Challenges in identifying asthma subgroups using unsupervised statistical learning techniques

Cross-Talk of Multiple Types of RNA Modification Regulators Uncovers the Tumor Microenvironment and Immune Infiltrates in Soft Tissue Sarcoma

Background

Soft-tissue sarcoma (STS) represents a rare and diverse cohort of solid tumors, and encompasses over 100 various histologic and molecular subtypes. In recent years, RNA modifications including m6A, m5C, m1A, and m7G have been demonstrated to regulate immune response and tumorigenesis. Nevertheless, the cross-talk among these RNA modification regulators and related effects upon the tumor microenvironment (TME), immune infiltrates, and immunotherapy in STS remain poorly understood.

Methods

In this study, we comprehensively investigated transcriptional and genetic alterations of 32 RNA modification regulators in STS patients from The Cancer Genome Atlas (TCGA) cohort and validated them in the Gene Expression Omnibus (GEO) cohort. Single-cell transcriptomes were introduced to identify regulators within specific cell types, with own sequencing data and RT-qPCR conducted for biological validation. Distinct regulator clusters and regulator gene subtypes were identified by using unsupervised consensus clustering analysis. We further built the regulator score model based on the prognostic regulator-related differentially expressed genes (DEGs), which could be used to quantitatively assess the risk for individual STS patients. The clinical and biological characteristics of different regulator score groups were further examined.

Results

A total of 455 patients with STS were included in this analysis. The network of 32 RNA modification regulators demonstrated significant correlations within multiple different RNA modification types. Distinct regulator clusters and regulator gene subtypes were characterized by markedly different prognoses and TME landscapes. The low regulator score group in the TCGA-SARC cohort was characterized by poor prognosis. The robustness of the scoring model was further confirmed by the external validation in GSE30929 and GSE17674. The regulator score was negatively correlated with CD4+ T cell, Th2 cell, and Treg cell recruitment and most immunotherapy-predicted pathways, and was also associated with immunotherapy efficacy.

Conclusions

Overall, our study is the first to demonstrate the cross-talk of RNA modification regulators and the potential roles in TME and immune infiltrates in STS. The individualized assessment based on the regulator score model could facilitate and optimize personalized treatment.

Treatment outcome clustering patterns correspond to discrete asthma phenotypes in children

Despite widely and regularly used therapy asthma in children is not fully controlled. Recognizing the complexity of asthma phenotypes and endotypes imposed the concept of precision medicine in asthma treatment. By applying machine learning algorithms assessed with respect to their accuracy in predicting treatment outcome, we have successfully identified 4 distinct clusters in a pediatric asthma cohort with specific treatment outcome patterns according to changes in lung function (FEV₁ and MEF₅₀), airway inflammation (FENO) and disease control likely affected by discrete phenotypes at initial disease presentation, differing in the type and level of inflammation, age of onset, comorbidities, certain genetic and other physiologic traits. The smallest and the largest of the 4 clusters- 1 (N = 58) and 3 (N = 138) had better treatment outcomes compared to clusters 2 and 4 and were characterized by more prominent atopic markers and a predominant allelic (A allele) effect for rs37973 in the GLCCI1 gene previously associated with positive treatment outcomes in asthmatics. These patients also had a relatively later onset of disease (6 + yrs). Clusters 2 (N = 87) and 4 (N = 64) had poorer treatment success, but varied in the type of inflammation (predominantly neutrophilic for cluster 4 and likely mixed-type for cluster 2), comorbidities (obesity for cluster 2), level of systemic inflammation (highest hsCRP for cluster 2) and platelet count (lowest for cluster 4). The results of this study emphasize the issues in asthma management due to the overgeneralized approach to the disease, not taking into account specific disease phenotypes.

Asthma clustering methods: a literature-informed application to the children's health study data

OBJECTIVE

The heterogeneity of asthma has inspired widespread application of statistical clustering algorithms to a variety of datasets for identification of potentially clinically meaningful phenotypes. There has not been a standardized data analysis approach for asthma clustering, which can affect reproducibility and clinical translation of results. Our objective was to identify common and effective data analysis practices in the asthma clustering literature and apply them to data from a Southern California population-based cohort of schoolchildren with asthma.

METHODS

As of January 1, 2020, we reviewed key statistical elements of 77 asthma clustering studies. Guided by the literature, we used 12 input variables and three clustering methods (hierarchical clustering, k-medoids, and latent class analysis) to identify clusters in 598 schoolchildren with asthma from the Southern California Children's Health Study (CHS).

RESULTS

Clusters of children identified by latent class analysis were characterized by exhaled nitric oxide, FEV₁/FVC, FEV₁ percent predicted, asthma control and allergy score; and were predictive of control at two year follow up. Clusters from the other two methods were less clinically remarkable, primarily differentiated by sex and race/ethnicity and less predictive of asthma control over time.

CONCLUSION

Upon review of the asthma phenotyping literature, common approaches of data clustering emerged. When applying these elements to the Children's Health Study data, latent class analysis clusters-represented by exhaled nitric oxide and spirometry measures-had clinical relevance over time.

Artificial intelligence in the diagnosis of pediatric allergic diseases

Artificial intelligence (AI) is a field of data science pertaining to advanced computing machines capable of learning from data and interacting with the human world. Early diagnosis and diagnostics, self-care, prevention and wellness, clinical decision support, care delivery, and chronic care management have been identified within the healthcare areas that could benefit from introducing AI. In pediatric allergy research, the recent developments in AI approach provided new perspectives for characterizing the heterogeneity of allergic diseases among patients. Moreover, the increasing use of electronic health records and personal healthcare records highlighted the relevance of AI in improving data quality and processing and setting-up advanced algorithms to interpret the data. This review aimed to summarize current knowledge about AI and discuss its impact on the diagnostic framework of pediatric allergic diseases such as eczema, food allergy, and respiratory allergy, along with the future opportunities that AI research can offer in this medical area.

Machine learning in asthma research: moving toward a more integrated approach

Introduction: Big data are reshaping the future of medicine. The growing availability and increasing complexity of data have favored the adoption of modern analytical and computational methodologies in every area of medicine. Over the past decades, asthma research has been characterized by a shift in the way studies are conducted and data are analyzed. Motivated by the assumptions that 'data will speak for themselves', hypothesis-driven approaches have been replaced by data-driven hypotheses-generating methods to explore hidden patterns and underlying mechanisms. However, even with all the advancement in technologies and the new important insight that we gained to understand and characterize asthma heterogeneity, very few research findings have been translated into clinically actionable solutions.Areas covered: To investigate some of the fundamental analytical approaches adopted in the current literature and appraise their impact and usefulness in medicine, we conducted a bibliometric analysis of big data analytics in asthma research in the past 50 years.Expert opinion: No single data source or methodology can uncover the complexity of human health and disease. To fully capitalize on the potential of 'big data', we will have to embrace the collaborative science and encourage the creation of integrated cross-disciplinary teams brought together around technological advances.

