Eosinophil‐derived neurotoxin levels in early childhood and association with preschool asthma – A prospective observational study

Eosinophil‐derived neurotoxin (EDN) is related to childhood asthma, while normal values are lacking. We aimed to document serum EDN levels at 1 and 3 years in general and in non‐atopic children, and explore if EDN levels differed by sex or were associated with preschool asthma at 3 years.


| INTRODUC TI ON
Asthma is a chronic obstructive airway disease that often presents early in life, 1 affecting around 11% of preschool children. 2 Common features of asthma in early childhood include recurrent episodes of wheeze, coughing and chest tightness, often triggered by viral infections, physical activity, allergen exposure and other environmental factors. 3Due to the heterogeneity of asthma and lack of objective tests to document the disease, the diagnostic criteria in young children are mainly based on patterns of symptoms, risk factors, response to treatment and exclusion of alternative diagnoses. 3sinophils play a key role in the development of asthma 4 and other allergic diseases. 5Triggered by exposure to allergens or by infections, eosinophils in mucosal tissue are activated releasing pro-inflammatory degranulation products such as eosinophilderived neurotoxin (EDN). 5This eosinophilic-induced inflammation is also seen in asthmatic children, 4 causing damage and dysfunction of lung tissues leading to airway inflammation. 6The and 20.1 (87.8) μg/L at 3 years (n = 857).Non-atopic children had EDN levels of 24.0 (107) μg/L at 1 year (n = 147), and 17.3 (84.6) μg/L at 3 years (n = 173).EDN levels were higher in boys compared to girls; 32.0 (133) versus 24.5 (97.0) μg/L at 1 year, and 20.9 (96.3) versus 19.0 (72.4) μg/L at 3 years.Preschool asthma was observed in 109/892 (12.2%) children.Higher EDN levels at 1 (>26.7 μg/L) and 3 (≥20.5 μg/L) years were associated with preschool asthma; adjusted OR (95% CI) 2.20 (1.09, 4.41) and 4.68 (2.29, 9.55), respectively.

Conclusion and Clinical Relevance:
We report EDN values in early childhood, demonstrating higher levels at 1 compared to 3 years and in boys compared to girls at both ages.Higher EDN levels at both ages were associated with preschool asthma.
However, EDN cut-off levels for preschool asthma were overall lower than the ULN of non-atopic children, limiting translation into clinical practice.

K E Y W O R D S
asthma, child, eosinophil-derived neurotoxin, PreventADALL, sex, type 2 inflammation, wheeze

G R A P H I C A L A B S T R A C T
In a paediatric population of 1233 children, serum eosinophil-derived neurotoxin (EDN) levels were higher at 1 compared to 3 years, and in boys compared to girls.Higher EDN levels at both ages were linked to preschool asthma.

Key Messages
• Eosinophil-derived neurotoxin (EDN) levels were higher at 1 compared to at 3 years.
• Boys had higher EDN levels at both ages compared to girls.
• Elevated EDN levels in infancy was linked to preschool asthma.
biomarker EDN has no clear circadian rhythm and appears to display stable serum concentrations over time and at various storage temperatures. 7EDN levels reflect eosinophil count [8][9][10] and likely also inflammation as elevated levels have been observed in tissues and body fluids weeks before asthmatic exacerbations manifest. 4,5Compared to eosinophil count, EDN may be a better indicator of asthma control status. 11Case-control studies have reported higher EDN levels in preschool-and school-aged children with asthma 9,[12][13][14][15][16][17] and allergic asthma, 14,18 as well as in asthmatic children without continuous corticosteroid treatment. 16In children with asthma, EDN was associated with respiratory symptoms, 14,15 allergic sensitization (AS), 19 atopic dermatitis (AD), 16 and correlated to symptom burden. 13,14,20In children recovering from airway infections, EDN was higher among wheezing compared to non-wheezing children, 21 and correlated to future wheezing episodes. 22Thus, EDN may be useful in diagnosis and management of childhood asthma, 4 but larger population-based studies are needed to confirm its potential role in clinical practice.
Although several case-control studies have established a link between EDN levels and various allergic diseases, normal values for children are lacking. 6In adult populations, higher EDN levels have been observed in males compared to females, 8 but in children sex differences remains unexplored.Similarly, the relationship between early-life EDN levels in low-risk paediatric populations and the development of allergic diseases is largely undetermined.
Retrieved from a longitudinal birth cohort, the first aim was to document serum EDN levels at 1 and 3 years in general, and in non-atopic children specifically, and second, to explore possible sex differences in EDN levels at both ages.The third aim was to investigate potential associations between EDN levels at 1 and 3 years and any wheeze, preschool asthma, and preschool asthma with AS and/or AD at 3 years.

