Research highlights from the 2018 European Respiratory Society International Congress: airway disease

Basic and translational data

Asthma is one of the most common inflammatory diseases globally and affects over 200 million people worldwide [1]; however, little is known regarding the detailed mechanisms driving the disease [2]. Asthmatic disease shows a high degree of heterogeneity [3] and this topic was elegantly addressed by Prof. Sally Wenzel in her talk “Molecular phenotyping of asthma comes of age” in the symposium “State of the art session: airways disease”. In line with this, Nawijn et al. [4] showed, using single cell RNA sequencing, that direct comparison of airway wall cells from asthmatic patients and healthy controls revealed selective loss of well-differentiated ciliated cells accompanied by an increase in basal cells in asthma. Specific inflammatory cell patterns were also observed in asthma and Tokunaga et al. [5] presented that CD163, a scavenger receptor expressed on macrophages, was significantly increased in asthma death patients compared to control subjects. They also confirmed in a mouse model of allergic inflammation that airway hyper-responsiveness (AHR) and the number of eosinophils in bronchoalveolar lavage fluid (BALF) were significantly decreased in CD163 deficient mice when compared with control wild-type mice. This indicates an important role for altered phenotypes of macrophages not only in chronic obstructive pulmonary disease (COPD), as previously reported, but also in asthma. It is evident that overlap between the inflammatory response in asthma and COPD exists, especially when it comes to patients who smoke. Silberbrandt et al. [6] reported that, in a severe asthma cohort with predominantly late-onset asthma, airway eosinophilia was significantly associated with a smoking history of ≥10 pack-years. They concluded that further research was needed to explore underlying mechanisms driving eosinophilic airway inflammation in severe asthma. Another factor in phenotyping asthma is age of onset and Mistry et al. [7] showed that there are phenotypic differences between early-onset and adult-onset difficult asthma. They showed that early-onset difficult asthma is more common in females and associated with more atopy, functional comorbidity, biologic therapy and healthcare need. Adult-onset difficult asthma meanwhile is more common in males associated with smoking and comorbid lung disease. Mapping clinical phenotypes and correlating them to molecular endotypes and inflammatory patterns will be of great importance to increase our knowledge of mechanisms driving asthmatic disease as well as treatment strategies.

Much focus during the 2018 European Respiratory Society (ERS) International Congress was on the mechanistic understanding of cell–microbe communications in the lung micro-environment and how they promote airway disease or protect patients from it. The symposium “Cross-talk in the lung microenvironment: implications for chronic lung disease” (session 218) focused on different aspects of the cellular communication mechanisms in the lung that participate in health and disease, with presentations by Profs. D. Davies, P. Hansbro, S. Krauss-Etschmann and S. Gabrielsson. The 2018 ERS International Congress also had a “Meet the Expert” session on the topic “Lung microbiome: study design and standardisation” (ME2). Lately, a shift in focus away from microbial involvement in infections and toward a role in physiology has been observed. For a long time the lungs of healthy individuals were considered to be free from microbes and hence to lack a microbiome. However, while the lung microbiota has a low density, the maintenance of small numbers of bacteria seems to be a critical determinant of good health [8]. Potaczek et al. [9] demonstrated that interleukin (IL)-6, a central cytokine of innate immunity responses, seems to be important for the ability to modulate the adaptive immune system in a non-allergic Type-2 T-helper cell (Th2) direction. They also showed that IL-6 plays a specific role in mediating the development of protective, anti-allergic mechanisms related to prior Acinetobacter lwoffii exposure. Probiotics can be considered nonspecific adjuvants of innate immune response by modulating the Type-1 T-helper cell (Th1)/Th2 balance. Aimbire et al. [10] showed that probiotic Lactobacillus bulgaricus mediated an anti-inflammatory effect in an experimental mouse model of allergic inflammation. They showed that dendritic cell expression of CD86 as well as the Toll-like 4 receptor was increased after oral administration of L. bulgaricus. In addition, L. bulgaricus attenuated eosinophil infiltration, mucus production and specific IgE concentration in serum, and also inhibited secretion of IL-4, IL-5, IL-13 and eotaxin in BALF. In a clinical study de boer et al. [11] presented interesting data from the BREATHE study where bacterial lysates showed positive effects in the prevention of asthma exacerbations in patients with severe asthma.

