Asthma and lower airway disease
Prognostic value of cluster analysis of severe asthma phenotypes

https://doi.org/10.1016/j.jaci.2014.04.038Get rights and content

Background

Cross-sectional severe asthma cluster analysis identified different phenotypes. We tested the hypothesis that these clusters will follow different courses.

Objective

We aimed to identify which asthma outcomes are specific and coherently associated with these different phenotypes in a prospective longitudinal cohort.

Methods

In a longitudinal cohort of 112 patients with severe asthma, the 5 Severe Asthma Research Program (SARP) clusters were identified by means of algorithm application. Because patients of the present cohort all had severe asthma compared with the SARP cohort, homemade clusters were identified and also tested. At the subsequent visit, we investigated several outcomes related to asthma control at 1 year (6-item Asthma Control Questionnaire [ACQ-6], lung function, and medication requirement) and then recorded the 3-year exacerbations rate and time to first exacerbation.

Results

The SARP algorithm discriminated the 5 clusters at entry for age, asthma duration, lung function, blood eosinophil measurement, ACQ-6 scores, and diabetes comorbidity. Four homemade clusters were mostly segregated by best ever achieved FEV1 values and discriminated the groups by a few clinical characteristics. Nonetheless, all these clusters shared similar asthma outcomes related to asthma control as follows. The ACQ-6 score did not change in any cluster. Exacerbation rate and time to first exacerbation were similar, as were treatment requirements.

Conclusion

Severe asthma phenotypes identified by using a previously reported cluster analysis or newly homemade clusters do not behave differently concerning asthma control–related outcomes, which are used to assess the response to innovative therapies. This study demonstrates a potential limitation of the cluster analysis approach in the field of severe asthma.

Section snippets

Patients

A cohort of patients with severe asthma was built during a 1-year recruitment period after ethics committee approval in the outpatient clinic of Arnaud de Villeneuve's Hospital, Montpellier, France. This cohort aimed to identify relevant outcomes in patients with severe asthma based on thorough initial investigations. Thus patients with a secured diagnosis of severe asthma according to Innovative Medicine Initiative consensus were solicited to participate, and data were recorded once patients

Results

Patients' characteristics are presented in Table I. The cohort consisted of 112 patients with severe asthma with a mean age of 56 years, most of whom (54.4%) were women. Sixty-one percent had received an annual dose of steroids greater than 2 g the year before inclusion. The mean exacerbation rate was 2.33 per patient at inclusion, ranging from 0 to 7.

Table II presents the distribution according to the algorithm tree proposed in the SARP study reproduced in our own cohort, which allows

Discussion

In the present study we could not discriminate future risks in the different phenotypes of severe asthma in a real-life cohort of patients with severe asthma using longitudinal follow-up. Specifically, exacerbations, asthma symptoms, treatment load, and the level of airway obstruction did not follow different trends among the identified groups.

Cluster analyses of severe asthma are used to better understand the heterogeneity of the disease and the constant failure to adequately control symptoms

References (23)

  • W.C. Moore et al.

    Identification of asthma phenotypes using cluster analysis in the Severe Asthma Research Program

    Am J Respir Crit Care Med

    (2010)
  • Cited by (62)

    • Definition and assessment of severe asthma

      2020, Revue des Maladies Respiratoires Actualites
    • Identification and prospective stability of electronic nose (eNose)–derived inflammatory phenotypes in patients with severe asthma

      2019, Journal of Allergy and Clinical Immunology
      Citation Excerpt :

      The second novelty of this study is the assessment of stability of patients in those clusters during longitudinal follow-up. Previously, clustering techniques have been used to describe and monitor phenotypes of asthma based on clinical characteristics12,48-51 and for eNose technology focused on phenotyping broad cohorts of obstructive pulmonary diseases.27,52 However, the present study merged both approaches by using unbiased molecular profiling as a starting point for the follow-up of severe asthma phenotypes.

    View all citing articles on Scopus

    Disclosure of potential conflict of interest: A. Bourdin has board memberships with Boehringer Ingelheim and GlaxoSmithKline. F. Paganin is a board member for Novartis. P. Chanez has board memberships with GlaxoSmithKline, AstraZeneca, Chiesi, and Boston Scientific; has consultant arrangements with Novartis; has received research support from Jansen; has received payment for lectures from AstraZeneca, GlaxoSmithKline, Boston Scientific; and has received travel support from Chiesi, Novartis, and AstraZeneca. The rest of the authors declare that they have no relevant conflicts of interest.

    View full text