Asthma is a chronic heterogenous airway disease characterized by excessive airway calibre fluctuation, which usually develops with local inflammation.1 Despite multiple classificatory phenotypes, recent guidelines have strengthened the utility of determining the inflammatory phenotype, especially in severe asthma, in order to establish the best therapeutic approach.1,2 The coexistence of asthma and comorbidities is also very frequent. In fact, the widespread treatable traits approach along with local and international consensus documents emphasize the need to identify coexisting comorbidities with the aim of treating them to achieve adequate asthma control.3

Interestingly, exacerbations are one the most relevant predictors of poor prognosis in asthma.4 Sex, age, race, lung function, smoking habit and exposure to respiratory viruses have been related to a higher prevalence of asthma exacerbations, although the causality is not clear.5–9 In this line, many authors have reported the predominance of the type 2 (T2) inflammatory phenotype in exacerbation-prone asthma (EPA)7,10,11 In fact, large population studies have reported blood eosinophils and exhaled nitric oxide levels (FeNO) as risk factors for the prevalence of exacerbation.12,13 However, the risk of having frequent exacerbations is less well documented in cohort studies of patients with severe asthma on existing therapy. Moreover, other aspects including frequent comorbidities, such as bronchiectasis, atopy status or an occupational origin of asthma could have an impact on the development of exacerbations, beyond the presence or absence of the T2 endotype.

Thus, the objective of the present study was to compare the clinical and inflammatory characteristics of patients with moderate or severe asthma with a history of two or more exacerbations in the previous year vs. those who present just one or no exacerbation.

Three hundred ten patients were visited in the Severe Asthma Unit in 2018. In 214 (69%) patients the sputum differential cell count was not obtained due to impossibility to perform SI in the context of FEV1 values <1.5L, or unsuccessful SI manoeuvre or patient refusal to undergo SI. In 96 (31%) patients the total IgE, atopy status, blood eosinophils and sputum differential cell count were obtained. There were no statistically significative differences in terms of basal characteristics between excluded patients and the 96 patients corresponding to the included population. In this latter group 46 patient (47%) presented EPA while 50 (53%) patients presented 1 or no exacerbations (Fig. 1).

Fig. 1.

Flow chart of the study population. Abbreviations: IE: induced sputum; FEV1: forced expiratory volume in 1s.

(0.14MB).

There were no differences between patients with EPA vs. non-exacerbators regarding the prevalence of the T2 inflammatory phenotype (n=26 (56.5%) vs. n=30 (60%) respectively 19.6; p=0.730) (Fig. 2). Neither were there differences related to IgE, blood eosinophil counts or blood neutrophil counts. The prevalence of atopy was lower in the EPA group compared to non-exacerbators.

Fig. 2.

Characteristics of the inflammatory phenotype in the study population. Abbreviation: EPA: exacerbation-prone asthma.

(0.12MB).

There were no differences in terms of FeNO or airway eosinophil count in SI in patients with EPA vs. non-exacerbators. No differences were found between the two groups in terms of the sputum neutrophil counts (Table 2 and Fig. 3).

Fig. 3.

Eosinophil and neutrophil levels according to the presence of asthma exacerbations (≥2 in the previous year), in blood (a and b) and in sputum (c and d). Abbreviation: EPA: exacerbation-prone asthma.

(0.18MB).

There were no significant differences in age, sex, age at diagnosis, smoking habit, or body mass index between the two groups.

Bronchiectasis was more common in the EPA group, although there were no differences in the Reef grade between groups.

EPA patients had lower asthma control test (ACT) values and were more frequently treated with long-acting muscarinic antagonists (LAMA), azithromycin, nebulized colistin, and oral corticosteroids compared to patients presenting one or no moderate or severe exacerbation (Table 1).

The multivariate logistic regression model revealed that ACT values less than 20 and the use of systemic steroids were independently associated with EPA (Table 3).

