Abstract
Background In clinical trials, the two anti-interleukin (IL)-5 monoclonal antibodies (mAbs: mepolizumab and reslizumab) approved to treat severe eosinophilic asthma reduce exacerbations by ∼50–60%.
Objective To observe response to anti-IL-5 mAbs in a real-life clinical setting, and to evaluate predictors of suboptimal response.
Methods In four Canadian academic centres, predefined clinical end-points in 250 carefully characterised moderate-to-severe asthmatic patients were collected prospectively to assess response to the two anti-IL-5 mAbs. Suboptimal response was determined based on failure to reduce maintenance corticosteroid (MCS) or asthma symptoms scores (Asthma Control Questionnaire (ACQ)) or exacerbations, in addition to persistence of sputum/blood eosinophils. Worsening in suboptimal responders was assessed based on reduced lung function by 25% or increase in MCS/ACQ. A representative subset of 39 patients was evaluated for inflammatory mediators, autoantibodies and complement activation in sputum (by ELISA) and for immune-complex deposition by immunostaining formalin-fixed paraffin-embedded sputum plugs.
Results Suboptimal responses were observed in 42.8% (107 out of 250) patients treated with either mepolizumab or reslizumab. Daily prednisone requirement, sinus disease and late-onset asthma diagnoses were the strongest predictors of suboptimal response. Asthma worsened in 13.6% (34 out of 250) of these patients. The majority (79%) of them were prednisone-dependent. Presence of sputum anti-eosinophil peroxidase immunoglobulin (Ig)G was a predictor of suboptimal response to an anti-IL-5 mAb. An increase in sputum C3c (marker of complement activation) and deposition of C1q-bound/IL-5-bound IgG were observed in the sputa of those patients who worsened on therapy, suggesting an underlying autoimmune-mediated pathology.
Conclusion A significant number of patients who meet currently approved indications for anti-IL5 mAbs show suboptimal response to them in real-life clinical practice, particularly if they are on high doses of prednisone. Monitoring blood eosinophil count is not helpful to identify these patients. The concern of worsening of symptoms associated with immune-complex mediated complement activation in a small proportion of these patients highlights the relevance of recognising airway autoimmune phenomena and this requires further evaluation.
Abstract
Severe asthmatics with adult-onset asthma, sinus disease and requiring daily prednisone, are at higher risk for responding suboptimally to current doses of anti-IL-5 mAbs, with further risk of worsening on an IgG1 mAb if they have sputum autoantibodies https://bit.ly/2Ahpvsm
Introduction
Targeting the interleukin (IL)-5 signalling pathway is now an established therapeutic strategy for patients with severe asthma whose severity is predominantly driven by eosinophils. Mepolizumab and reslizumab are neutralising monoclonal antibodies (mAbs) directed against IL-5, while benralizumab is an afucosylated mAb directed against the IL-5 receptor [1]. All three molecules effectively deplete blood eosinophil levels and, on average, reduce exacerbations by ∼50–60% [2–4]. There could be a number of reasons why the response is not more impressive. These include 1) selection of patients based on peripheral blood eosinophil counts which may not always be associated with airway eosinophilia [5, 6]; 2) inadequate dosing that may not suppress airway eosinophils [7]; 3) IL-5 not being the dominant cytokine driving eosinophilia [8]; and 4) airway autoimmune mechanisms (reported in the airways of up to one-third of patients with eosinophilic asthma [9]) that may interfere with the effects of the biologics [7, 10]. Local autoimmune response may interfere when IL-5 in the airway (which is a predictor of response to anti-IL-5 therapies [7]) is not adequately neutralised by anti-IL-5 mAb, and instead, when in a zone of equivalence with the antigen [11] may form IL-5–anti-IL-5 heterocomplexes with endogenous immunoglobulin (Ig)G autoantibodies and consequently activate the complement pathways [12]. This could lead to not just a suboptimal response, but potentially worsening of asthma. These phenomena would not be observed with benralizumab (since it is not directed against the antigen, IL-5, rather against the receptor), and therefore this article does not report clinical responses to benralizumab.