Endotypes identified by cluster analysis in asthmatics and non-asthmatics and their clinical characteristics at follow-up: the case-control EGEA study

Background

Identifying relevant asthma endotypes may be the first step towards improving asthma management. We aimed identifying respiratory endotypes in adults using a cluster analysis and to compare their clinical characteristics at follow-up.

Methods

The analysis was performed separately among current asthmatics (CA, n=402) and never asthmatics (NA, n=666) from the first follow-up of the French EGEA study (EGEA2). Cluster analysis jointly considered 4 demographic, 22 clinical/functional (respiratory symptoms, asthma treatments, lung function) and four blood biological (allergy-related, inflammation-related and oxidative stress-related biomarkers) characteristics at EGEA2. The clinical characteristics at follow-up (EGEA3) were compared according to the endotype identified at EGEA2.

Results

We identified five respiratory endotypes, three among CA and two among NA: CA1 (n=53) with active treated adult-onset asthma, poor lung function, chronic cough and phlegm and dyspnoea, high body mass index, and high blood neutrophil count and fluorescent oxidation products level; CA2 (n=219) with mild asthma and rhinitis; CA3 (n=130) with inactive/mild untreated allergic childhood-onset asthma, high frequency of current smokers and low frequency of attacks of breathlessness at rest, and high IgE level; NA1 (n=489) asymptomatic, and NA2 (n=177) with respiratory symptoms, high blood neutrophil and eosinophil counts. CA1 had poor asthma control and high leptin level, CA2 had hyper-responsiveness and high interleukin (IL)-1Ra, IL-5, IL-7, IL-8, IL-10, IL-13 and TNF-α levels, and NA2 had high leptin and C reactive protein levels. Ten years later, asthmatics in CA1 had worse clinical characteristics whereas those in CA3 had better respiratory outcomes than CA2; NA in NA2 had more respiratory symptoms and higher rate of incident asthma than those in NA1.

Conclusion

These results highlight the interest to jointly consider clinical and biological characteristics in cluster analyses to identify endotypes among adults with or without asthma.

Challenges of Clustering Multimodal Clinical Data: Review of Applications in Asthma Subtyping

Background

In the current era of personalized medicine, there is increasing interest in understanding the heterogeneity in disease populations. Cluster analysis is a method commonly used to identify subtypes in heterogeneous disease populations. The clinical data used in such applications are typically multimodal, which can make the application of traditional cluster analysis methods challenging.

Objective

This study aimed to review the research literature on the application of clustering multimodal clinical data to identify asthma subtypes. We assessed common problems and shortcomings in the application of cluster analysis methods in determining asthma subtypes, such that they can be brought to the attention of the research community and avoided in future studies.

Methods

We searched PubMed and Scopus bibliographic databases with terms related to cluster analysis and asthma to identify studies that applied dissimilarity-based cluster analysis methods. We recorded the analytic methods used in each study at each step of the cluster analysis process.

Results

Our literature search identified 63 studies that applied cluster analysis to multimodal clinical data to identify asthma subtypes. The features fed into the cluster algorithms were of a mixed type in 47 (75%) studies and continuous in 12 (19%), and the feature type was unclear in the remaining 4 (6%) studies. A total of 23 (37%) studies used hierarchical clustering with Ward linkage, and 22 (35%) studies used k-means clustering. Of these 45 studies, 39 had mixed-type features, but only 5 specified dissimilarity measures that could handle mixed-type features. A further 9 (14%) studies used a preclustering step to create small clusters to feed on a hierarchical method. The original sample sizes in these 9 studies ranged from 84 to 349. The remaining studies used hierarchical clustering with other linkages (n=3), medoid-based methods (n=3), spectral clustering (n=1), and multiple kernel k-means clustering (n=1), and in 1 study, the methods were unclear. Of 63 studies, 54 (86%) explained the methods used to determine the number of clusters, 24 (38%) studies tested the quality of their cluster solution, and 11 (17%) studies tested the stability of their solution. Reporting of the cluster analysis was generally poor in terms of the methods employed and their justification.

Conclusions

This review highlights common issues in the application of cluster analysis to multimodal clinical data to identify asthma subtypes. Some of these issues were related to the multimodal nature of the data, but many were more general issues in the application of cluster analysis. Although cluster analysis may be a useful tool for investigating disease subtypes, we recommend that future studies carefully consider the implications of clustering multimodal data, the cluster analysis process itself, and the reporting of methods to facilitate replication and interpretation of findings.

Heterogeneity of Mild to Moderate Persistent Asthma in Children: Confirmation by Latent Class Analysis and Association with 1-Year Outcomes

BACKGROUND

Compared with adults, phenotypic characterization of children with asthma is still limited and it remains difficult to predict which children with asthma are at highest risk for poor outcomes.

OBJECTIVE

To identify latent classes in a large population of treatment-adherent children with mild to moderate asthma enrolled in clinical trials and determine whether latent class assignment predicts future lung function abnormalities and exacerbation rate.

METHODS

Latent class analysis was performed on 2593 children with mild to moderate asthma aged 5 18 years, with 19 variables encompassing demographic characteristics, medical history, symptoms, lung function, allergic sensitization, and type 2 inflammation. Outcomes included lung function and the annualized exacerbation rate at 12 months of follow-up.

RESULTS

Five latent classes were identified with differing demographic features, asthma control, sensitization, type 2 inflammatory markers, and lung function. Exacerbation rates were 1.30 ± 0.12 for class 1 (multiple sensitization with partially reversible airflow limitation), 0.90 ± 0.05 for class 2 (multiple sensitization with reversible airflow limitation), 0.87 ± 0.08 for class 3 (lesser sensitization with reversible airflow limitation), 0.87 ± 0.05 for class 4 (multiple sensitization with normal lung function), and 0.71 ± 0.06 for class 5 (lesser sensitization with normal lung function). Lung function abnormalities persisted in class 1 at 12 months.