Clinical studies

Asthma is a heterogeneous inflammatory airway disease and the majority of asthmatics (59%) treated in general practice are uncontrolled [12]. When physicians recognise poor asthma control, an appropriate change to therapy must be considered. At this year's Congress, Van der Meer et al. [13] presented their research on dynamic hyperinflation, which, independent of asthma severity, is associated with poorer overall health, less well-being and impaired activity in daily life. Due to its major impact on activity in everyday life, dynamic hyperinflation should be an important target for treatment in asthma patients in the future.

It is also well known that asthma is associated with accelerated rate of lung function decline. The relationship between decline and airway inflammation among asthmatics has important therapeutic implications. In a large cohort, Backman et al. [14] found that adult asthmatics with higher levels of blood eosinophils had a history of excess forced expiratory volume in 1 s (FEV₁) decline compared to non-eosinophilic asthmatics, independent of other factors such as inhaled corticosteroid (ICS) use. Graff et al. [15] presented similar results and their observations go further, showing that an evolution towards an increase in blood eosinophils over time predicts accelerated FEV₁ decline. They further confirmed that this was independent of inflammatory phenotype or ICS treatment category.

A significant proportion of patients with difficult-to-control asthma remained nonadherent to corticosteroid therapy [16]. Alahmadi et al. [17] presented data from the U-BIOPRED study with the aim of investigating the adherence of asthmatics according to the Medication Adherence Report Scale (MARS) and urinary tests. One third of asthmatics reported sub-optimal adherence, while 42% of the severe asthmatics did not have detectable urinary prednisolone or metabolites. In this study, there was very poor agreement between MARS and urinary testing, suggesting that both approaches should be performed as part of a difficult asthma assessment and are important before prescribing expensive novel biological therapies.

Moreover, multiple comorbidities are associated with asthma control, healthcare utilisation and quality of life (QoL) in difficult asthma, and the data presented by Azim et al. [18] demonstrated a high prevalence of atopy, obesity, rhinitis and gastro-oesophageal reflux disease (GORD) in the Wessex Asthma Cohort. In the area of personalised medicine there is increasing interest in the identification of treatable traits. The analysis of the Australasian Severe Asthma Web-based Database revealed that treatable traits, such as allergic sensitisation, upper-airway disease, airflow limitation, eosinophilic inflammation and frequent exacerbations, are more common in severe asthma. Ten traits predicted exacerbation risk with the strongest being frequent exacerbations, depression, inhaler-device polypharmacy, vocal cord dysfunction and obstructive sleep apnoea (OSA) [19]. Psychological stress exacerbates clinical symptoms in patients with asthma and stress has been previously shown to amplify the immune response to asthma triggers [20]. The study by Bacon et al. [21] provides evidence that acute stress negatively influences the physiology of asthmatics. Stress did not affect the caliber of bronchi but was associated with decreased carbon dioxide production, potentially suggesting hypocapnia and hyperventilation.

There are a lot of treatment options in the management of asthma; however, inhaled glucocorticoids remain a cornerstone because of their efficacy, tolerability and rapid onset of action. Unfortunately, there are patients in clinical practice who do not respond to ICS and predictors of such a poor response, such as the absence of eosinophilic inflammation, have been highlighted previously [22]. Marcon et al. [23] presented a study looking at FEV₁ decline in asthma patients treated with ICS. Here, ICS use was associated with increased FEV₁ decline among subjects without sensitisation but not in sensitised individuals. This suggests that a long-term ICS therapy may be more effective in preventing FEV₁ decline in allergic compared to non-allergic asthma [24].

After monitoring and treating comorbidities, 5–10% of the asthmatic population remains uncontrolled despite high-dose glucocorticosteroids and bronchodilator medications. Here, we focus on data for biologics presented at the 2018 ERS International Congress. Targeting IgE, omalizumab is able to reduce exacerbations and improve asthma control and QoL in severe allergic asthmatics. The objective of the study presented by Hannania et al. [25] was to evaluate the reduction of exacerbation in patients treated with omalizumab according to their blood eosinophil counts (BECs). The reduction in exacerbations was 53% in patients with BECs greater than 300 per mm³ and 42% in patients with BECs less than 300 per mm³, suggesting that BEC is not a good predictor of response to anti-IgE.