The primary objective of our study was to examine the prevalence and clinical predictors of suboptimal response and worsening of asthma in a real-world setting in 250 patients with moderate-to-severe asthma from four Canadian university hospitals who were prescribed mepolizumab or reslizumab based on best clinical practice and national regulatory guidelines. The secondary objective was to examine the molecular predictors, including autoimmune responses, in those patients who were suboptimal responders. In a smaller, but representative, subset of patients with available sputum samples, we assessed inflammatory mediators (cytokines and released eosinophil products). We also analysed autoantibodies, immune-complex formation/deposition and complement activation. Finally, this study was not designed to test the superiority of one mAb over another, but only to report the rate of response to anti-IL-5 mAbs based on patient-related outcomes routinely used by clinicians in a pragmatic real-world scenario. Some of the results of this analysis have been reported previously in the form of an abstract [13].
Methods
Study patients and collection of clinical data
Either mepolizumab or reslizumab was prescribed as add-on therapy in patients with severe asthma (adult) with significant eosinophilia (blood and/or sputum), inadequately controlled despite treatment with a high-dose inhaled corticosteroid (ICS) plus another controller (equivalent to the Global Initiative for Asthma criteria 4 and 5) [14], according to the Health Canada prescription guidelines. Mepolizumab was available in Canada for patients as early as January 2016, whereas reslizumab became available ∼12 months later. All the initial prescriptions were for mepolizumab. When both drugs were available, prescriptions of the anti-IL-5 therapy was at the physician's discretion; dependent on the patient's insurance coverage and logistical considerations of administering subcutaneous versus intravenous dosing. Forced expiratory volume in 1 s (FEV1) and Asthma Control Questionnaire (ACQ)-5 [15] were obtained as a part of routine follow-up in the clinical centres. In 160 subjects, sputum was induced and processed [16] as a part of routine clinical assessment. Fractional exhaled nitric oxide (FENO) was not routinely collected as part of clinical assessment, except in patients who could not produce sputum. These data are not included in this article. The clinical data were collected prospectively as and when the patients were routinely followed in their respective clinics, from the time when they were started on the adjunct anti-IL-5 therapy (pre-treatment value or baseline) until the database was locked or they were taken off the treatment due to inadequate/suboptimal response (between November 2015 and January 2019) (post-treatment values). For 250 patients who were included in the final analysis (figure 1) any additional data (refer to table 1 for list of clinical parameters) were collected as a part of retrospective chart review with ethical permission from the respective local ethics board. Matched pre- and post-treatment sputum for 39 patients (with adequate volume) for molecular analysis was available at the McMaster site (figure 1) and all experimental procedures were reviewed and approved by the Hamilton integrated research ethics board.
Steroid reduction strategy
After three doses of the anti-IL-5 mAb (i.e. 3 months), the maintenance dose of steroid was reduced by 5 mg every month, and asthma control was assessed with 3-month follow-up visits. For patients on long-term (>2 years) high-dose maintenance prednisone (>15–20 mg), tapering was halted at 5 mg·day−1, and then patients were tested for adrenal insufficiency before completely weaning them off. For patients maintained on high-dose ICS (≥1500 μg fluticasone propionate equivalent), the tapering was done at 500 μg (reduction by 1 puff twice daily every clinic visit). Asthma control during steroid tapering was monitored every 3 months using ACQ, sputum and/or spirometry.
Assessment of response to anti-IL-5 mAb therapy
Suboptimal response (NR) to therapy was determined by one of the three clinical criteria (failure to reduce maintenance corticosteroid by 50%, failure to reduce ACQ-5 ≤1.5, failure to reduce exacerbations by 50%) plus persistence of sputum eosinophils >3% or blood eosinophils ≥400 cells·µL−1 after ≥4 months of treatment (adapted from a previous study [7]). Without a biological criterion (reduction in eosinophilia), the presence of any two of the three clinical criteria were accepted as designating NR status. Worsening on treatment was defined as patients identified as NR having a further one of the following criteria: reduction in FEV1 from pre-treatment baseline by ≥25%; any increase in maintenance corticosteroid; and increase in ACQ by 0.5 (minimal clinically important difference).