CONCLUSIONS

Children with mild to moderate asthma are a heterogeneous group. Allergic sensitization and lung function may be particularly useful in identifying children at the greatest risk for future exacerbation. Additional studies are needed to determine whether latent classes correspond to meaningful phenotypes for the purpose of personalized treatment.

Longitudinal trajectories of severe wheeze exacerbations from infancy to school age and their association with early-life risk factors and late asthma outcomes

INTRODUCTION

Exacerbation-prone asthma subtype has been reported in studies using data-driven methodologies. However, patterns of severe exacerbations have not been studied.

OBJECTIVE

To investigate longitudinal trajectories of severe wheeze exacerbations from infancy to school age.

METHODS

We applied longitudinal k-means clustering to derive exacerbation trajectories among 887 participants from a population-based birth cohort with severe wheeze exacerbations confirmed in healthcare records. We examined early-life risk factors of the derived trajectories, and their asthma-related outcomes and lung function in adolescence.

RESULTS

498/887 children (56%) had physician-confirmed wheeze by age 8 years, of whom 160 had at least one severe exacerbation. A two-cluster model provided the optimal solution for severe exacerbation trajectories among these 160 children: "Infrequent exacerbations (IE)" (n = 150, 93.7%) and "Early-onset frequent exacerbations (FE)" (n = 10, 6.3%). Shorter duration of breastfeeding was the strongest early-life risk factor for FE (weeks, median [IQR]: FE, 0 [0-1.75] vs. IE, 6 [0-20], P < .001). Specific airway resistance (sR_aw ) was significantly higher in FE compared with IE trajectory throughout childhood. We then compared children in the two exacerbation trajectories with those who have never wheezed (NW, n = 389) or have wheezed but had no severe exacerbations (WNE, n = 338). At age 8 years, FEV₁ /FVC was significantly lower and FeNO significantly higher among FE children compared with all other groups. By adolescence (age 16), subjects in FE trajectory were significantly more likely to have current asthma (67% FE vs. 30% IE vs. 13% WNE, P < .001) and use inhaled corticosteroids (77% FE vs. 15% IE vs. 18% WNE, P < .001). Lung function was significantly diminished in the FE trajectory (FEV₁ /FVC, mean [95%CI]: 89.9% [89.3-90.5] vs. 88.1% [87.3-88.8] vs. 85.1% [83.4-86.7] vs. 74.7% [61.5-87.8], NW, WNE, IE, FE respectively, P < .001).

CONCLUSION

We have identified two distinct trajectories of severe exacerbations during childhood with different early-life risk factors and asthma-related outcomes in adolescence.

Childhood Asthma: Advances Using Machine Learning and Mechanistic Studies

A paradigm shift brought by the recognition that childhood asthma is an aggregated diagnosis that comprises several different endotypes underpinned by different pathophysiology, coupled with advances in understanding potentially important causal mechanisms, offers a real opportunity for a step change to reduce the burden of the disease on individual children, families, and society. Data-driven methodologies facilitate the discovery of "hidden" structures within "big healthcare data" to help generate new hypotheses. These findings can be translated into clinical practice by linking discovered "phenotypes" to specific mechanisms and clinical presentations. Epidemiological studies have provided important clues about mechanistic avenues that should be pursued to identify interventions to prevent the development or alter the natural history of asthma-related diseases. Findings from cohort studies followed by mechanistic studies in humans and in neonatal mouse models provided evidence that environments such as traditional farming may offer protection by modulating innate immune responses and that impaired innate immunity may increase susceptibility. The key question of which component of these exposures can be translated into interventions requires confirmation. Increasing mechanistic evidence is demonstrating that shaping the microbiome in early life may modulate immune function to confer protection. Iterative dialogue and continuous interaction between experts with different but complementary skill sets, including data scientists who generate information about the hidden structures within "big data" assets, and medical professionals, epidemiologists, basic scientists, and geneticists who provide critical clinical and mechanistic insights about the mechanisms underpinning the architecture of the heterogeneity, are keys to delivering mechanism-based stratified treatments and prevention.

Differential associations of allergic disease genetic variants with developmental profiles of eczema, wheeze and rhinitis

BACKGROUND

Allergic diseases (eczema, wheeze and rhinitis) in children often present as heterogeneous phenotypes. Understanding genetic associations of specific patterns of symptoms might facilitate understanding of the underlying biological mechanisms.

OBJECTIVE

To examine associations between allergic disease-related variants identified in a recent genome-wide association study and latent classes of allergic diseases (LCADs) in two population-based birth cohorts.

METHODS

Eight previously defined LCADs between birth and 11 years: "No disease," "Atopic march," "Persistent eczema and wheeze," "Persistent eczema with later-onset rhinitis," "Persistent wheeze with later-onset rhinitis," "Transient wheeze," "Eczema only" and "Rhinitis only" were used as the study outcome. Weighted multinomial logistic regression was used to estimate associations between 135 SNPs (and a polygenic risk score, PRS) and LCADs among 6345 individuals from The Avon Longitudinal Study of Parents and Children (ALSPAC). Heterogeneity across LCADs was assessed before and after Bonferroni correction. Results were replicated in Manchester Asthma and Allergy Study (MAAS) (n = 896) and pooled in a meta-analysis.

RESULTS

We found strong evidence for differential genetic associations across the LCADs; pooled PRS heterogeneity P-value = 3.3 × 10-14 , excluding "no disease" class. The associations between the PRS and LCADs in MAAS were remarkably similar to ALSPAC. Two SNPs (a protein-truncating variant in FLG and a SNP within an intron of GSDMB) had evidence for differential association (pooled P-values ≤ 0.006). The FLG locus was differentially associated across LCADs that included eczema, with stronger associations for LCADs with comorbid wheeze and rhinitis. The GSDMB locus in contrast was equally associated across LCADs that included wheeze.

CONCLUSIONS AND CLINICAL RELEVANCE

We have shown complex, but distinct patterns of genetic associations with LCADs, suggesting that heterogeneous mechanisms underlie individual disease trajectories. Establishing the combination of allergic diseases with which each genetic variant is associated may inform therapeutic development and/or predictive modelling.

