Abstract
U-BIOPRED aims to characterise paediatric and adult severe asthma using conventional and innovative systems biology approaches.
A total of 99 school-age children with severe asthma and 81 preschoolers with severe wheeze were compared with 49 school-age children with mild/moderate asthma and 53 preschoolers with mild/moderate wheeze in a cross-sectional study.
Despite high-dose treatment, the severe cohorts had more severe exacerbations compared with the mild/moderate ones (annual medians: school-aged 3.0 versus 1.1, preschool 3.9 versus 1.8; p<0.001). Exhaled tobacco exposure was common in the severe wheeze cohort. Almost all participants in each cohort were atopic and had a normal body mass index. Asthma-related quality of life, as assessed by the Paediatric Asthma Quality of Life Questionnaire (PAQLQ) and the Paediatric Asthma Caregiver's Quality of Life Questionnaire (PACQLQ), was worse in the severe cohorts (mean±se school-age PAQLQ: 4.77±0.15 versus 5.80±0.19; preschool PACQLQ: 4.27±0.18 versus 6.04±0.18; both p≤0.001); however, mild/moderate cohorts also had significant morbidity. Impaired quality of life was associated with poor control and airway obstruction. Otherwise, the severe and mild/moderate cohorts were clinically very similar.
Children with severe preschool wheeze or severe asthma are usually atopic and have impaired quality of life that is associated with poor control and airflow limitation: a very different phenotype from adult severe asthma. In-depth phenotyping of these children, integrating clinical data with high-dimensional biomarkers, may help to improve and tailor their clinical management.
Abstract
Children with severe preschool wheeze or severe asthma are usually atopic and have impaired quality of life http://ow.ly/RrrGE
Introduction
Asthma is one of the most common chronic diseases in childhood. Although many achieve control with currently available therapies, an estimated 5–10% of patients remain symptomatic despite receiving large amounts of treatment. These children with severe asthma [1] have poor quality of life (QoL), frequent asthma attacks and lung function impairment, are at high risk of side-effects from medication and account for significant medical and societal costs.
It is increasingly recognised that asthma, and particularly severe asthma, is not one single disease entity. Data in adults have been available for some time [2] but evidence now exists in children to suggest that there are a number of different clinical manifestations of severe asthma that are driven by a variety of pathophysiological mechanisms [3, 4]. Phenotypic classifications in children have primarily been based on clinical data, lung function measurement and assessment of allergic status. The small number of studies that have included biological samples have described important differences in the underlying pathobiology of severe asthma in children compared with adults [5–7]. Some but not all preschool children with severe wheeze have evidence of airway remodelling and inflammation from an early age [8], consistent with established asthma, but little is known about the underlying mechanisms, which in many cases appear to be very different from school-age and adult asthma. These early changes do not always predict a progression to asthma [9]. These observations are indicative of considerable heterogeneity amongst children with severe school-aged asthma or severe preschool wheeze.
In order to capture the relevant phenotypes of children with severe asthma or severe wheeze, careful and extensive clinical characterisation is required. This provides the basis for future integration with biological disease markers. The Unbiased Biomarkers for the Prediction of Respiratory Disease Outcomes (U-BIOPRED) project is a public–private partnership, within the framework of the Innovative Medicines Initiative, bringing together academic institutions and European Federation of Pharmaceutical Industries and Associations partners from across Europe. It was set up in 2009 in order to take advantage of the combination of extensive clinical characterisation and biological fingerprinting by “omics” technologies for the unbiased discovery of phenotypes in both adult and paediatric severe asthma [10]. The paediatric arm of the U-BIOPRED study used the same thorough clinical characterisation and innovative systems biology approach as the adult study [11] to integrate clinical, physiological and inflammatory data and patient/parent-reported outcomes with the high-dimensional data of “omics” technologies (transcriptomic, proteomic, lipidomic and metabolomic) obtained from blood, urine, breath and airway samples [12].
The main objective of this first report of the paediatric U-BIOPRED study was to fully clinically characterise the cohorts of children with severe asthma and preschool wheeze and mild/moderate cohorts based on cross-sectional baseline data. The second objective was to investigate the burden of severe asthma and severe preschool wheeze, as described by QoL, and the clinical factors that relate to this burden.