| Study design
The present study addressed observational research questions based on information prospectively collected from the Preventing Atopic Dermatitis and ALLergies in children (PreventADALL) study.This Scandinavian mother-child cohort, recruiting from the general population, with two randomized controlled interventions; skin care and early food introduction, is described elsewhere. 23,24iefly, the mothers were recruited during the routine ultrasound examination in mid-pregnancy in Norway and Sweden, between 2014-2016.The children, born without serious illnesses at gestational age (GA) ≥35 weeks, were enrolled during the first days of life.The children were invited to clinical follow-ups at 3 and 6 months, and at 1, 2 and 3 years.Informed written consent was obtained from the mothers at enrolment, and from both parents at newborn inclusion.

| Study population
The present study population consisted of 1233/2394 children from the PreventADALL study with serum EDN levels available at 1 and/ or 3 years (Figure 1).

| Serum EDN
Blood samples, collected at 1 and 3 years, were kept in room temperature for 60-90 min before centrifugation and separation of serum, thereafter kept at −20°C for 5-7 days before transferred for storage at −80°C.Serum EDN was measured using ImmunoCAP (Thermo Fisher Scientific, Uppsala, Sweden), according to protocol. 9

| Preschool asthma and wheeze
Information on doctor diagnosed asthma and wheeze was collected from questionnaires (3, 6, 9, 12, 18, 24, 30 and 36 months) and interviews (24 and 36 months).Asthma medication use was retrieved from questionnaires (12, 18, 24, 30 and 36 months).To compensate for missing information in some participants, the children's medical records were reviewed for the prescription of asthma medication between 9 months through 3 years.

| Allergic sensitization
Skin prick tests (SPT) were performed at 3 years using standard allergen solutions for egg, cow's milk, peanut, wheat, soy, cod, F I G U R E 1 Flow chart of the study population (n = 1233).EDN, eosinophil-derived neurotoxin; PreventADALL, Preventing Atopic Dermatitis and ALLergies in children.
birch, grass, dog, cat and house dust mite (Soluprick ALK-Albelló, Hørsholm, Denmark).In addition, at 3 years analyses for specific serum immunoglobulin E (IgE) to egg, cow's milk, peanut, wheat, soy, cod, birch, grass, mugwort, dog, cat, horse, house dust mite and mould (Cladosporium herbarium) were performed using ImmunoCAP (Thermo Fisher Scientific).Allergic sensitization was defined as having a median SPT wheal diameter of ≥3 mm after 15 min subtracting the diameter of the negative control and/or a specific IgE level >0.35 kU A /L towards at least one allergen.

| Clinical skin assessments
Examinations were carried out at 3 years by trained study personnel educated together at workshops to minimize inter-observer variability, using the United Kingdom Working Party (UKWP) and/or Hanifin and Rajka criteria to diagnose children with AD. 25

| Birth and background characteristics
Background data on socio-demographic characteristics, lifestyle and environmental factors were gathered from electronic questionnaires at 18 and 34 weeks of pregnancy.Birth data was collected from electronic medical records.

| Non-atopic children
Non-atopic children at 3 years were defined as children with no history of asthma, wheeze, rBO, doctor diagnosed asthma, use of and/or prescribed asthma medications, AS or AD between birth and 3 years.

| Primary outcome
Serum EDN levels at 1 and 3 years are presented with median, interquartile range (IQR), minimum-maximum (min-max), and the 95th percentile, due to the skewed distributions of EDN levels.The latter measure is defined as the upper limit of normal (ULN) in non-atopic children.

| Secondary outcomes
Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.
Preschool asthma at 3 years was defined as children with ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/ or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years. 26eschool asthma with AS and/or AD at 3 years was defined as having a positive SPT (≥3 mm) and/or specific IgE test (>0.35kU A /L) towards any allergen and/or fulfilling the diagnostic criteria for AD through either UKWP and/or Hanifin and Rajka at 3 years, in addition to meeting the criteria for preschool asthma.