Challenges and choices in the pharmacological treatment of non-severe pediatric asthma: A commentary for the practicing physician

In recent years, asthma research has focused intensely on the severe part of the disease spectrum, leading to new treatments, mostly therapeutic monoclonal antibodies. However, severe asthma accounts for not more than 2% of asthma in the pediatric population. Therefore, non-severe asthma remains a major health problem in children, not only for patients and parents but also for healthcare professionals such as general practitioners, pediatricians and allergists who take care of these patients. It is thus essential to identify and put in context novel concepts, applicable to the treatment of these patients. Recent evidence suggests benefits from using anti-inflammatory treatment even for the mildest cases, for whom until now only symptomatic bronchodilation was recommended. Likewise, “reliever” medication may be better combined with an inhaled corticosteroid (ICS). Among “new” treatments (for children), ICS formulation in ultrafine particles has showed promise and tiotropium is gaining access to the pediatric population. Maintenance and reliever therapy (MART) is an option for moderate disease. Most importantly, personalized response to medications appears to be considerable, therefore, it may need to be taken into account. Overall, these new options provide opportunities for multiple new management strategies. The deployment of such strategies in different populations remains to be evaluated.

Data-driven adult asthma phenotypes based on clinical characteristics are associated with asthma outcomes twenty years later

BACKGROUND

Research based on cluster analyses led to the identification of particular phenotypes confirming phenotypic heterogeneity of asthma. The long-term clinical course of asthma phenotypes defined by clustering analysis remains unknown, although it is a key aspect to underpin their clinical relevance. We aimed to estimate risk of poor asthma events between asthma clusters identified 20 years earlier.

METHODS

The study relied on two cohorts of adults with asthma with 20-year follow-up, ECRHS (European Community Respiratory Health Survey) and EGEA (Epidemiological study on Genetics and Environment of Asthma). Regression models were used to compare asthma characteristics (current asthma, asthma exacerbations, asthma control, quality of life, and FEV₁ ) at follow-up and the course of FEV₁  between seven cluster-based asthma phenotypes identified 20 years earlier.

RESULTS

The analysis included 1325 adults with ever asthma. For each asthma characteristic assessed at follow-up, the risk for adverse outcomes differed significantly between the seven asthma clusters identified at baseline. As compared with the mildest asthma phenotype, ORs (95% CI) for asthma exacerbations varied from 0.9 (0.4 to 2.0) to 4.0 (2.0 to 7.8) and the regression estimates (95% CI) for FEV₁ % predicted varied from 0.6 (-3.5 to 4.6) to -9.9 (-14.2 to -5.5) between clusters. Change in FEV₁ over time did not differ significantly across clusters.

CONCLUSION

Our findings show that the long-term risk for poor asthma outcomes differed between comprehensive adult asthma phenotypes identified 20 years earlier, and suggest a strong tracking of asthma activity and impaired lung function over time.

Comparison of hypothesis- and data-driven asthma phenotypes in NHANES 2007-2012: the importance of comprehensive data availability

Background

Half of the adults with current asthma among the US National Health and Nutrition Examination Survey (NHANES) participants could be classified in more than one hypothesis-driven phenotype. A data-driven approach applied to the same subjects may allow a more useful classification compared to the hypothesis-driven one.

Aim

To compare previously defined hypothesis-driven with newly derived data-driven asthma phenotypes, identified by latent class analysis (LCA), in adults with current asthma from NHANES 2007–2012.

Methods

Adults (≥ 18 years) with current asthma from the NHANES were included (n = 1059). LCA included variables commonly used to subdivide asthma. LCA models were derived independently according to age groups: < 40 and ≥ 40 years old.

Results

Two data-driven phenotypes were identified among adults with current asthma, for both age groups. The proportions of the hypothesis-driven phenotypes were similar among the two data-driven phenotypes (p > 0.05). Class A < 40 years (n = 285; 75%) and Class A ≥ 40 years (n = 462; 73%), respectively, were characterized by a predominance of highly symptomatic asthma subjects with poor lung function, compared to Class B < 40 years (n = 94; 25%) and Class B ≥ 40 years (n = 170; 27%). Inflammatory biomarkers, smoking status, presence of obesity and hay fever did not markedly differ between the phenotypes.

Conclusion

Both data- and hypothesis-driven approaches using clinical and physiological variables commonly used to characterize asthma are suboptimal to identify asthma phenotypes among adults from the general population. Further studies based on more comprehensive disease features are required to identify asthma phenotypes in population-based studies.

Electronic supplementary material

The online version of this article (10.1186/s13601-019-0258-7) contains supplementary material, which is available to authorized users.

Evolution of IgE responses to multiple allergen components throughout childhood

BACKGROUND

There is a paucity of information about longitudinal patterns of IgE responses to allergenic proteins (components) from multiple sources.

OBJECTIVES

This study sought to investigate temporal patterns of component-specific IgE responses from infancy to adolescence, and their relationship with allergic diseases.

METHODS

In a population-based birth cohort, we measured IgE to 112 components at 6 follow-ups during childhood. We used a Bayesian method to discover cross-sectional sensitization patterns and their longitudinal trajectories, and we related these patterns to asthma and rhinitis in adolescence.

RESULTS

We identified 1 sensitization cluster at age 1, 3 at age 3, 4 at ages 5 and 8, 5 at age 11, and 6 at age 16 years. "Broad" cluster was the only cluster present at every follow-up, comprising components from multiple sources. "Dust mite" cluster formed at age 3 years and remained unchanged to adolescence. At age 3 years, a single-component "Grass" cluster emerged, which at age 5 years absorbed additional grass components and Fel d 1 to form the "Grass/cat" cluster. Two new clusters formed at age 11 years: "Cat" cluster and "PR-10/profilin" (which divided at age 16 years into "PR-10" and "Profilin"). The strongest contemporaneous associate of asthma at age 16 years was sensitization to dust mite cluster (odds ratio: 2.6; 95% CI: 1.2-6.1; P < .05), but the strongest early life predictor of subsequent asthma was sensitization to grass/cat cluster (odds ratio: 3.5; 95% CI: 1.6-7.4; P < .01).

CONCLUSIONS

We describe the architecture of the evolution of IgE responses to multiple allergen components throughout childhood, which may facilitate development of better diagnostic and prognostic biomarkers for allergic diseases.

Features of asthma which provide meaningful insights for understanding the disease heterogeneity

BACKGROUND

Data-driven methods such as hierarchical clustering (HC) and principal component analysis (PCA) have been used to identify asthma subtypes, with inconsistent results.

OBJECTIVE

To develop a framework for the discovery of stable and clinically meaningful asthma subtypes.

METHODS

We performed HC in a rich data set from 613 asthmatic children, using 45 clinical variables (Model 1), and after PCA dimensionality reduction (Model 2). Clinical experts then identified a set of asthma features/domains which informed clusters in the two analyses. In Model 3, we reclustered the data using these features to ascertain whether this improved the discovery process.