Monoclonal antibodies targeting cytokines or their receptors may induce the production of anti-drug antibodies. Ortega et al. [26] have looked at the immunogenicity profile of mepolizumab during the Phase III clinical development programme in severe asthma (COSMOS and COLUMBA). There was no correlation between anti-drug antibody titres and change in BECs. Adverse events evaluated as potential systemic allergic reactions were uncommon (2%) and unrelated to the study drug in antibody-positive patients. Wenzel et al. [27] conducted a post hoc analysis of pooled data from the SIROCCO and CALIMA studies on the effect of benralizumab in patients classified according to age of onset (<18, ≥18 to <40 and ≥40 years). They found a higher effect of benralizumab in adult-onset and late-onset asthma in terms of FEV₁ improvement or reduction in exacerbations as compared to early-onset asthma. This suggests that adult-onset asthma with less atopy is the prime target population for anti-IL-5 therapy.

Approved for the treatment of moderate-to-severe atopic dermatitis, dupilumab is a fully human, anti-IL-4R monoclonal antibody (mAb) that inhibits IL-4/IL-13. During the Congress, Corren et al. [28] presented their findings that dupilumab was more effective in asthmatics with chronic rhinosinusitis and nasal polyps. In this sub-population, dupilumab significantly improved asthma control and QoL.

Bronchial thermoplasty (BT) offers an alternative therapeutic option for severe asthmatics with persistent symptoms requiring step five treatment in specialised centres. There is still a lack of data on predictors of response to BT. During this Congress, Torrego et al. [29] reported the results of the real-world BT Global Registry study (including 159 patients). They suggested that healthcare utilisation decreased, clinical outcomes improved and medication usage was reduced at 1 year after BT. Frix et al. [30] also presented that thermoplasty was well tolerated, with an improvement in pre- and post-FEV₁ together with a reduction in exacerbation rate and in the average dose of oral corticosteroids (OCS) in severe non-T2-asthmatics. In addition, Goorsenberg et al. [31] presented data on BT-treated patients and showed that lung function parameters were unchanged after BT. However, a higher airway resistance was observed in non-responders suggesting that a high resistance at baseline could be a negative predictor for BT response. Furthermore, Sano et al. [32] presented BT data showing that responders to BT exhibited higher R5 and R5-R20 (where R5 is the total respiratory system resistance, measured at 5 Hz and R20 is the central resistance, measured at 20 Hz) at baseline compared to the non-responders. The responders also exhibited a decrease in resistance compared to the non-responders 1 year after BT treatment.

Biomarkers

There is an unmet clinical need in COPD for biomarkers which reflect disease activity (i.e. future exacerbations and rate of lung function decline) or predict treatment responsiveness. Blood eosinophils are surrogate markers for airway eosinophilia [33] and previous studies have showed that patients with higher BECs benefit from a reduction in exacerbations when treated with ICS-containing regimes (ICS/long-acting β-agonist (LABA) or ICS/LABA/long-acting muscarinic antagonist (LAMA)). There are still some uncertainties which have to be clarified regarding BEC, as follows: 1) studies are primarily based on Western populations and we do not know if proposed BEC cut-off values (i.e. >300 cells·mm⁻³) can be generalised for all populations; 2) the stability of eosinophil-low and eosinophil-high phenotypes has to be established; 3) the impact of undiagnosed fungal and parasitic infections needs to be determined; and 4) the effect of previous ICS treatment on BEC and its prognostic value need to be determined. Pascoe et al. [34] presented data from the IMPACT trial which compared ICS/LAMA/LABA to ICS/LABA and LABA/LAMA treatments. The authors concluded that increasing BECs were associated with increased ICS effects on exacerbation on a continuous scale. Mathioudakis et al. [35] investigated the data from the ISOLDE trial and confirmed that BEC is predictive for ICS response in ICS-naïve COPD patients [36]. Furthermore, direct ICS withdrawal from triple therapy led to increased risk of moderate/severe exacerbations in patients with consistently high BECs in the SUNSET trial [37].

Markers predicting future mortality in COPD are warranted to identify patients at high risk who may benefit from early interventions and close follow-up. Waschki et al. [38] presented results from the COSYCONET cohort where serum high-sensitivity troponin I was evaluated in stable COPD patients and it was found that serum levels >6 ng·L⁻¹ were associated with all-cause mortality even after adjustment for cardiovascular (CV) factors, body mass index (BMI), lung function, modified Medical Research Council (mMRC) dyspnoea score and 6-min walk distance (6MWD) [39]. In addition, Agustí et al. [40] presented data from the ECLIPSE cohort evaluating the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2017 disease groups (A, B, C and D) and their relation to mortality. The authors reported an association between mortality and the highly symptomatic groups (B and D). This highlights the clinical relevance of symptom assessment as an important therapeutic goal in COPD patients [41]. Furthermore, Prudente et al. [42] reported that BEC may have a long-term predictive value, as high BECs were associated with increased 9-year mortality in a Brazilian cohort of COPD patients.