Methods
This was a multicentre, prospective, observational cohort study following the life course of asthma. Full details of the adult cohorts are described in the companion paper by Shaw et al. [11].
Cohorts
Seven centres in five European countries recruited preschool (age 1–5 years) and school-age (age 6–17 years) children. Four paediatric cohorts were recruited by approaching consecutive patients attending respiratory and general paediatric clinics who fulfilled the following inclusion criteria. 1) Severe school-aged asthma (SA): ongoing poorly controlled asthma (persistent symptoms, frequent exacerbations or persistent airflow limitation) despite high-dose inhaled corticosteroids (ICS) and at least two other controller medications [13]. 2) Mild/moderate school-aged asthma (MMA): controlled or partly controlled asthma, prescribed low-dose ICS and no other or one additional controller medication. 3) Severe preschool wheeze (SW): persistent symptoms and frequent exacerbations despite current or failed high-dose ICS and a leukotriene receptor antagonist (LTRA). 4) Mild/moderate preschool wheeze (MMW): controlled or partially controlled symptoms prescribed no treatment or low-dose ICS and/or a LTRA.
Full cohort descriptions, inclusion and exclusion criteria are shown in table S2 in the online supplementary material. All children in the severe cohorts (SA and SW) had been under follow-up with a respiratory paediatrician for ≥6 months before enrolling in the study. During this time, assessments were undertaken to exclude other diagnoses, treat comorbidities, optimise asthma control, assess medication adherence (e.g. checking prescription uptake) and reduce allergen exposure in sensitised individuals [13].
Study design
The study was approved by the local ethics committees (see table S1). Parents/caregivers provided written consent; children gave assent where appropriate. The study is registered with ClinicalTrials.gov (NCT01982162). The study aims and outcomes have been published on the U-BIOPRED website (www.europeanlung.org/en/projects-and-research/projects/u-biopred/home).
All participants were identified and recruited locally and attended a screening and baseline visit. Clinical data and biological samples were collected from all cohorts (figure 1). Full details are provided in the supplementary material.
Visit schedule. SA: severe school-aged asthma; SW: severe preschool wheeze.
Study assessments
Baseline data included demographics, current and past medical history (including detailed asthma and atopic disease history), medications, birth history, family history, exposure to environmental tobacco smoke, and known clinical and environmental risk factors. Asthma control was assessed using the Asthma Control Test (ACT) (for children ≥12 years of age) [14] or the Childhood Asthma Control Test (cACT) for children <12 years [15]. Non-scheduled healthcare utilisation was assessed by documenting exacerbations. QoL was assessed using the Paediatric Asthma Quality of Life Questionnaire (PAQLQ) (school-aged children only) and the Paediatric Asthma Caregiver's Quality of Life Questionnaire (PACQLQ) [16, 17]. Adherence was evaluated using the Medication Adherence Report Scale (MARS) [18]. A summary of the assessments carried out in each of the cohorts is shown in table S3.
In all cohorts, total IgE, specific IgE tests and/or skin prick testing (SPT) to six common allergens were undertaken. Atopy was defined as the presence of sensitisation on SPT (wheal ≥3 mm) or serum specific IgE (≥0.35 kU·L−1). Spirometry before and after bronchodilator [19] and exhaled nitric oxide fraction (FeNO) were measured where possible. Sputum induction was performed in the school-aged cohorts and differential cell counts were determined. Exposure to environmental tobacco smoke was assessed by measuring urinary cotinine. In selected centres, forced oscillation technique and plethysmography were undertaken.
Full details of the methods are provided in the supplementary material, including samples collected for future “omic” analysis. A centralised biobank was selected for the study and operated in accordance with its own Standard Operating Procedures, as described in the supplementary material.
Data management and statistics
Data were entered into an electronic clinical record form before being transferred to the tranSMART system for quality control checks [20]. Missing data were not imputed.
The cohort sizes of 97 and 43 (comparing SA and MMA), and 77 and 54 (comparing SW and MMW), both provide 80% power to detect a difference in means of half a standard deviation, assuming standard normally distributed data, in a two-sided t-test at the 5% significance level [21].
Due to the descriptive (as opposed to inferential) nature of the analyses presented, raw, unadjusted p-values are reported throughout the manuscript. Those in tables 1–4 were derived using logistic regression (binary variables) or general linear regression (continuous variables). Continuous variables exhibiting positive skew were summarised by the median and interquartile range (IQR), and were log-transformed prior to association testing. Where appropriate, tests of association were performed both with and without adjustments for age and sex.
