COBRAPed cohort: Do sensitization patterns differentiate children with severe asthma from those with a milder disease?

It is unclear whether sensitization patterns differentiate children with severe recurrent wheeze (SRW)/severe asthma (SA) from those with non‐severe recurrent wheeze (NSRW)/non‐severe asthma (NSA). Our objective was to determine whether sensitization patterns can discriminate between children from the French COBRAPed cohort with NSRW/NSA and those with SRW/SA.


| INTRODUC TI ON
6][7][8][9] However, it is still unclear whether severity in preschool and school-age children is underpinned by different patterns of sensitization. 10 Component-resolved diagnostics (CRD) detects IgE specific to vs. 4%, p < .01)and associated with an FEV1/FVC < −1.64 z-score.Among sensitized children, seven clusters with varying patterns were identified.The two broader clusters identified in each age group were characterized by "few sensitizations, mainly to HDM."One cluster (n = 4) with "multiple sensitizations, mainly to grass pollen, HDM, PR-10, and nsLTP" was associated with SA in school-age children.

Conclusions:
Although children with wheeze/asthma display frequent occurrences and high levels of sensitization, sensitization patterns did not provide strong signals to discriminate children with severe disease from those with milder disease.These results suggest that the severity of wheeze/asthma may depend on both IgE-and non-IgE-mediated mechanisms.

K E Y W O R D S
asthma, preschool, school-age, sensitization, severe asthma

G R A P H I C A L A B S T R A C T
The contents of this page will be used as part of the graphical abstract of html only.It will not be published as part of main article.IgE sensitization patterns in severe recurrent wheeze/school-age asthma.
7]10 Previous results from the Pediatric Cohort of Bronchial Obstruction and Asthma (COBRAPed) of preschool and school-age children with recurrent wheeze/asthma suggest a role for both environmental factors and atopy in asthma severity. 11The description of sensitization profiles using CRD provides an opportunity to further study the relationship between allergic sensitization and asthma severity during childhood.
We aimed to determine whether sensitization patterns can discriminate between children with SA/SRW and those with milder disease.

| Study design and participants
A description of the cohort has been published 11,12 and is available in the Online Supplement.Ethical approval and written informed consent were obtained.The study is registered in ClinicalTrial.gov (NCT02114034).
Children were assigned to four groups: non-severe preschool recurrent wheezers (NSRW), severe preschool recurrent wheezers (SRW), non-severe school-age asthmatic children (NSA), and severe school-age asthmatic children (SA).Atopy was defined as having at least one positive skin-prick test and/or specific IgE levels (≥0.35 kuA/L) against airborne and/or food allergens.Patients with SRW and SA receiving omalizumab were excluded from this analysis.

| Detection and classification of component-specific IgE antibodies
IgE to 112 allergenic components were measured using an ImmunoCAP Immuno Solid-Phase Allergen Chip (ISAC) (Thermo Fisher/Phadia A, Uppsala, Sweden).Levels of component-specific IgE (c-sIgE) antibodies were reported in ISAC Standardized Units (ISU).To determine sensitization at the c-sIgE level, depending on the nature of the analysis, we dichotomized c-sIgE using a binary threshold (< or ≥0.30ISU) or based on the supplier's four-group categorical classification (negative: <0.3 ISU, low: 0.3-1 ISU, medium/ high: ≥1-15 ISU, very high: ≥15 ISU) (Figure 1). 10,13,14Sensitization was also defined at the biological source level based on the food/airborne biological sources (e.g., egg, cow's milk, etc.) or molecular family for cross-reactive components (e.g., PR-10: pathogenesis-related protein family 10 (PR-10), etc.).Multivariable logistic regression analyses were built with the inclusion of all biological sources with univariate p-values <0.1.Resulting odds ratios (OR) were reported with a 95% confidence interval (CI)

Key message
Children with wheeze/asthma display frequent occurrences and high levels of sensitization, but c-sIgE sensitization patterns did not provide strong signals to discriminate between non-severe and severe recurrent wheeze/asthma.Sensitization to non-specific lipid transfer protein (nsLTP) components was more frequent among SA than NSA and was associated with lung function impairment.Cluster analysis of the results for sensitized children identified seven clusters, of which the two largest were characterized by "few sensitizations, mainly to house dust mite (HDM)."Only one small cluster consisting of "multiple sensitizations, including to nsLTP," was associated with severe asthma at school age.

