunol 1999;104:322–8.doi:10.1016/S0091-6749(99)70374-X
      OpenUrlCrossRefPubMedWeb of Science
    78. ↵
      2. Sinha R ,
      3. Weissenburger-Moser LA ,
      4. Clarke JL , et al
      . Short term dynamics of the sputum microbiome among COPD patients. PLoS One 2018;13:e0191499.doi:10.1371/journal.pone.0191499
    79. ↵
      2. Vargas AJ ,
      3. Harris CC
      . Biomarker development in the precision medicine era: lung cancer as a case study. Nat Rev Cancer 2016;16:525–37.doi:10.1038/nrc.2016.56
      OpenUrlCrossRefPubMed
    80. ↵
      2. Selleck MJ ,
      3. Senthil M ,
      4. Wall NR
      . Making meaningful clinical use of biomarkers. Biomark Insights 2017;12:1177271917715236.doi:10.1177/1177271917715236
    81. ↵
      2. Bafadhel M ,
      3. Mckenna S ,
      4. Agbetile J , et al
      . Aspergillus fumigatus during stable state and exacerbations of COPD. Eur Respir J 2014;43:64–71.doi:10.1183/09031936.00162912
      OpenUrlAbstract/FREE Full Text
    82. ↵
      2. Nguyen LDN ,
      3. Viscogliosi E ,
      4. Delhaes L
      . The lung mycobiome: an emerging field of the human respiratory microbiome. Front Microbiol 2015;6:89.doi:10.3389/fmicb.2015.00089
    83. ↵
      2. Deveau A ,
      3. Bonito G ,
      4. Uehling J , et al
      . Bacterial–fungal interactions: ecology, mechanisms and challenges. FEMS Microbiol Rev 2018;42:335–52.doi:10.1093/femsre/fuy008
      OpenUrlCrossRef
    84. ↵
      2. Almand EA ,
      3. Moore MD ,
      4. Jaykus L-A
      . Virus-Bacteria interactions: an emerging topic in human infection. Viruses 2017;9:58.doi:10.3390/v9030058
    85. ↵
      2. Molyneaux PL ,
      3. Mallia P ,
      4. Cox MJ , et al
      . Outgrowth of the bacterial airway microbiome after rhinovirus exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2013;188:1224–31.doi:10.1164/rccm.201302-0341OC
      OpenUrlCrossRefPubMedWeb of Science
    86. ↵
      2. Hanada S ,
      3. Pirzadeh M ,
      4. Carver KY , et al
      . Respiratory viral infection-induced microbiome alterations and secondary bacterial pneumonia. Front Immunol 2018;9:2640.doi:10.3389/fimmu.2018.02640
    87. ↵
      2. Langelier C ,
      3. Kalantar KL ,
      4. Moazed F , et al
      . Integrating host response and unbiased microbe detection for lower respiratory tract infection diagnosis in critically ill adults. Proc Natl Acad Sci U S A 2018;115:E12353–62.doi:10.1073/pnas.1809700115
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors BD wrote the initial draft of the manuscript with additional content provided by SC, CB and AF and critical revisions from all authors.

    • Funding The submitted work is cofinanced by the Ministry of Economic Affairs and Climate Policy by means of the PPP. AF was supported by a junior longfond grant (4.2.16.132JO).

    • Competing interests SC reports personal fees from AstraZeneca, personal fees from GlaxoSmithKline, personal fees from Amgen, personal fees from Glenmark, personal fees from Sunovion, personal fees and non-financial support from Genentech, non-financial support from Medimmune, personal fees from UpToDate, outside the submitted work. CB reports grants and personal fees from GSK, grants and personal fees from AZ/MedImmune, grants and personal fees from Novartis, grants and personal fees from Chiesi, grants from Roche/Genentech, grants and personal fees from Mologic, grants and personal fees from Gossamer, grants and personal fees from BI, grants and personal fees from 4DPharma, personal fees from Sanofi, personal fees from Regeneron, personal fees from Theravance, outside the submitted work. JR reports personal fees from IDbyDNA, from null, outside the submitted work. HAMK reports an unrestricted research grant, and fees for participation in advisory boards from GlaxoSmithKline, the sponsor of this study as well as from Boehringer Ingelheim, and Novartis. He also reports fees advisory board participation for AstraZeneca and Chiesi. All above paid to his institution,outside the submitted work. BD, AF and MvdB have nothing to disclose

    • Patient consent for publication Not required.

    • Provenance and peer review Not commissioned; externally peer reviewed.