Baseline demographic characteristics and medical history
Asthma history and treatment
Lung function and airway inflammation
Asthma quality of life, exacerbations and control
Associations between key potential facets of asthma (forced expiratory volume in 1 s (FEV1) z-score, FEV1/forced vital capacity (FVC), age of onset/diagnosis, number of exacerbations in preceding 12 months, ACT z-score, body mass index (BMI), MARS, hay fever, eczema, atopy, smoking and white race) were each assessed singly for association with asthma burden, as quantified by QoL, using linear regression. Adjustments for age and sex were not applied at this stage due to a lack of univariate association between either age or sex. QoL contour plots were derived for continuous variables with p<0.05, using two-dimensional kernel density estimation with a bivariate normal kernel, evaluated at 50 grid points in each direction [23]. The variables were also modelled jointly in a multivariate general linear model. Backwards stepwise regression using the Akaike Information Criterion was then applied, in order to derive a parsimonious model.
Analyses were performed using R version 2.15.2 (R Core Team; www.r-project.org). The present report is based on cross-sectional analysis of the baseline data.
Results
Participants
A total of 298 children and teenagers with asthma or wheeze were screened to recruit 282 participants. Numbers of participants in each cohort that provided baseline questionnaire data, spirometry, blood samples and sputum samples are detailed in figure 2.
Consort diagram for participants in the paediatric Unbiased Biomarkers for the Prediction of Respiratory Disease Outcomes (U-BIOPRED) study. GWAS: genome-wide association study.
Cohorts SA and MMA were well matched for age (mean 12.21 and 11.26 years, respectively), as were cohorts SW and MMW (mean 3.56 and 3.46 years, respectively). Exposure to environmental tobacco smoke was reported by 15.8–22.8% of each cohort. More of the SW cohort were positive for urinary cotinine than of the MMW cohort (19.4% versus 4.3%; p=0.035) (table 1).
Atopy
Most of the school-age participants in both cohorts (SA, MMA) were atopic (85.4% and 89.5%, respectively) (table 1). Rates of atopy were lower in both preschool cohorts (36.5% and 37.5%) (table 1). The majority of the school-age children (SA, MMA) had a diagnosis of eczema, hay fever or allergic rhinitis (table 1). Most of the preschoolers had a co-existing diagnosis of eczema with a third also having allergic rhinitis. In the preschool children, significantly more SW than MMW participants had a diagnosis of hay fever (58.8% versus 36.1%, respectively; p=0.04). A sizeable minority of participants reported symptoms of food allergy (40.2% for SA versus 32.6% for MMA, p=0.39; 21.1% for SW versus 27.8% for MMW, p=0.38).
Asthma history and treatment
The mean age at diagnosis was in the fourth year of life for both school-aged cohorts, whereas for the preschool ones it was in their second year (table 2). There were significant differences in the triggers for respiratory symptoms between the severe and mild/moderate cohorts (table 2). While almost all of cohorts SA, MMA and SW were treated with ICS, they were prescribed for less than half of MMW as most had failed to respond to ICS therapy. Additionally, 23.7% of SA and 5.2% of SW were receiving maintenance oral corticosteroid therapy. Parent/participant-reported adherence to therapy was good in all cohorts (table 2).
Lung function and airway inflammation
Lung function and bronchodilator reversibility in the SA and MMA cohorts were similar at baseline when participants were well (table 3). For preschool participants able to perform spirometry, results were again similar for severe and mild/moderate cohorts. There was a trend towards specific airway conductance being lower in the SA cohort compared with the MMA cohort (1.58 versus 1.95 kPa·s−1; p=0.054) (table 3). We were only able to collect induced sputum from a minority of school-aged participants so we could not make a meaningful comparison between the cohorts (table 3).
Asthma burden: QoL, control and exacerbations
Asthma-related QoL was used as the primary measure of burden. The mean result for the PAQLQ for the SA cohort was 4.77, equivalent to “somewhat bothered”, significantly worse than for the MMA cohort (5.8, equivalent to “bothered a bit”; p<0.001). Similar differences were found for the symptoms, emotions and activity domains (table 4). For the preschool cohorts the PACQLQ was used as a proxy, given that there is no validated QoL tool for preschool wheeze. For SW the mean was 4.27 (“some of the time” or “somewhat worried/concerned”), again significantly worse than for MMW (6.04, “hardly ever” or “hardly worried/concerned”; p<0.001).