The present study describes the clinical, inflammatory, lung function and the presence of bronchiectasis in the HRCT of moderate and severe asthma patients in a well-phenotyped ambispective cohort. This study revealed that airway eosinophilic inflammation did not differ between EPA and non-exacerbators in patients with moderate or severe asthma and ongoing therapy, and that it was not a risk factor or a predictor for EPA in our cohort. Neither was the blood eosinophil count a risk factor for presenting asthma exacerbations. Our study also shows that an ACT score less than 20 and treatment with systemic steroids were independent predictors of EPA.

Some authors have linked the presence of sputum eosinophilia and elevated FeNO with the risk of asthma exacerbations.7 In this line, Domínguez-Ortega et al.24 reported eosinophilia and a high-T2 inflammatory pattern being significantly associated with AE in the severe asthma Spanish MEGA cohort, encompassing 486 patients. Parallel, Peters et al.25 described a 406 adult severe asthma cohort with a 3-year follow-up, where 21% presented at least one exacerbation each year. In this cohort, the risk of AE was related to increasing values of eosinophils. However, several large cohorts of patients with severe asthma have not been able to confirm these findings.11 This discrepancy between studies probably reflects the impact of the heterogeneity of asthma, the burden of comorbidities and the bias of patients being or not under controller therapy. In this sense, the response of inflammatory biomarkers to ongoing treatment could differ depending on the nature of each specific treatment. Bateman et al.26 performed an analysis of three clinical trials including more than 7000 patients with asthma, receiving treatment with long-acting beta-adrenergic agonists (LABA), inhaled corticosteroids (ICS) and LAMA and unlike the inflammatory phenotype, only disease severity, symptom control and history of smoking were predictors of exacerbations. Along this line, Denlinger et al.11 were not able to relate the presence of sputum eosinophils or FeNO with the risk of exacerbations in the 709 patient Severe Asthma Research Program-3 (SARP-3) cohort. Again, these patients were already being treated with controller therapy and not surprisingly, in the SARP-3 cohort, the EPA group had the highest medication burden similar to what was found in the present study.

Our study found no differences in blood eosinophil counts between EPA and patients with asthma and one or no exacerbation. Similar to what occurs with the sputum eosinophil count, many authors have related a higher blood eosinophil count has been related to EPA.12,13,27 These finding are coherent with blood eosinophils being an acceptable predictor of airway eosinophils.28 However, EPA patients probably have other features that are independent of T2 inflammation biomarkers. In this line, our study likely reflects aspects such as current ongoing treatment or the presence of comorbidities such as bronchiectasis or atopy.

Sixty-one (64%) patients in our cohort presented bronchiectasis in the HRCT, most being mild (Reiff score 1–6). Bronchiectasis is a relatively common comorbidity in asthma patients, having a much higher prevalence in the severe forms of the disease.29 It is likely that the existence of uncontrolled bronchial inflammation produces an increase in mucus production and local mucosa damage leading to structural alterations.30–32 Despite the frequent coexistence of bronchiectasis and asthma, there are still many gaps in the knowledge related to the interaction with asthma in terms of aetiology, clinical features, management and even exacerbations. Although bronchiectasis is a usual phenomenon leading to frequent exacerbations in these patients, little is known about the impact of the inflammatory phenotype on asthma exacerbations in patients with asthma-bronchiectasis overlap. In this context, the presence of bronchiectasis could have a direct impact on the inflammatory phenotype of these patients, distorting the classical belief of eosinophils in sputum induction being a predictor of exacerbations. In fact, García-Clemente et al.33 recently described lower levels of FeNO in patients with severe asthma and bronchiectasis compared to those with severe asthma and a normal HRCT. Parallelly, Mao et al.34 recently reported that age and FEV1, and not the T2 inflammatory phenotype, were factors associated with exacerbations in patients with asthma and bronchiectasis, and that the presence Pseudomonas aeruginosa was an independent factor associated with exacerbations in this population.