RESULTS

Cluster stability was poor in Models 1 and 2. Clinical experts highlighted four asthma features/domains which differentiated the clusters in two models: age of onset, allergic sensitization, severity, and recent exacerbations. In Model 3 (HC using these four features), cluster stability improved substantially. The cluster assignment changed, providing more clinically interpretable results. In a 5-cluster model, we labelled the clusters as: "Difficult asthma" (n = 132); "Early-onset mild atopic" (n = 210); "Early-onset mild non-atopic: (n = 153); "Late-onset" (n = 105); and "Exacerbation-prone asthma" (n = 13). Multinomial regression demonstrated that lung function was significantly diminished among children with "Difficult asthma"; blood eosinophilia was a significant feature of "Difficult," "Early-onset mild atopic," and "Late-onset asthma." Children with moderate-to-severe asthma were present in each cluster.

CONCLUSIONS AND CLINICAL RELEVANCE

An integrative approach of blending the data with clinical expert domain knowledge identified four features, which may be informative for ascertaining asthma endotypes. These findings suggest that variables which are key determinants of asthma presence, severity, or control may not be the most informative for determining asthma subtypes. Our results indicate that exacerbation-prone asthma may be a separate asthma endotype and that severe asthma is not a single entity, but an extreme end of the spectrum of several different asthma endotypes.

Functional phenotypes determined by fluctuation-based clustering of lung function measurements in healthy and asthmatic cohort participants

RATIONALE

Asthma is characterised by inflammation and reversible airway obstruction. However, these features are not always closely related. Fluctuations of daily lung function contain information on asthma phenotypes, exacerbation risk and response to long-acting β-agonists.

OBJECTIVES

In search of subgroups of asthmatic participants with specific lung functional features, we developed and validated a novel clustering approach to asthma phenotyping, which exploits the information contained within the fluctuating behaviour of twice-daily lung function measurements.

METHODS

Forced expiratory volume during the first second (FEV₁) and peak expiratory flow (PEF) were prospectively measured over 4 weeks in 696 healthy and asthmatic school children (Protection Against Allergy - Study in Rural Environments (PASTURE)/EFRAIM cohort), and over 1 year in 138 asthmatic adults with mild-to-moderate or severe asthma (Pan-European Longitudinal Assessment of Clinical Course and BIOmarkers in Severe Chronic AIRway Disease (BIOAIR) cohort). Using enrichment analysis, we explored whether the method identifies clinically meaningful, distinct clusters of participants with different lung functional fluctuation patterns.

MEASUREMENTS AND MAIN RESULTS

In the PASTURE/EFRAIM dataset, we found four distinct clusters. Two clusters were enriched in children with well-known clinical characteristics of asthma. In cluster 3, children from a farming environment predominated, whereas cluster 4 mainly consisted of healthy controls. About 79% of cluster 3 carried the asthma-risk allele rs7216389 of the 17q21 locus. In the BIOAIR dataset, we found two distinct clusters clearly discriminating between individuals with mild-to-moderate and severe asthma.

CONCLUSIONS

Our method identified dynamic functional asthma and healthy phenotypes, partly independent of atopy and inflammation but related to genetic markers on the 17q21 locus. The method can be used for disease phenotyping and possibly endotyping. It may identify participants with specific functional abnormalities, potentially needing a different therapeutic approach.

Diagnosing eosinophilic asthma using a multivariate prediction model based on blood granulocyte responsiveness

BACKGROUND

The identification of inflammatory asthma phenotypes, using sputum analysis, has proven its value in diagnosis and disease monitoring. However due to technical limitations of sputum analysis, there is a strong need for fast and noninvasive diagnostics. This study included the activation state of eosinophils and neutrophils in peripheral blood to phenotype and monitor asthma.

OBJECTIVES

To (i) construct a multivariable model using the activation state of blood granulocytes, (ii) compare its diagnostic value with sputum eosinophilia as gold standard and (iii) validate the model in an independent patient cohort.

METHODS

Clinical parameters, activation of blood granulocytes and sputum characteristics were assessed in 115 adult patients with asthma (training cohort/Utrecht) and 34 patients (validation cohort/Oxford).

RESULTS

The combination of blood eosinophil count, fractional exhaled nitric oxide, Asthma Control Questionnaire, medication use, nasal polyposis, aspirin sensitivity and neutrophil/eosinophil responsiveness upon stimulation with formyl-methionyl-leucyl phenylalanine was found to identify sputum eosinophilia with 90.5% sensitivity and 91.5% specificity in the training cohort and with 77% sensitivity and 71% specificity in the validation cohort (relatively high percentage on oral corticosteroids [OCS]).

CONCLUSIONS

The proposed prediction model identifies eosinophilic asthma without the need for sputum induction. The model forms a noninvasive and externally validated test to assess eosinophilic asthma in patients not on OCS.

Systems Biology and Clinical Practice in Respiratory Medicine. The Twain Shall Meet

Respiratory diseases are highly complex, being driven by host-environment interactions and manifested by inflammatory, structural, and functional abnormalities that vary over time. Traditional reductionist approaches have contributed vastly to our knowledge of biological systems in health and disease to date; however, they are insufficient to provide an understanding of the behavior of the system as a whole. In this Pulmonary Perspective, we discuss systems biology approaches, especially but not limited to the study of the lung as a complex system. Such integrative approaches take into account the large number of dynamic subunits and their interactions found in biological systems. Borrowing methods from physics and mathematics, it is possible to study the collective behavior of these systems over time and in a multidimensional manner. We first examine the physiological basis for complexity in the respiratory system and its implications for disease. We then expand on the potential applications of systems biology methods to study complex systems, within the context of diagnosis and monitoring of respiratory diseases including asthma, chronic obstructive pulmonary disease (COPD), and critical illness. We summarize the significant advances made in recent years using systems approaches for disease phenotyping, applied to data ranging from the molecular to clinical level, obtained from large-scale asthma and COPD networks. We describe new studies using temporal complexity patterns to characterize asthma and COPD and predict exacerbations as well as predict adverse outcomes in critical care. We highlight new methods that are emerging with this approach and discuss remaining questions that merit greater attention in the field.

Asthma phenotypes in inner-city children

BACKGROUND

Children with asthma in low-income urban areas have high morbidity. Phenotypic analysis in these children is lacking, but may identify characteristics to inform successful tailored management approaches.