The burden of asthma was also illustrated by the ACT results, which assessed ongoing symptoms and rescue medication. Most of the severe cohorts were uncontrolled (74.6% in SA compared with 29.2% in MMA, p<0.001; 78.0% in SW compared with 18.2% in MMW, p<0.001). This was reflected in the number of exacerbations in the year prior to assessment. In the previous year, the SA cohort had a median of three exacerbations (IQR two to five), compared with one (IQR zero to two) in the MMA cohort (p<0.001). A similar difference was seen between the SW and MMW cohorts (table 4). However, there was still an important asthma burden in the mild/moderate cohorts.
Which factors are associated with asthma burden?
Asthma burden is described as asthma-related QoL, with z-scores used to give a combined variable for all age groups. Pre-bronchodilator FEV1, but not FEV1/FVC ratio, was significantly related to QoL (regression coefficient 0.151, p=0.002) (table 5). The number of exacerbations in the previous year was significantly inversely associated with asthma QoL (−0.52, p<0.001). Asthma control (measured by ACT and cACT z-score) was significantly related to asthma QoL (0.730, p<0.001). BMI was inversely associated with asthma QoL (−0.036, p=0.011). These are illustrated in figure 3. Results were similar when PAQLQ and PACQLQ were considered separately (table S4).
Factors associated with asthma burden as measured by quality of life. Figures represent scatter plots describing the relationship between each factor and the combined asthma-related quality of life z-score. The contour lines are coloured blue to red, to indicate increasing density of points in the graph to overcome the issue of overlying data points. The contour plots show c) a strong positive relationship between quality of life and asthma control (Asthma Control Test (ACT)) with a) a weaker positive relationship between quality of life and lung function (forced expiratory volume in 1 s (FEV1) z-score). Additionally there is b) a strong negative relationship between quality of life and exacerbation rate plus d) a weaker negative relationship between quality of life and body mass index (BMI). The combined z-score merges the Paediatric Asthma Quality of Life Questionnaire (PAQLQ) and the Paediatric Asthma Caregiver's Quality of Life Questionnaire (PACQLQ). The density was modelled using two-dimensional kernel density estimation.
Factors associated with asthma burden as measured by quality of life
To assess which factors were independent predictors of asthma-related QoL, a backward stepwise regression analysis was performed for FEV1 z-score, FEV1/FVC, age of onset/diagnosis, number of exacerbations in preceding 12 months, ACT z-score, BMI, MARS, hay fever, eczema, atopy, smoking and white race. Significant factors in the reduced model were ACT z-score (regression coefficient 0.76, p<0.001) and FEV1 z-score (0.11, p=0.036).
Discussion
This article presents the detailed clinical characteristics of 282 children in four paediatric cohorts, including preschool and school-age children with both severe and mild/moderate wheeze and asthma across Europe. Standard Operating Procedures and Good Clinical Practice criteria were used to ensure consistency and quality across sites, with data collected on an online platform and stored in a single online repository. The severe cohorts by definition had a significantly higher treatment burden than the mild/moderate ones, despite which they remained poorly controlled with frequent severe exacerbations and low ACT scores. Children with severe disease, and their caregivers, had significantly lower QoL scores across all domains than the mild/moderate cohorts. Asthma control and airway obstruction were found to be significantly associated with QoL. Exposure to environmental tobacco smoke in the SW cohort was a striking finding and will be an important concomitant factor in future analyses. Otherwise, the severe and mild/moderate cohorts were very similar; this is in contrast to the adult severe and mild/moderate U-BIOPRED cohorts [11] and suggests that paediatricians should be cautious about extrapolating from adult studies. The vast majority of children were atopic. The rates of reported food allergy were high, although the rate of actual food allergy is expected to be much lower [24]. Most had a normal BMI, unlike the typical adult severe asthma phenotype. Also conspicuous was the morbidity in the mild/moderate paediatric groups; although they were clearly differentiated from the severe groups, a number are clearly sub-optimally treated. These data demonstrate that we succeeded in recruiting severe paediatric cohorts and provide a comprehensive view of the clinical burden of severe asthma or wheeze in childhood.