Assessment of inflammatory markers
Type 2 T-helper cell cytokines and mediators were evaluated using the Discovery assay (Eve Technologies, Calgary, AB, Canada), as described previously [9, 17], while eosinophil peroxidase (EPX) as a measure of airway eosinophil activity was assessed using ELISA [18, 19]. Anti-EPX IgG [9] and Ig-bound C1q (using ELISA, #ab170246; Abcam, Cambridge, MA, USA) were detected in the sputa along with C3c, a marker of complement activation (#HK368-01; HyCult Biotech Inc., Wayne, PA, USA. Detailed methodology for detection and quantification of immune-complexes in the paraffin-embedded formalin-fixed sputum plugs is given in the supplementary material.
Statistical analysis
Detailed information on statistical analysis and strategies are provided in the supplementary material for both clinical and experimental cohorts reported here. Briefly, we calculated descriptive statistics of demographic and baseline clinical characteristics within groups of responders and suboptimal responders to anti-IL-5 therapy. Means±sd were used to summarise continuous variables, while frequencies and percentages were used to summarise categorical variables. From the logistic regression models, we reported the odds ratio and 95% confidence intervals associated with each clinical predictor (detailed method in the supplementary material). For the molecular end-points, the Kruskal–Wallis test with Dunn's multiple comparison was used between responders, suboptimal or nonresponders (NR) and those who worsened (WR). Prism (version 8; GraphPad, La Jolla, CA, USA) was used for statistical analysis and generation of plots.
Results
Study patients (recruitment and inclusion into analysis)
Based on the recent consensus statement by Buhl et al. [20] that severe eosinophilic asthmatics prescribed biologics should receive treatment for ≥4 months, in the principal analysis we included 250 patients (out of 289) with at least four injections of either reslizumab/mepolizumab. The baseline demographic data for these 250 patients (figure 1) stratified by their response to anti-IL-5 mAb therapies is provided in table 1. In addition, the baseline data distribution between the four academic sites and baseline demographics for 148 patients who had ≥12 injections are provided in supplementary tables E1 and E2, respectively.
Patient response to anti-IL-5 mAb treatment: suboptimal response
43% of patients showed suboptimal response to anti-IL-5 mAbs (table 1). Furthermore, for 148 patients who received ≥12 months of therapy, the rate of suboptimal response, 42.5% (n=63), was comparable (Chi-squared, 0.002, degrees of freedom (df) 1; p=0.96). The clinical variables determined to designate clinical response are stratified based on the treatment groups in table 2 (subgroup analysis: supplementary table E4).
In ICS-dependent asthma: frequency of suboptimal response
The proportion of patients with suboptimal response was 33% (n=36) in 109 patients maintained on daily high-dose ICS (median dose 1500 μg of fluticasone equivalent), with 20 patients failing to reduce exacerbations by 50% and 12 failing to reduce at all (or documented an increase; table 2).
In oral corticosteroid-dependent asthma: frequency of suboptimal response
Out of 141 oral corticosteroid (OCS)-dependent asthmatics (median daily dose 10 mg), 50.3% showed inadequate response to either of the prescribed anti-IL-5 mAbs, with 52% (62 out of 119) nonresponse to mepolizumab and 41% (nine out of 22) to reslizumab (figure 1). Out of the 141 OCS-dependent patients who showed a suboptimal response to anti-IL-5 therapy, ∼34 (24%) could not reduce their OCS dose at all, and 56 (40%) failed to reduce it by 50% (table 2). The rate of NRs in the OCS group was higher than those maintained on ICS only (Chi-squared 7.4, df 1, z=2.6; p=0.008). The mean reduction of OCS in the responders was 74.3% compared to 12.2% in the suboptimal group (p<0.0001).
