Identifying adult asthma phenotypes using a clustering approach

Siroux, Valérie; Basagaña, Xavier; Boudier, Anne; Pin, Isabelle; Garcia‐Aymerich, Judith; Vesin, Aurélien; Slama, Rémy; Jarvis, Déborah; Antó, Josep M.; Kauffmann, F.; Sunyer, Jordi

doi:10.1183/09031936.00120810

Cited by 235 publications

(178 citation statements)

References 30 publications

Supporting

Mentioning

155

Contrasting

Unclassified

Order By: Relevance

“…This diversification was highlighted in previous studies, both in adults (6)(7)(8) and children (9,10). However, the stability of the components was weak, and less robust to changes in the model assumptions (like rotations) or variable discretization policies.…”

Section: Comparison With Previous Work and Interpretationmentioning

confidence: 69%

“…A data-driven approach with unsupervised statistical learning techniques can be used for discovery of latent asthma phenotypes, which can be derived based on a series of observable disease manifestations, instead of using predetermined classifications proposed by committees of experts. Several previous studies applied principal components analysis, exploratory factor analysis (EFA), partitioning clustering, hierarchical clustering (HC), and other techniques to identify latent groups and associated symptom patterns among adults (6)(7)(8) and children (9-12) with asthma. The results have been inconsistent.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Challenges in Identifying Asthma Subgroups Using Unsupervised Statistical Learning Techniques

Prosperi

Şahiner

Belgrave

et al. 2013

Am J Respir Crit Care Med

View full text Add to dashboard Cite

Rationale: Unsupervised statistical learning techniques, such as exploratory factor analysis (EFA) and hierarchical clustering (HC), have been used to identify asthma phenotypes, with partly consistent results. Some of the inconsistency is caused by the variable selection and demographic and clinical differences among study populations. Objectives: To investigate the effects of the choice of statistical method and different preparations of data on the clustering results; and to relate these to disease severity. Methods: Several variants of EFA and HC were applied and compared using various sets of variables and different encodings and transformations within a dataset of 383 children with asthma. Variables included lung function, inflammatory and allergy markers, family history, environmental exposures, and medications. Clusters and original variables were related to asthma severity (logistic regression and Bayesian network analysis). Measurements and Main Results: EFA identified five components (eigenvalues > 1) explaining 35% of the overall variance. Variations of the HC (as linkage-distance functions) did not affect the cluster inference; however, using different variable encodings and transformations did. The derived clusters predicted asthma severity less than the original variables. Prognostic factors of severity were medication usage, current symptoms, lung function, paternal asthma, body mass index, and age of asthma onset. Bayesian networks indicated conditional dependence among variables. Conclusions: The use of different unsupervised statistical learning methods and different variable sets and encodings can lead to multiple and inconsistent subgroupings of asthma, not necessarily correlated with severity. The search for asthma phenotypes needs more careful selection of markers, consistent across different study populations, and more cautious interpretation of results from unsupervised learning.

show abstract

Section: Comparison With Previous Work and Interpretationmentioning

confidence: 69%

mentioning

confidence: 99%

Challenges in Identifying Asthma Subgroups Using Unsupervised Statistical Learning Techniques

Prosperi

Şahiner

Belgrave

et al. 2013

Am J Respir Crit Care Med

View full text Add to dashboard Cite

show abstract

“…Asthma is a common complex respiratory disorder with various overlapping phenotypes [1,2]. Common features include fluctuating respiratory symptoms associated with variable airflow limitation and bronchial hyperresponsiveness (BHR) due to inflammation of the airways.…”

Section: Introductionmentioning

confidence: 99%

Determinants of persistent asthma in young adults

Traulsen

Halling

Bælum

et al. 2018

European Clinical Respiratory Journal

View full text Add to dashboard Cite

Objective: The aim of the study was to evaluate determinants for the prognosis of asthma in a population-based cohort of young adults.Design: The study was a nine-year clinical follow up of 239 asthmatic subjects from an enriched population-based sample of 1,191 young adults, aged 20–44 years, who participated in an interviewer-administered questionnaire and clinical examination at baseline in 2003–2006. From the interview, an asthma score was generated as the simple sum of affirmative answers to five main asthma-like symptoms in order to analyse symptoms of asthma as a continuum. The clinical examination comprised spirometry, bronchial challenge or bronchodilation, and skin prick test.Results: Among the 239 individuals with asthma at baseline 164 (69%) had persistent asthma at follow up, while 68 (28%) achieved remission of asthma and seven (3%) were diagnosed with COPD solely. Determinants for persistent asthma were use of medication for breathing within the last 12 months: Short-acting beta-adrenoceptor agonists (SABA) only (OR 3.39; 95%CI: 1.47–7.82) and inhaled corticosteroids (ICS) and/or long-acting beta-adrenoceptor agonists (LABA) (8.95; 3.87–20.69). Stratified by age of onset determinants for persistence in individuals with early-onset asthma (age less than 16 years) were FEV₁ below predicted (7.12; 1.61–31.50), asthma score at baseline (2.06; 1.15–3.68) and use of ICS and/or LABA within 12 months (9.87; 1.95–49.98). In individuals with late-onset asthma the determinant was use of ICS and/or LABA within 12 months (6.84; 2.09–22.37).Conclusions: Pulmonary function below predicted, severity of disease expressed by asthma score and use of ICS and/or LABA were all determinants for persistent early-onset asthma, whereas only use of ICS and/or LABA was a determinant in late-onset asthma. A high asthma score indicated insufficient disease control in a substantial proportion of these young adults.

show abstract

“…Recently, novel statistical and mathematical methods such as cluster or factor analysis and principal component techniques have been used to phenotype asthma, suggesting novel phenotypes in both adults and children (14,15). Machinelearning approaches have also been applied to cluster phenotypes in childhood asthma (16).…”

mentioning

confidence: 99%

Untangling asthma phenotypes and endotypes

Agache

Akdiş

Jutel

et al. 2012

Allergy

292

235

View full text Add to dashboard Cite

Asthma phenotypes have been developed to address the complexities of the disease. However, owing to a lack of longitudinal studies, little is known about the onset as well as the stability of phenotypes. Distinguishing phenotypes with regard to the severity or duration of the disease is essential. A phenotype covers the clinically relevant properties of the disease, but does not show the direct relationship to disease etiology and pathophysiology. Different pathogenetic mechanisms might cause similar asthma symptoms and might be operant in a certain phenotype. These putative mechanisms are addressed by the term 'endotype'. Classification of asthma based on endotypes provides advantages for epidemiological, genetic, and drug-related studies. A successful definition of endotypes should link key pathogenic mechanisms with the asthma phenotype. Thus, the identification of corresponding molecular biomarkers for individual pathogenic mechanism underlying phenotypes or subgroups within a phenotype is important. Whether newly defined asthma endotypes predict the individual course of asthma has to be validated in longitudinal studies. The accurate endotyping reflects natural history of asthma and should help to predict treatment response. Thus, understanding asthma endotypes might be useful in clinical practice.

show abstract

Identifying adult asthma phenotypes using a clustering approach

Cited by 235 publications

References 30 publications

Challenges in Identifying Asthma Subgroups Using Unsupervised Statistical Learning Techniques

Challenges in Identifying Asthma Subgroups Using Unsupervised Statistical Learning Techniques

Determinants of persistent asthma in young adults

Untangling asthma phenotypes and endotypes

Contact Info

Product

Resources

About