Children in the U-BIOPRED SA and SW cohorts have frequent symptoms and severe exacerbations that adversely impact on QoL and carry a high treatment burden; almost 17% of the SA cohort was prescribed omalizumab and 24% prescribed maintenance oral corticosteroids. This is in keeping with a previous study, which reported a strong association between health-related QoL and ACT score in children with problematic severe asthma [25]. In our study, the impact on QoL was seen to be greatest for the SW cohort. A significant impact on QoL was also seen in the mild/moderate cohorts, highlighting the often overlooked influence of asthma on the lives of children and their families. Allergic sensitisation and other atopic diseases were a frequent finding across all cohorts, more so in school-aged children, adding further to the treatment burden. Lung function was not significantly different in the school-aged cohorts, possibly due to good treatment adherence, being between exacerbations and the fact that FEV1 is not a good discriminator of severity.
U-BIOPRED builds upon previous severe asthma cohort studies [26–29]. However, to our knowledge, this is the first study to recruit preschool wheeze cohorts on the basis of a consensus definition, which can be directly compared with parallel school-age and adult cohorts [13]. Most studies of preschool wheeze have been based on birth cohorts. A small number of studies have focused on severe preschool wheeze [9, 30] and they have provided valuable insights into the underlying pathophysiology and natural history of preschool wheeze. In common with the TENOR (The Epidemiology and Natural History of Asthma: Outcomes and Treatment Regimens) [27] and SARP (Severe Asthma Research Program) [26, 28] severe asthma cohorts, U-BIOPRED children with severe asthma were commonly atopic, had high healthcare utilisation and a high treatment burden. In the TENOR study there were far more boys than girls (63% versus 37%) in the severe cohort but, in common with SARP, we did not see these sex differences. Unlike SARP, children in the SA cohort did not have significantly higher FeNO levels than those in the MMA cohort; however, FeNO measurements were made off-line in SARP, making it difficult to make direct comparisons.
There are a number of limitations to this study. There were no healthy controls recruited to the paediatric cohorts; however, as the aim was to understand what makes asthma severe, the mild/moderate cohorts are the most appropriate comparator. The mild/moderate asthma group were all on prophylactic medication and participants were recruited from general paediatric and respiratory clinics so they are not completely representative of the children with mild/moderate asthma or wheeze seen in primary care. Also, as this is a multicentre pan-European study, it is likely that there were differences in patients recruited into each cohort between centres. Feasibility and safety considerations meant that assessments such as airway hyperresponsiveness were not included. Additionally, preschool children were unable to perform lung function, induced sputum and FeNO. We were not able to reach the target of 100 preschool severe wheeze children; many had not been under tertiary follow-up for ≥6 months, did not reach the treatment threshold or did not meet the stringent inclusion criteria at screening due to the intermittent nature of their symptoms. There was no objective measure of adherence during the study; however, this was a pragmatic study of real-life severe asthma where clinics had tried to exclude adherence issues, and the high MARS scores suggest a good level of adherence.
Despite advances in recent years in our understanding and management of severe asthma, the data presented here highlight the ongoing unmet needs. Both severe asthma and severe wheeze are heterogeneous diseases. Single or even clustered biomarkers have had limited impact in predicting clinical course or therapeutic efficacy in children: for example, the SA cohort is not distinguishable from MMA by classical lung function and airway inflammatory phenotypes [5, 7, 9]. Classification of preschool wheeze phenotypes is at an even more basic level, limited to symptom pattern [31, 32] and progression to asthma determined retrospectively. Analysis of samples from these cohorts will provide high-dimensional biological (“omics”) data, which can be integrated with clinical characteristics to define multidimensional handprints of severe asthma. This approach has the potential to allow a step change in our understanding of asthma, identify more relevant prognostic and therapeutic targets and enable a personalised, phenotype-driven approach to management to address the unmet burden.
Acknowledgements
This paper is presented on behalf of the U-BIOPRED Study Group with input from the U-BIOPRED Patient Input Platform, Ethics Board and Safety Management Board.