| Description of the population
Among the 295 children with available ISAC data, 47 were classified as NSRW, 78 as SRW, 108 as NSA, and 62 as SA (Figure 1).Their main characteristics are presented in Table 1.Briefly, children with SRW were more frequently exposed to second-hand smoke and visible mold/dampness.Children with SA had a more frequent history of food allergy and atopic dermatitis than those with NSA.Atopy status was similar between NSRW and SRW or NSA and SA, respectively.

| Sensitization profile differences between non-severe and severe patients
We observed individual c-sIgE sensitization (at least one positive c-sIgE ≥0.30ISU) for 51.4% of preschool children and 75.3% of schoolage children.

| Age and sensitization profiles
We observed an increase in the numbers and levels of c-sIgE sensitization with age, both among non-severe and severe patients (Figure 2).There was an increase in the number of positive biological sources for airborne (RR 1.14 [1.08-1.20],p < .0001)and cross-reactive c-sIgE (RR 1.18 [1.07-1.30]per one-year increase, p = .00098),but not for food biological sources (Figure S2, Table S4).

| Lung function and sensitization profiles
Among the 235 participants with available data on LF, there was no relationship between c-sIgE sensitization and the forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) z-score, except for the frequency of nsLTP sensitization, higher for the participants with a FEV1/FVC z-score < −1.64 than in the others (16.7% vs. 5.2%, p = .017)(Table S5).

Asthma outcomes
Age at first wheeze (months)    (Continues)

| Supervised multivariate analysis
PCA was performed with the c-IgE values for the preschool children.PC1 accounted for 20.3% of the variance and PC2 for 11.7%.
Overall, PCA did not allow differentiation between NSRW and SRW.
Similarly, the random forest did not allow discrimination between NSRW and SRW, with an estimated out-of-bag error rate of 43.1% and a ROC AUC of 0.56 (Figure S3).
Among school-age children, PCA, with PC1 explaining 24.1% of the variance and PC2 10.4%, did not allow differentiation between NSA and SA.Similarly, the random forest did not allow discrimination between NSA and SA, with an estimated out-of-bag error rate of 33.9% and a ROC AUC of 0.53 (Figure S4).

| Main results
We aimed to determine whether sensitization profiles of children with SRW or SA could be distinguishable from those with NSRW or NSA using a CRD multiplex assay.Overall, the patterns of biological source sensitization did not discriminate between children with  .23 Note: c-sIgE were dichotomized using a binary threshold (positive ≥0.30ISU).Abbreviations: CCD, cross-reactive carbohydrate determinants; HDM, house dust mite; NSA, non-severe school-age asthmatic children; nsLTP, non-specific lipid transfer protein; NSRW, non-severe preschool recurrent wheezers; PR-10, pathogenesis-related protein family 10; SRW, severe preschool recurrent wheezers; TLP, thaumatin-like proteins.

TA B L E 2 (Continued)
and nsLTP was more frequent among children with SA, and sensitization to nsLTP was associated with impaired LF.Unsupervised clustering confirmed the heterogeneity in sensitization profiles, identifying three clusters for preschoolers and four for school-age children with shared patterns but also some specificities (grass and PR-10 among preschoolers and nsLTP among school-age children).
Only one small cluster with multiple airborne and nsLTP sensitization was associated with asthma severity at school age.

| Most sensitized children with recurrent wheezing/asthma show comparable patterns
Although preschoolers were less frequently sensitized than schoolage children, the sensitization profiles in the two age groups showed strong similarities.The two broader clusters identified in each age group were characterized by few sensitizations, mainly to HDM, and were comparable to clusters described in the U-BIOPRED cohort. 10nsitization to HDM and multi-sensitization were even more frequent among preschoolers with NSRW than those with SRW, supporting that disease severity is associated with exposure to mold and cigarette smoke rather than atopy in this age group. 11This suggests that the drivers of inflammation may differ between NSRW and SRW.In this regard, in a previous paper by our teams, airway inflammation in SRW was found to be more neutrophilic than eosinophilic. 16One could hypothesize that the imbalance between type 2 and non-type 2 mechanisms in the preschool years favors a more severe presentation in SRW.The finding that patterns of sensitization to biological sources did not discriminate between children with SA/SRW and those with milder disease confirms the results from the U-BIOPRED cohort. 10The similarity of sensitization profiles between children from the two groups suggests that, at least among sensitized children, asthma in school-age children may share common features with wheezing in preschoolers.Abbreviations: FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; HDM, house dust mite; PR-10, pathogenesis-related protein family 10.
These results confirm that sensitization patterns may not be useful biomarkers of disease severity in children when described in terms of numbers/levels of c-sIgE sensitization at a single time point. 10More complex endotypic mechanisms than simple allergenic sensitization may underpin asthma severity during childhood.