Patient response to anti-IL-5 mAb treatment: worsening of symptoms
34 (13.6%) patients worsened while on prescribed anti-IL-5 mAbs (table 2), of whom 27 (79%) were maintained on daily prednisone. Based on the three criteria for designating worsening, 1) 12 (35.2%) showed an increase in ACQ (median increase by 0.8 points); 2) 14 (41.1%) had their maintenance corticosteroid dose increased (median increase in prednisone by 7.5 mg, ICS by 500 μg·day−1; 3) 11 (32.3%) recorded a fall in FEV1 by ≥25% (median fall of 31.2%, range 25–59%). The mean FEV1 post-treatment recorded for 11 patients was 50.1±18% pred (Δ877±385 mL). Finally, only three patients were designated as “worsened” based on more than one of the three clinical criteria.
Clinical predictors for suboptimal response to anti-IL-5 treatment
Use of daily prednisone (OR 1.92, 95% CI 1.15–3.22), dose of prednisone (OR 1.04, 95% CI 1–1.08) and presence of sinus disease (OR 1.88, 95% CI 1.13–3.14) could independently, in a univariate analysis predict “suboptimal” response to an anti-IL-5 mAb (table 3). This is further reflected in the baseline difference (table 1) for OCS (p=0.02) and sinus disease (p=0.02) between the two response groups (table 1). A multivariate regression model (table 3) showed that there was an increased possibility of a suboptimal response in late-onset asthmatics (OR 4.5, 95% CI 1.14–17.79) with evidence of sinus disease (OR 3.47, 95% CI 1.23–9.78) and/or atopy (OR 3.28, 95% CI 1.11–9.68). In the OCS-dependent group, subgroup analysis revealed ∼28-fold increase in risk of suboptimal response (OR 27.87, 95% CI 1.3–599.25) if the patient has adult-onset asthma, and ~12-fold risk (OR 12.66, 95% CI 2.17–74.02) with evidence of sinus disease.
Molecular mechanisms underlying inadequate response to anti-IL-5 treatment
For molecular investigations, a smaller subset of prototype patients was recruited with clinical characteristics representative of the larger clinical cohort of 250 patients (supplementary table E3).
Assessing eosinophil-associated inflammatory mediators as fluid phase markers
An increase in sputum IL-5 was noted in the mepolizumab suboptimal responders post-treatment (figure 2) (p=0.04), indicating inadequate neutralisation of the antigen. Univariate analysis showed a significant predictive value only for granulocyte–macrophage colony-stimulating factor (GM-CSF) (table 4) (estimate±se −6.3±2.9, p<0.03). However, the absolute values (figure 2d) are predominantly below the lower limit of detection, and hence the role of GM-CSF remains inconclusive. As depicted in figure 2g, sputum levels of EPX were unremarkable at baseline and were not computed to be a predictor of suboptimal response (table 4) (OR 0.97, 95% CI 0.77–1.22; p=0.8).
Assessing autoimmune-mediated inflammation (fluid phase)
Anti-EPX IgG was used as a marker of localised autoimmunity [7, 9] (figure 3a). Levels of anti-EPX IgG were elevated in the airways (p=0.03) of patients who failed to respond adequately to mepolizumab (figure 3). Additionally, in mepolizumab-treated patients who showed worsening of symptoms, both C1q-IgG and C3c levels were significantly elevated from baseline (figure 3).
Commensurate with the distribution plots for all inflammatory markers assessed (figures 2 and 3), a multivariate regression analysis confirmed baseline anti-EPX IgG levels (at 1:2 titre) to be a predictor for suboptimal response to an anti-IL-5 mAb (estimate±se −1.77±0.7, z-value −2.2; p=0.02) (table 4 and supplementary table S3).