The members of the U-BIOPRED Study Group are as follows. Kamran Tariq and Patrick Dennison: NIHR Southampton Respiratory Biomedical Research Unit, Clinical and Experimental Sciences, NIHR-Wellcome Trust Clinical Research Facility, Faculty of Medicine, University of Southampton, UK; Annelie F. Behndig: Umeå University, Sweden; Wim van Aalderen, Rene Lutter, Ariane Wagener, Kees van Drunen, Pieter-Paul Hekking, Paul Brinkman and Koos Zwinderman: Academic Medical Centre, University of Amsterdam, The Netherlands; Nadia Mores, Giuseppe Santini and Salvatore Valente: Università Cattolica del Sacro Cuore, Italy; Christos Rossios, David Gibeon and Uruj Hoda: Imperial College London, UK; João Pedro Carvalho da Purificação Rocha, Adesimbo Sogbesan, Julaiha Gent and Andrew Menzies-Gow: Royal Brompton and Harefield NHS Foundation Trust, UK; Davide Campagna: University of Catania, Italy; Jeannette Bigler, Michael J. Boedigheimer, Wen Yu and Xugang Hu: Amgen Inc.; Klaus Fichtner, Katja Nething, Damijan Erzen, Ralf Sigmund, Kathrin Riemann, Alix Berton and Matthias Klüglich: Boehringer Ingelheim Pharma GmbH & Co. KG, Germany; Martina Gahlemann, Boehringer Ingelheim (Schweiz) GmbH, Switzerland; Jens Hohlfeld, Philipp Badorrek and Cornelia Faulenbach: Fraunhofer ITEM; Anna James, Elisabeth Henriksson and Roelinde Middelveld: Karolinska Institutet, Sweden; Ann-Sofie Lantz and Karin Strandberg: Karolinska University Hospital & Karolinska Institutet, Sweden; Gabriella Galffy, M. Szentkereszty and Lilla Tamasi: Semmelweis University, Budapest, Hungary; Katherine M. Smith: University of Nottingham, UK; Bertrand De Meulder and Diane Lefaudeux, University of Lyon, France.
The members of the U-BIOPRED Paediatric Study Group are as follows. Ann Berglind, Jon R. Konradsen and Päivi Söderman: Karolinska University Hospital & Karolinska Institutet, Sweden; Klaus Bønnelykke, Nadja H. Vissing and Jakob Stokholm: University of Copenhagen & Danish Pediatric Asthma Center, Gentofte Hospital, University of Copenhagen, Denmark; Gina Kerry, Lesley Lowe, Katy Johnson, Simon Stephan and Meera Sunthar: University of Manchester, UK; Emily Guillmant-Farry and Sheila Nsubuga: Royal Brompton and Harefield NHS Foundation Trust, UK.
The authors would like to acknowledge Elizabeth Juniper (McMaster University, Hamilton, ON, Canada) (PAQLQ and PACQLQ) and Robert Horne (University College London, London, UK) (MARS) for the use of their questionnaires. They would also like to acknowledge help in data and knowledge management from the eTRIKS project, which is funded by the Innovative Medicines Initiative. Additionally, the authors would like to acknowledge the support of the University Hospital NHS Foundation Trust and the NIHR-Wellcome Trust Clinical Research Facility, Southampton, UK.
Footnotes
This article has supplementary material available from erj.ersjournals.com
This study is registered on ClinicalTrials.gov (NCT01982162).
Support statement: The research leading to these results has received support from the Innovative Medicines Initiative (IMI) Joint Undertaking, under grant agreement no. 115010, resources for which are composed of financial contribution from the European Union's Seventh Framework Programme (FP7/2007–2013) and kind contributions from companies in the European Federation of Pharmaceutical Industries and Associations (EFPIA) (www.imi.europa.eu). A. Bush [CP5] was supported by the NIHR Respiratory Disease Biomedical Research Unit at the Royal Brompton and Harefield NHS Foundation Trust and Imperial College London. Help in data and knowledge management was received from the IMI-funded eTRIKS project (EU grant code no. 115446). In Southampton, support was received from the University Hospital NHS Foundation Trust and the NIHR-Wellcome Trust Clinical Research Facility.
Conflict of interest: Disclosures can be found alongside the online version of this article at erj.ersjournals.com
- Received May 18, 2015.
- Accepted August 23, 2015.
- Copyright ©ERS 2015
References