Atopy status is another factor that should be taken into account to interpret the results. In our study, the prevalence of atopy was lower in the EPA group compared to non-exacerbators. In this line, in the SARP-3 cohort an inverse relationship between exacerbations and IgE levels or degree of allergen sensitization was described.11 These findings may be related to a better response to ICS in patients presenting this phenotype, although other authors have described a lack of association between IgE or atopy status and asthma control.35 Again, the underlaying reality is probably much more complex since it is necessary to consider the interaction of atopy and allergy with other factors such as viruses, considered to be the trigger of more than the 50% of asthma exacerbations, or pollutants.36

Our study also showed that an ACT value less than 20 and systemic corticosteroid use were related to EPA. These results are in line with previous studies reporting similar results.7 In this context, it seems that the need for intensive anti-inflammatory therapy is related to prior exacerbations, worse asthma control and a higher risk of future exacerbations, since these are all characteristics linked to more severe asthma.26

This study has several potential limitations. First, the triggers of the exacerbations were retrospectively and not systematically collected, which could have contributed to misinterpreting the inflammatory status of the patients included. Second, sputum induction was obtained in only 30% of the patients of our cohort, which may bias the characteristics of the patients included towards those with higher mucus hypersecretion and more severe disease. Moreover, bronchiectasis was present in 60% of the patients studied, which seems to be a slightly higher percentage than that observed by previous authors, with the potential displacement of the inflammatory biomarkers in one particular direction. Third, although sputum samples were collected, cytokines such as interleukin 4, 5 and 13 or proteins such as interferons were not measured, which might also have been of use to understand the underlying mechanisms. Fourth, the presence of ongoing treatment could be interpreted as a limitation. However, we believe that this is a reliable reflection of the patients that professionals observe in daily clinical practice, being an assessment that needs to be taken.

In conclusion, our study reveals that airway or blood eosinophilic T2 inflammation status is not a valid tool to recognize EPA or predict asthma exacerbations in moderate or severe patients following controller therapy. While high levels of T2 inflammatory biomarkers may be related to a higher risk of exacerbation, patients with EPA receiving ongoing treatment may not present high levels of T2 inflammation at the time of assessment. Future studies aimed at determining the molecular pathobiology of EPA are needed to further understand the discrepancies existing between asthma phenotypes and their prognostic implications.

MFP, MJC, DM, CRM, DE, AO, MAR, XM and IO contributed substantially to the study design, data analysis and interpretation, and the writing of the manuscript.

MFP, MJC, DM, CRM, DE, AO and MAR do not have conflicts of interest related to this paper.

IO declares to have received honoraria in the last three years for participating as a speaker in meetings sponsored by Astrazeneca, Boehringuer-Ingelheim, Chiesi, and Novartis and as a consultant for Astrazeneca, GlaxoSmithKlein, Puretech and Sanofi. He has received financial aid from Astrazeneca, Bial and Chiesi for congress attendance and has received grants from Sanofi for research projects.

XM reports a relationship with Sanofi that includes: funding grants, speaking and lecture fees, and travel reimbursement. Xavier Munoz reports a relationship with GlaxoSmithKline that includes: funding grants, speaking and lecture fees, and travel reimbursement. Xavier Munoz reports a relationship with Novartis Pharmaceuticals Corporation that includes: funding grants, speaking and lecture fees, and travel reimbursement. Xavier Munoz reports a relationship with Boehringer Ingelheim Pharmaceuticals Inc that includes: speaking and lecture fees. Xavier Munoz reports a relationship with Laboratorios Gebro Pharma SA that includes: speaking and lecture fees. Xavier Munoz reports a relationship with Menarini Laboratories that includes: travel reimbursement. Xavier Munoz reports a relationship with Faes Farma that includes: travel reimbursement. Xavier Munoz reports a relationship with Chiesi Pharmaceuticals Inc that includes: speaking and lecture fees.