Assessing in situ immune-complex deposition in fixed sputum plugs
Figure 4 details the co-localisation of C1q-IgG and IL-5 IgG in the airways of patients who suboptimally responded to mepolizumab; this was not done in reslizumab since there was no evidence of C1q-IgG or C3c in their sputa (figure 3). The absolute values of the differential cell count for the respective sputum samples (from which the plug was selected for embedding) have been given. As is evident, there is no particular increase in a cell-type which limits us to the assessment of whether there is a particular cell associated with the C1q-IgG/IL-5 IgG co-localisation. A significant correlation (r=0.8, p<0.0001) between the IL-5+IgG+/C1q+IgG+ dual-positive cells indicates a mutually inclusive event (figure 4c, representative image in supplementary figure S3).
Discussion
We report suboptimal response in a population of moderate-to-severe eosinophilic asthmatics to adjunct anti-IL-5 mAb treatment (107 (42.8%) out of 250). In addition, we report a novel observation of worsening of asthma in 13.6% (34 out of 250) of patients with “eosinophilic asthma” on anti-IL-5 neutralising mAbs. Oral corticosteroid dependence, late onset of asthma, and sinus disease (paradoxically, these are the best indicators of an IL-5-driven eosinophilia) were the strongest predictors of suboptimal response. Presence of autoimmune responses in the airways, formation of heterocomplexes and complement activation contributed to the worsening of asthma.
The response rate to anti-IL-5 mAb therapies reported in this study is consistent with phase III clinical trials [2–4] and real-life cohort studies [21–25]. However, the criteria used in our study and the other observational studies were different, and therefore it is challenging to contrast our data with the other publications. Unlike our study, the other observational studies do not provide any insight into mechanisms of poor response. Our investigations reveal that suboptimal responders to both anti-IL-5 therapies had increased baseline anti-EPX IgG levels. Furthermore, anti-EPX IgG was observed to be the only molecular factor that could predict a possible nonresponse to anti-IL-5 mAbs (table 4). A significant positive correlation between EPX levels and anti-EPX IgG (r=0.45, p=0.02) suggested an ongoing autoimmune response that sustains the ongoing eosinophilic activity, and the events remain uncurbed despite the high-dose corticosteroid therapy and additional anti-IL-5 mAb. The clinical indicators for nonresponse to anti-IL-5 therapy (e.g. OCS use and sinus disease) further reflects a population that is prone to localised autoimmune inflammation as reported in recent studies [9, 26]. This phenomenon is localised to the airways as there was no increase noted in the systemic markers of autoimmune responses (anti-neutrophil cytoplasmic antibodies, rheumatoid factors or complement activation). C-reactive protein, a marker of acute inflammation, was unremarkable between the pre-treatment and post-treatment levels (pre-Rx median (interquartile range (IQR)) 2.3 (5.5); post-Rx median (IQR) 3.9 (8.9); p=0.36).
These autoantibodies, being of the IgG subtype, can bind to complement (C1q). In the event that the drug is inadequate for the target antigen [11], there may be immune-complexes formed with the IgG1 mAb as well as the IgG autoantibodies, forming heteroimmune-complexes (for example, IL-5–IgG/EPX–anti-EPX IgG), bound to the six heads of a C1q molecule [27]. C1q activation can either induce the complement cascade or heighten inflammation by recruiting immune cells via FcRγ receptors without binding other complement factors [28]. Increased anti-EPX IgG, C3c and C1-q/IgG levels in the fluid phase, along with C1q–IgG/IL-5–IgG dual-positive cells (in the sputum plugs, see supplementary figure E4 for representative pre- versus post-mepolizumab image), strongly suggest an autoimmune-triggered immune-complex mediated pathology in the patients who worsened on mepolizumab. Worsening on reslizumab could not be conclusively linked to a similar pathology due to limited sample size. The three reported cases of worsening (representative image of low/no immune-complex deposition is given in supplementary figure E5) is unlikely to be mediated by immune-complex mediated complement activation, possibly because reslizumab has an IgG4 backbone which does not bind (C1q) complement [29].