OBJECTIVE

We sought to identify distinct asthma phenotypes among inner-city children receiving guidelines-based management.

METHODS

Nine inner-city asthma consortium centers enrolled 717 children aged 6 to 17 years. Data were collected at baseline and prospectively every 2 months for 1 year. Participants' asthma and rhinitis were optimally managed by study physicians on the basis of guidelines. Cluster analysis using 50 baseline and 12 longitudinal variables was performed in 616 participants completing 4 or more follow-up visits.

RESULTS

Five clusters (designated A through E) were distinguished by indicators of asthma and rhinitis severity, pulmonary physiology, allergy (sensitization and total serum IgE), and allergic inflammation. In comparison to other clusters, cluster A was distinguished by lower allergy/inflammation, minimally symptomatic asthma and rhinitis, and normal pulmonary physiology. Cluster B had highly symptomatic asthma despite high step-level treatment, lower allergy and inflammation, and mildly altered pulmonary physiology. Cluster C had minimally symptomatic asthma and rhinitis, intermediate allergy and inflammation, and mildly impaired pulmonary physiology. Clusters D and E exhibited progressively higher asthma and rhinitis symptoms and allergy/inflammation. Cluster E had the most symptomatic asthma while receiving high step-level treatment and had the highest total serum IgE level (median, 733 kU/L), blood eosinophil count (median, 400 cells/mm³), and allergen sensitizations (15 of 22 tested).

CONCLUSIONS

Allergy distinguishes asthma phenotypes in urban children. Severe asthma often coclusters with highly allergic children. However, a symptomatic phenotype with little allergy or allergic inflammation was identified.

The soft computing-based approach to investigate allergic diseases: a systematic review

BACKGROUND

Early recognition of inflammatory markers and their relation to asthma, adverse drug reactions, allergic rhinitis, atopic dermatitis and other allergic diseases is an important goal in allergy. The vast majority of studies in the literature are based on classic statistical methods; however, developments in computational techniques such as soft computing-based approaches hold new promise in this field.

OBJECTIVE

The aim of this manuscript is to systematically review the main soft computing-based techniques such as artificial neural networks, support vector machines, bayesian networks and fuzzy logic to investigate their performances in the field of allergic diseases.

METHODS

The review was conducted following PRISMA guidelines and the protocol was registered within PROSPERO database (CRD42016038894). The research was performed on PubMed and ScienceDirect, covering the period starting from September 1, 1990 through April 19, 2016.

RESULTS

The review included 27 studies related to allergic diseases and soft computing performances. We observed promising results with an overall accuracy of 86.5%, mainly focused on asthmatic disease. The review reveals that soft computing-based approaches are suitable for big data analysis and can be very powerful, especially when dealing with uncertainty and poorly characterized parameters. Furthermore, they can provide valuable support in case of lack of data and entangled cause-effect relationships, which make it difficult to assess the evolution of disease.

CONCLUSIONS

Although most works deal with asthma, we believe the soft computing approach could be a real breakthrough and foster new insights into other allergic diseases as well.

Disaggregating asthma: Big investigation versus big data

We are facing a major challenge in bridging the gap between identifying subtypes of asthma to understand causal mechanisms and translating this knowledge into personalized prevention and management strategies. In recent years, "big data" has been sold as a panacea for generating hypotheses and driving new frontiers of health care; the idea that the data must and will speak for themselves is fast becoming a new dogma. One of the dangers of ready accessibility of health care data and computational tools for data analysis is that the process of data mining can become uncoupled from the scientific process of clinical interpretation, understanding the provenance of the data, and external validation. Although advances in computational methods can be valuable for using unexpected structure in data to generate hypotheses, there remains a need for testing hypotheses and interpreting results with scientific rigor. We argue for combining data- and hypothesis-driven methods in a careful synergy, and the importance of carefully characterized birth and patient cohorts with genetic, phenotypic, biological, and molecular data in this process cannot be overemphasized. The main challenge on the road ahead is to harness bigger health care data in ways that produce meaningful clinical interpretation and to translate this into better diagnoses and properly personalized prevention and treatment plans. There is a pressing need for cross-disciplinary research with an integrative approach to data science, whereby basic scientists, clinicians, data analysts, and epidemiologists work together to understand the heterogeneity of asthma.

Subtypes of asthma based on asthma control and severity: a latent class analysis

BACKGROUND

Asthma subtyping is a complex new field of study. Usually both etiological and outcome factors of asthma have been used simultaneously for subtyping thus making the interpretation of the results difficult. Identification of subtypes of asthma based on questionnaire data only will be useful for both treatment of asthma and for research. Our objective was to identify asthma subtypes that capture both asthma control and severity based on easily accessible variables.

METHODS

We applied latent class analysis for the 1995 adult asthmatics, 692 men and 1303 women, of the Northern Finnish Asthma Study (NoFAS). The classifying variables included use of asthma medication within the last 12 months, St. George's Respiratory Questionnaire score, and asthma-related healthcare use within the last 12 months. Covariates adjusted for included COPD, allergic rhinitis/allergic eczema, BMI, age and sex. All information was based on self-administered questionnaires.

RESULTS

We identified four subtypes for women: Controlled, mild asthma (41% of participants); Partly controlled, moderate asthma (24%); Uncontrolled asthma, unknown severity (26%), and Uncontrolled, severe asthma (9%). For men we identified three subtypes: Controlled, mild asthma (31%); Poorly controlled asthma, unknown severity (53%); and Partly controlled, severe asthma (17%). For almost 96% of the subjects this subtyping was accurate. The covariates fitted in the model were based on clinical judgment and were good predictors of class membership.

CONCLUSIONS

Our results show that it is possible to form meaningful and accurate asthma subtypes based on questionnaire data, and that separate classification should be applied for men and women.

Relationship between cytokine expression patterns and clinical outcomes: two population-based birth cohorts

BACKGROUND

Models that incorporate patterns of multiple cytokine responses to allergens, rather than individual cytokine production, may better predict sensitization and asthma.

OBJECTIVE

To characterize the patterns of peripheral blood mononuclear cells' (PBMCs) cytokine responses to house dust mite (HDM) allergens among children from two population-based birth cohorts using machine learning techniques.

METHODS

PBMCs collected at 8 years of age from the UK Manchester Asthma and Allergy Study (n = 268) and at 14 years of age from the Australian Raine Study (n = 1374) were cultured with HDM extract (10 μg/ml). Cytokine expression (IL-13, IL-5, IFN-γ, and IL10) was measured in the supernatant. Cytokine patterns were identified using a Gaussian mixture model clustering, and classification stability was assessed by bootstrapping.