| Sensitization to certain single components is associated with severity, in particular to nsLTP
Sensitization to Gal d 1 and Cor a 9 and Can f 1 and Can f 2 was more frequent among children with SA than those with NSA.These results confirm that Can f 2 sensitization and multi-sensitization to lipocalins are more frequent among children with SA than those with milder disease. 17,18Interestingly, sensitization to the nsLTPs Cor a 8 and Art v 3 was associated with SA, and sensitization to Pru p 3, a major nsLTP, also tended to be more frequent among SA.In addition, nsLTP sensitization was associated with lower LF.In contrast to the multi-sensitization pattern shown in cluster 5, nsLTP sensitization was a characteristic of the sensitization profile shown in cluster 4, which was the only cluster associated with SA.Among preschoolers, 75% of children from cluster 2, also characterized by nsLTP sensitization, had SRW.This association of nsLTP sensitization with asthma severity has not been described elsewhere.0][21] Sensitization toward nsLTP from pollen and food was observed.It is yet to be determined whether sensitization to nsLTP primarily occurs through pollen or food exposure. 20Although these results need confirmation, they highlight how geographical variation might affect asthma severity.

| The longitudinal follow-up of the cohort will allow the comparison of sensitization patterns as biomarkers of disease trajectories
We observed an increase in sensitization between the ages of 3 and 12 years.3][24] However, mold sensitization was retained in only a limited number of children in our study which did not allow full exploration of its association with severity because of lack of power.The follow-up of this cohort will make it possible to analyze sensitization trajectories.An unbalanced immune reaction biased toward a response involving type 2 helper T cells may be involved in children with early and multiple sensitizations 9,25 and exacerbated interferon production in response to viral infections in children with late-onset sensitization and asthma. 25

| Strengths and limitations
The CobraPed cohort has enrolled a subsequent and well-characterized population.In particular, preschoolers represent a significant number, of whom 61 could be included in the cluster analysis. 10This study had several limitations.We excluded patients receiving omalizumab for obvious reasons, thus severe and often highly atopic patients were excluded. 26,27However, omalizumab being mostly offered to schoolage children, did not influence results for the preschoolers.Because our analysis was exploratory, with no a priori hypothesis, we have not corrected the p-values for multiple testing, which can be seen as a limitation of our study.If we had applied this correction, it would have probably shown null results, further reinforcing our conclusion that overall, sensitization patterns may not be useful biomarkers of disease severity in children and that the severity of asthma may rely on more complex mechanisms than sensitization.Abbreviations: FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; HDM, house dust mite; PR-10, pathogenesis-related protein family 10.

R
V3.3.1 was used for statistical analysis.Continuous variables are presented as medians [interquartile range], and categorical variables as numbers (%).Comparisons of quantitative data were performed using Wilcoxon-Mann-Whitney tests.Categorical variables were analyzed using the chi-square test or Fisher exact test as appropriate.

2
Differences in the sensitization profile between (A) NSRW and (B) SRW and between (C) NSA and (D) SA.
NSRW and SRW or with NSA and SA.At the c-sIgE level, sensitization to airborne allergens, especially towards HDM components, and multi-sensitization, were approximately twice as frequent among preschoolers with NSRW than with SRW.At school age, sensitization to Gal d 1, hazelnut 2S globulin, dog salivary lipocalin proteins,

HDM, PR-10, and nsLTP n = 4 Cluster 5 multiple, mainly airborne including grass pollens and HDM n = 6 Cluster 6 multiple, mainly grass pollens, HDM and PR-10 N = 24 Cluster 7 few, mainly HDM n = 94 p-value
Severity and lung function by cluster in school-age children.