In addition, we tested the hypothesis that poor response might have been due to IL-5 not being the dominant driver of eosinophilia in these patients. Evidence from sputum transcriptomics data from the U-BioPred study [30] strongly suggest that in addition to IL-5, other cytokines related to innate lymphoid cell group 2 biology such as IL-33, thymic stromal lymphopoietin and IL-13 may be mediators of/contributors to eosinophilia. However, except for GM-CSF, none of the other cytokines related to eosinophil biology were predictors of suboptimal response. We believe that IL-5 is indeed the dominant cytokine in the severe eosinophilic patients as we observed an increase in sputum IL-5 in the patients who showed inadequate responses to mepolizumab (figure 2a), indicative of poor neutralisation of target antigen, and perpetuation of ongoing eosinophilic inflammation. Finally, in patients who worsened (indicated by red symbols in figure 2c) there was an increase in IL-13 in the sputa, which agrees with the cytokine signature we previously reported for asthmatics with sputum autoantibodies [9].
Finally, absolute blood eosinophil counts, currently recommended to be the best biomarker for anti-IL-5 mAb treatment response, is not supported by our current observation in a dataset of 250 patients. Indeed, post-treatment blood eosinophils were elevated (≥400 cells·µL−1) in only 8% of the suboptimal responders (table 2). In contrast, 76% of the 67 suboptimal responders with available airway inflammometry data (table 2), showed increased sputum eosinophils ≥3%, indicating unsuppressed airway eosinophilia. Furthermore, 68.6% of these patients had sputum eosinophils despite normalisation of blood eosinophils. These further agree with previous reports on discordance between blood and sputum eosinophils [5], and the former to be an inadequate biomarker for monitoring therapeutic response to anti-IL-5 mAbs [7, 31, 32], particularly in prednisone-dependent asthmatics, as depicted in table 2. In similar prospective real-life studies, baseline blood eosinophil levels were not predictive of response to mepolizumab [21, 22, 33]. Although FENO correlates with airway eosinophilia in steroid-naïve patients, it is not helpful to monitor response to therapies with anti-IL5 mAb [34] and therefore, it was not routinely measured in this study.
Although this study is one of the largest detailed description of clinical responses to anti-IL-5 mAbs, we acknowledge the limitation due to incomplete data collection of a few variables in approximately a third of patients, particularly sputum cytology, as this is not part of routine clinical assessment in all academic centres. We have attempted to address this in our statistical analysis outlined in the supplementary material. A second limitation is the low numbers of patients who were on reslizumab, and therefore the current study does not attempt to make any direct comparisons between the clinical efficacies of the two mAb therapies. However, even with the limited samples, we were able to establish that the underlying factors leading to inadequate response and/or worsening for these two anti-IL-5 mAbs appear to be different. Finally, although we speculate that underdosing may be responsible for the suboptimal response to the anti-IL-5 mAbs, we have not provided pharmacokinetic data to support this hypothesis. Validated assays of mAb concentrations in airway secretions are not commercially available and we were not able to obtain these assays from the manufacturers of the said therapeutic molecules.
In summary, we report a significant prevalence of suboptimal response to the two currently approved anti-IL-5 neutralising mAbs in moderate-to-severe asthmatics associated with blood and/or sputum eosinophilia. This is unlikely to be due to IL-5 not being the dominant cytokine in perpetuating eosinophilia. Indirect evidence suggests that inadequate neutralisation of IL-5 in the airways may be relevant. More importantly, we report the presence of an alternative inflammatory event, airway autoimmunity, that may compromise anti-IL-5 treatment efficacy, and in some leads to worsening of symptoms and airway obstruction via immune-complex mediated injury. Monitoring of blood eosinophil count may not be helpful to identify this. Clinicians ought to be mindful of this possibility while prescribing anti-IL-5 mAb therapy to asthmatics who are prednisone-dependent, with late onset asthma diagnosis, evidence of a sinus disease, and therefore likely to have a higher burden of IL-5 in their airways.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-00117-2020.SUPPLEMENT
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-00117-2020.Shareable
Acknowledgements
We acknowledge technical and administrative support from Naheed Rezzae, Janice Halbecki, Nicola LaVigne (McMaster University, Hamilton, ON, Canada) and Jocelyne L'Archevêque (Sacré-Coeur Hospital of Montreal, Montreal, QC, Canada).