RESULTS

A six-class model indicated complex latent structure of cytokine expression. Based on the characteristics of each class, we designated them as follows: 'Nonresponders' (n = 905, 55%); 'IL-10 responders' (n = 49, 3%); 'IFN-γ and IL-13 medium responders' (n = 56, 3.4%); 'IL-13 medium responders' (n = 351, 21.4%); 'IL-5 and IL-13 medium responders' (n = 77, 4.7%); and 'IL-13 and IL-5 high responders' (n = 204, 12.4%). 'IL-13 and IL-5 high responders' were at much higher risk of HDM sensitization and asthma compared to all other classes, with 88% of children assigned to this class being sensitized and 28.5% having asthma.

CONCLUSION

Using model-based clustering, we identified several distinct patterns of cytokine response to HDM and observed interplay between cytokine expression level, cytokine patterns (especially IL-13 and IL-5), and clinical outcomes. 'IL-13 and IL-5 high responders' class was strongly associated with HDM sensitization. However, among HDM-sensitized children, one-third showed no PBMC response to HDM, and the majority of HDM-sensitized children did not have asthma or wheeze. Our findings suggest that positive HDM 'allergy tests' and asthma are associated with a broad range of immunophenotypes, which may have important implications for the use of cytokine-targeted treatment approaches.

Identification of Asthma Subtypes Using Clustering Methodologies

Asthma is a heterogeneous disease comprising a number of subtypes which may be caused by different pathophysiologic mechanisms (sometimes referred to as endotypes) but may share similar observed characteristics (phenotypes). The use of unsupervised clustering in adult and paediatric populations has identified subtypes of asthma based on observable characteristics such as symptoms, lung function, atopy, eosinophilia, obesity, and age of onset. Here we describe different clustering methods and demonstrate their contributions to our understanding of the spectrum of asthma syndrome. Precise identification of asthma subtypes and their pathophysiological mechanisms may lead to stratification of patients, thus enabling more precise therapeutic and prevention approaches.

Phenotyping of difficult asthma using longitudinal physiological and biomarker measurements reveals significant differences in stability between clusters

Background

Although the heterogeneous nature of asthma has prompted asthma phenotyping with physiological or biomarker data, these studies have been mostly cross-sectional. Longitudinal studies that assess the stability of phenotypes based on a combination of physiological, clinical and biomarker data are currently lacking.

Our objective was to assess the longitudinal stability of clusters derived from repeated measures of airway and physiological data over a 1-year period in moderate and severe asthmatics.

Methods

A total of 125 subjects, 48 with moderate asthma (MA) and 77 with severe asthma (SA) were evaluated every 3 months and monthly, respectively, over a 1-year period. At each 3-month time point, subjects were grouped into 4 asthma clusters (A, B, C, D) based on a combination of clinical (duration of asthma), physiological (FEV₁ and BMI) and biomarker (sputum eosinophil count) variables, using k-means clustering.

Results

Majority of subjects in clusters A and C had severe asthma (93 % of subjects in cluster A and 79.5 % of subjects in cluster C at baseline). Overall, a total of 59 subjects (47 %) had stable cluster membership, remaining in clusters with the same subjects at each evaluation time. Cluster A was the least stable (21 % stability) and cluster B was the most stable cluster (71 % stability). Cluster stability was not influenced by changes in the dosage of inhaled corticosteroids.

Conclusion

Asthma phenotyping based on clinical, physiologic and biomarker data identified clusters with significant differences in longitudinal stability over a 1-year period. This finding indicates that the majority of patients within stable clusters can be phenotyped with reasonable accuracy after a single measurement of lung function and sputum eosinophilia, while patients in unstable clusters will require more frequent evaluation of these variables to be properly characterized.

Atopic Dermatitis and Respiratory Allergy: What is the Link

Understanding the aetiology and progression of atopic dermatitis and respiratory allergy may elucidate early preventative and management strategies aimed towards reducing the global burden of asthma and allergic disease. In this article, we review the current opinion concerning the link between atopic dermatitis and the subsequent progression of respiratory allergies during childhood and into early adolescence. Advances in machine learning and statistical methodology have facilitated the discovery of more refined definitions of phenotypes for identifying biomarkers. Understanding the role of atopic dermatitis in the development of respiratory allergy may ultimately allow us to determine more effective treatment strategies, thus reducing the patient and economic burden associated with these conditions.

To what extent is allergen exposure a risk factor for the development of allergic disease?

The role of allergen exposure in the development of allergic disease has been a matter of considerable debate, and our understanding of the importance of allergens has evolved over the last 25 years. Several observational and primary prevention studies have investigated these relationships, and different studies reported inconsistent, and sometimes opposite findings. It has to be emphasized that a clear understanding of how aeroallergen exposure occurs, and accurate and reproducible measurement of exposure are essential prerequisites for understanding the role of exposure. However, our current understanding of how we get exposed to allergens is not based on solid evidence, but on a number of assumptions, and we urgently need to develop better proxy measures (or indices) of exposure. In addition, the relative importance of the timing of exposure (e.g. early compared to exposure in later life) is unknown. It is also unclear which route of exposure is the most relevant (e.g. inhaled vs. oral vs. transcutaneous). Available data suggest that the dose-response relationship between allergen exposure and allergic disease may differ between different allergens, dose ranges and exposure patterns, and these relationships may further differ between different populations and geographical areas. It is increasingly clear that childhood asthma and atopy are not single phenotypes, and it is likely that allergen exposure has different effect on distinct subgroups under the umbrella terms of 'sensitization' and 'asthma'. Susceptibility to allergen exposure, other environmental exposures and their interactions may also differ between individuals with different genetic predispositions. However, the precise nature of these complex relationships is unclear. We need a holistic approach offered by systems biology, with integration of information on the standardized and reliable measures of exposures (including allergens and other relevant exposures) with genetic and biological data to fully understand the role of allergens in the development of allergic disease.