Markers identifying exacerbated COPD patients at high risk are also clinically important outside of stable disease. According to the results of Jørgensen et al. [43], high serum soluble urokinase plasminogen activator receptor (suPAR) may highlight patients with an increased risk for 30-day mortality. Interestingly, Papathanassiou et al. [44] reported a significantly higher 1-year mortality in COPD patients who had low BECs at hospital admission. In contrast, low blood eosinophil levels were associated with lower rates of re-admission in a Russian study [45].

The new GOLD ABCD severity grading system

The updated GOLD 2017 classification of COPD severity grading generated debate and its impact was evaluated by several studies presented at the Congress. Compared to GOLD 2013, GOLD 2017 recommends a simplified ABCD assessment tool, using only exacerbation history and respiratory symptoms to guide treatment, while spirometry is used only for diagnosis and prognostic assessment [46]. Exclusion of spirometry from the ABCD assessment tool leads to the reclassification of a proportion of patients from GOLD groups C and D to groups A and B. As a result, less intensive treatment is now recommended for these patients, who have significantly affected pulmonary function (FEV₁ <50% predicted) but do not experience frequent exacerbations or burdensome symptoms at stable state. This approach is justified by the fact that currently available COPD pharmacotherapy targets respiratory symptoms and exacerbations but cannot delay lung function decline or prevent mortality. Therefore, it is anticipated that the newer GOLD classification will decrease the therapeutic burden of these patients without affecting their clinical outcomes. Application of GOLD 2017 criteria in a cross-sectional study involving over 19 000 patients from the UK Clinical Practice Research Datalink (CPRD) led to reclassification of 37% of all participants from groups C and D (according to GOLD 2013) to groups A and B [47, 48]. Only 16% of all patients with COPD were classified as GOLD groups C or D according to GOLD 2017, compared to 54% according to GOLD 2013 [47, 48]. A cross-sectional study in Poland involving 2597 patients with COPD, as identified by primary care physicians, confirmed that the application of GOLD 2017 leads to a decrease in patients allocated to GOLD groups C and D by 60% [49]. Less than 8% were classified as group C and 21% as group D. In the ECLIPSE cohort, only 7% and 15% of all participants were allocated to GOLD groups C and D, respectively, compared to 23% and 40% according to the GOLD 2013 classification [42]. Overall, application of GOLD 2017 criteria led to reclassification of 25–60% of all patients with COPD who were previously graded as GOLD groups C or D, but were now classified as GOLD groups A or B.

The recently published 2019 version of the GOLD document suggests that the ABCD system should be used to guide the initial pharmacological management of COPD, while follow-up management should be guided by a review-assess-adjust cycle [50]. The previously described studies evaluated patients who were already receiving treatments that should have limited their symptoms and/or exacerbation burden. Therefore, they must have underestimated the participants' baseline ABCD grade (prior to treatment initiation) and this is a limitation. However, GOLD 2019 also suggests that treatment de-escalation may be needed in cases of ineffective treatment and side effects, but also of symptom resolution or overtreatment. The aforementioned studies reveal a significant proportion of patients who do not experience exacerbations and/or significant symptoms (GOLD groups A and B) and who are potentially overtreated, mostly due to their previous classification as GOLD groups C and D. For example, in the UK CPRD cohort, 22% and 43% of patients classified as GOLD stages A and B, respectively (26% of all patients with COPD), were receiving triple therapy. These patients represent a prime target for a treatment de-escalation trial aimed at limiting therapeutic burden and real-life data on the safety of de-escalation should be assessed.

Spirometry, which is not included in the GOLD 2017 ABCD grading, is an accurate predictor of mortality. For this reason, GOLD now suggests the use of the ABCD system to guide treatment, but the combination of both the ABCD system and spirometry for prognostic evaluation [46]. A number of studies assessed the predictive value of the ABCD system with regards to mortality. First, mortality was evaluated in the 784 participants of the real-life COPD cohort from the Czech Republic [51]. The shift of patients from GOLD groups C and D to groups A and B when using GOLD 2017 resulted in an increase in the mortality rate of group B (25% compared to 18.7% with GOLD 2011) and group C (23.1% compared to 17.9% with GOLD 2011). Secondly, in the ECLIPSE cohort, GOLD 2017 classification resulted in a trend towards increased mortality in all GOLD groups [40]. This suggests that the mortality rate of those reclassified was lower than that of patients who were consistently in groups C or D and higher than that of those who were consistently in groups A or B.