Footnotes
This article has supplementary material available from erj.ersjournals.com
Author contributions: P. Nair and M. Mukherjee developed the concept, M. Mukherjee designed all experiments. P. Nair, J.G. Martin, C. Lemiere and L-P. Boulet recruited patients. S. Tran, M-E. Boulay, M. Bertrand, H. Al-Hayyan, J. Cherukat, M. Kjarsgaard and C. Huang collected and tabulated the data. K. Radford and M. Mukherjee performed molecular experiments. T. Javkar performed all immunostaining protocols while K. Ask, S.D. Revill and A. Ayoub did the microscopy and related validation of image analysis. A. Dvorkin-Gheva, N. Dendukuri and D.F. Forero undertook all statistical analysis. M. Mukherjee wrote the first draft. P. Nair, J.G. Martin, K. Ask, C. Lemiere, L-P. Boulet and N. Dendukuri edited and added to the development of the manuscript. All authors have read and agreed to the submitted manuscript. P. Nair and M. Mukherjee take overall guarantee of the manuscript.
Conflict of interest: M. Mukherjee reports grants from Canadian Institutes of Health Research and Canadian Allergy, Asthma, and Immunology Foundation, grants and personal fees from Methapharm Specialty Pharmaceuticals, and personal fees from Astrazeneca, outside the submitted work.
Conflict of interest: D.F. Forero has nothing to disclose.
Conflict of interest: S. Tran has nothing to disclose.
Conflict of interest: M-E. Boulay has nothing to disclose.
Conflict of interest: M. Bertrand has nothing to disclose.
Conflict of interest: A. Bhalla has nothing to disclose.
Conflict of interest: J. Cherukat has nothing to disclose.
Conflict of interest: H. Al-Hayyan has nothing to disclose.
Conflict of interest: A. Ayoub has nothing to disclose.
Conflict of interest: S.D. Revill has nothing to disclose.
Conflict of interest: T. Javkar has nothing to disclose.
Conflict of interest: K. Radford has nothing to disclose.
Conflict of interest: M. Kjarsgaard has nothing to disclose.
Conflict of interest: C. Huang has nothing to disclose.
Conflict of interest: A. Dvorkin-Gheva has nothing to disclose.
Conflict of interest: K. Ask reports grants from Canadian Pulmonary Fibrosis Foundation, Ontario Thoracic Society, Synairgen, GSK, Indalo, Unity, Avalyn, Canadian Institutes of Health Research and Synairgen, grants and personal fees from Boehringer Ingelheim, outside the submitted work.
Conflict of interest: R. Olivenstein has nothing to disclose.
Conflict of interest: N. Dendukuri has nothing to disclose.
Conflict of interest: C. Lemiere reports grants and personal fees for advisory board work from AstraZeneca and TEVA Innovation, personal fees for advisory board work from GlaxoSmithKline, Sanofi and Novartis, outside the submitted work.
Conflict of interest: L-P. Boulet has nothing to disclose.
Conflict of interest: J.G. Martin has nothing to disclose.
Conflict of interest: P. Nair reports grants and personal fees for lectures from AZ and Teva, grants from Novartis and Sanofi, grants and personal fees for consultancy from Roche, personal fees for lectures from Novartis, personal fees advisory board work from Merck and Equillium, outside the submitted work.
Support statement: This study was funded by a research grant awarded to P. Nair by the Canadian Institutes of Health Research. P. Nair is supported by the Frederick E. Hargreave Teva Innovation Chair in Airway Diseases. M. Mukherjee is supported by an investigator award from CIHR and Canadian Allergy, Asthma and Immunology Foundation/AllerGen NCE. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received January 20, 2020.
- Accepted May 8, 2020.
- Copyright ©ERS 2020