Defining phenotypes in asthma: a step towards personalized medicine

Asthma is a common disease with a complex pathophysiology. It can present in various clinical forms and with different levels of severity. Unbiased cluster analytic methods have unravelled several phenotypes in cohorts representative of the whole spectrum of severity. Clusters of severe asthma include those on high-dose corticosteroid treatment, often with both inhaled and oral treatment, usually associated with severe airflow obstruction. Phenotypes with concordance between symptoms and sputum eosinophilia have been reported, including an eosinophilic inflammation-predominant group with few symptoms and late-onset disease who have a high prevalence of rhinosinusitis, aspirin sensitivity, and exacerbations. Sputum eosinophilia is also a biomarker that can predict therapeutic responses to antibody-based treatments to block the effects of the T-helper (Th)-2 cytokine, interleukin (IL)-5. Low Th2-expression has been predictive of poor therapeutic response to inhaled corticosteroid therapy. Current asthma schedules emphasise a step-up approach to treating asthma in relation to increasing severity, but, in more severe disease, phenotyping or endotyping of asthma will be necessary to determine new treatment strategies as severe asthma is recognized as being a particularly heterogeneous disease. Much less is known about 'non-eosinophilic' asthma. Phenotypic characterisation of corticosteroid insensitivity and chronic airflow obstruction of severe asthma is also needed. Phenotype-driven treatment of asthma will be further boosted by the advent of transcriptomic and proteomic technologies, with the application of systems biology or medicine approaches to defining phenotypes and biomarkers of disease and therapeutic response. This will pave the way towards personalized medicine and healthcare for asthma.

Treatment responsiveness of phenotypes of symptomatic airways obstruction in adults

BACKGROUND

Asthma and chronic obstructive pulmonary disease (COPD) are heterogeneous disorders encompassing different phenotypes of airflow obstruction, which might differ in their response to treatment.

OBJECTIVE

The aim of this study was to determine distinct phenotypes comprising the syndromes of asthma and COPD and the treatment responsiveness of these phenotypes to inhaled β-agonist, antimuscarinic, and corticosteroid therapy.

METHODS

We undertook a cross-sectional study with 3 phases. In phase 1, 1,264 participants aged 18 to 75 years with self-reported current wheeze and breathlessness were identified from a random population sample of 16,459. In phase 2, 451 participants attended for detailed assessment, including responsiveness to inhaled salbutamol and ipratropium bromide. In phase 3, 168 steroid-naive participants were enrolled in a 12-week trial of inhaled budesonide. Cluster analysis was performed in 389 participants who completed phase 2 with full data. Treatment responsiveness was compared between phenotypes.

RESULTS

Cluster analysis identified 5 phenotypes: moderate-to-severe childhood-onset atopic asthma, asthma-COPD overlap, obese-comorbid, mild childhood-onset atopic asthma, and mild intermittent. Bronchodilation after salbutamol was equal to or greater than that after ipratropium for all phenotypes. The moderate-to-severe childhood-onset atopic asthma, asthma-COPD overlap, and obese-comorbid phenotypes had greater efficacy with inhaled corticosteroid treatment than the mild intermittent group.

CONCLUSION

Cluster analysis of adults with symptomatic airflow obstruction identifies 5 disease phenotypes, including asthma-COPD overlap and obese-comorbid phenotypes, and provides evidence that patients with the asthma-COPD overlap syndrome might benefit from inhaled corticosteroid therapy.

The role of the early-life environment in the development of allergic disease

A consensus has been reached that the development of allergic disorders is strongly influenced by early life exposures. An overview of several prenatal and early life factors that have been investigated for their associations with development of childhood allergy is presented. Delivery mode, the gut microbiome, vitamin D, folate, breastfeeding, pets, antibiotics, environmental tobacco smoke, and airborne traffic pollutants are discussed. Although many studies suggest an effect, overall, no risk factors clearly increase or reduce the risk of allergic outcomes.

Developmental profiles of eczema, wheeze, and rhinitis: two population-based birth cohort studies

BACKGROUND

The term "atopic march" has been used to imply a natural progression of a cascade of symptoms from eczema to asthma and rhinitis through childhood. We hypothesize that this expression does not adequately describe the natural history of eczema, wheeze, and rhinitis during childhood. We propose that this paradigm arose from cross-sectional analyses of longitudinal studies, and may reflect a population pattern that may not predominate at the individual level.

METHODS AND FINDINGS

Data from 9,801 children in two population-based birth cohorts were used to determine individual profiles of eczema, wheeze, and rhinitis and whether the manifestations of these symptoms followed an atopic march pattern. Children were assessed at ages 1, 3, 5, 8, and 11 y. We used Bayesian machine learning methods to identify distinct latent classes based on individual profiles of eczema, wheeze, and rhinitis. This approach allowed us to identify groups of children with similar patterns of eczema, wheeze, and rhinitis over time. Using a latent disease profile model, the data were best described by eight latent classes: no disease (51.3%), atopic march (3.1%), persistent eczema and wheeze (2.7%), persistent eczema with later-onset rhinitis (4.7%), persistent wheeze with later-onset rhinitis (5.7%), transient wheeze (7.7%), eczema only (15.3%), and rhinitis only (9.6%). When latent variable modelling was carried out separately for the two cohorts, similar results were obtained. Highly concordant patterns of sensitisation were associated with different profiles of eczema, rhinitis, and wheeze. The main limitation of this study was the difference in wording of the questions used to ascertain the presence of eczema, wheeze, and rhinitis in the two cohorts.

CONCLUSIONS

The developmental profiles of eczema, wheeze, and rhinitis are heterogeneous; only a small proportion of children (∼ 7% of those with symptoms) follow trajectory profiles resembling the atopic march. Please see later in the article for the Editors' Summary.

Identification of clinical phenotypes using cluster analyses in COPD patients with multiple comorbidities

Chronic obstructive pulmonary disease (COPD) is characterized by persistent airflow limitation, the severity of which is assessed using forced expiratory volume in 1 sec (FEV1, % predicted). Cohort studies have confirmed that COPD patients with similar levels of airflow limitation showed marked heterogeneity in clinical manifestations and outcomes. Chronic coexisting diseases, also called comorbidities, are highly prevalent in COPD patients and likely contribute to this heterogeneity. In recent years, investigators have used innovative statistical methods (e.g., cluster analyses) to examine the hypothesis that subgroups of COPD patients sharing clinically relevant characteristics (phenotypes) can be identified. The objectives of the present paper are to review recent studies that have used cluster analyses for defining phenotypes in observational cohorts of COPD patients. Strengths and weaknesses of these statistical approaches are briefly described. Description of the phenotypes that were reasonably reproducible across studies and received prospective validation in at least one study is provided, with a special focus on differences in age and comorbidities (including cardiovascular diseases). Finally, gaps in current knowledge are described, leading to proposals for future studies.