| Statistical analyses
Birth and background characteristics were compared using independent t-tests for continuous variables presented with means and standard deviations (SD), and χ 2 tests or Fisher's exact tests for categorical variables presented with numbers (n) and percentages (%).
The relationship between EDN levels at 1 and 3 years was examined in a robust linear regression model presented with β-coefficient with 95% confidence interval (95% CI), coefficient of determination (R 2 ) and correlation coefficient (r) with 95% CI.Similarly, among the children with paired samples at 1 and 3 years, EDN levels in general and in non-atopic children were compared with Wilcoxon's signed rank test.EDN levels at 1 and 3 years in children with any wheeze, preschool asthma, AS, AD, preschool asthma with AS and/or AS and any allergic disease at 3 years were compared to EDN levels in non-atopic children, using Mann-Whitney U-tests.To determine cut-off levels for 'higher EDN levels' at 1 and 3 years for any wheeze, preschool asthma and preschool asthma with AS and/or AD at 3 years, Youden Index 27 were estimated and EDN levels accordingly dichotomized.
Among the atopic and non-atopic children, associations between continuous and dichotomized EDN cut-off levels at 1 and 3 years and any wheeze, preschool asthma and preschool asthma with AS and AD at 3 years were investigated in univariate and multivariate logistic regression models, presented with odds ratio (OR) with 95% CI.Further details are described in Data S1.

| RE SULTS
The study population consisted of 1233 children of whom 573 (46.5%) were girls, with a mean (SD) age of 12.5 (0.73) and 38.0 (3.65) months at clinical follow-ups (Table 1).Birth and background characteristics among the 411 children with EDN levels available at both 1 and 3 years are described in Table S1.At 3 years, 151/810 (18.6%) had reported any wheeze while not fulfilling the preschool asthma criteria, 109/892 (12.2%) were defined with preschool asthma, 189/1090 (17.3%) were sensitized to at least one allergen, and 269/1108 (22.6%) were diagnosed with AD.In S1 A-B, the distribution of allergic diseases as well as non-atopic children are illustrated, respectively.Birth and background characteristics of non-atopic children and children with preschool asthma at 3 years are described in Table S2.

TA B L E 1
Birth and background characteristics of the included children with information on EDN at 1 and/or 3 years (n = 1233) compared with the remaining cohort without information on EDN at 1 and/or 3 years (n = 1161).

Characteristics Children with EDN at 1 and/or 3 years (n = 1233)
Remaining b Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.
c Preschool asthma at 3 years was defined as having ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years.
e Atopic dermatitis at 3 years was defined as fulfilling the United Kingdom Working Party and/or Hanifin and Rajka criteria at 3 years.
f Any allergic disease at 3 years was defined as children with either any wheeze, preschool asthma, allergic sensitization and/or atopic dermatitis at 3 years.

| EDN levels by sex
In all children, the median EDN levels (μg/L) were higher in boys compared to girls; 32.0 versus 24.5 at 1 year and 20.9 versus 19.0 at 3 years (p < .001),whereas in non-atopic children, the levels at 1 and 3 years were similar between the sexes (Table 3).In the 411 children with EDN levels available at both ages, higher median EDN levels

| EDN levels in children with preschool asthma and other allergic diseases
Compared to non-atopic children, the median EDN levels at 1 year were higher in children with any wheeze (p = .001),preschool asthma (p = .025),AS (p < .001),and AD (p < .001)(Figure 4A), as well as preschool asthma with AS and/or AD (p = .002),and any allergic disease (p < .001) at 3 years (Table 2).Compared to non-atopic children, the median EDN levels at 3 years were higher in children with any wheeze (p = .016),preschool asthma (p < .001),AS (p < .001)and AD (p = .001)(Figure 4B), as well as preschool asthma with AS and/or AD (P < .001),and any allergic disease (p < .001) at 3 years (Table 2).
Compared to girls, boys with preschool asthma, AD and any allergic disease at 3 years had higher median EDN levels at 1 year, while median EDN levels at 3 years were higher in boys with any wheeze at 3 years (p-values < .05;Table 3).

| EDN levels in relation to preschool asthma and other allergic diseases
Among all children, higher EDN levels (μg/L) were identified at 1 and 3 years for any wheeze as ≥27.8 and ≥20.6, respectively, for preschool asthma ≥26.7 and ≥20.5, respectively, and for preschool asthma with AS and/or AD ≥27.8 and ≥20.6 at 3 years, respectively.
Higher EDN levels at 1 year were associated with any wheeze, preschool asthma and preschool asthma with AS and/or AD at  4).Higher EDN levels at 3 years were associated with all three clinical outcomes at 3 years, see Table 4. Including paired EDN levels (1 and 3 years) in the analyses, only the associations between EDN levels at 1 year for any wheeze as well as EDN levels at 3 years for preschool asthma at 3 years were significant (Table S3).The linear relationship between continuous EDN levels at 1 and 3 years by atopic status at 3 years is shown in Figure S2.
Higher EDN levels at 1 year for differentiating preschool asthma at 3 years from non-atopic children had a sensitivity of 61.8%, specificity of 57.1% and a ROC-AUC (95% CI) of 0.60 (0.51, 0.68).The predictive accuracy of EDN levels for differentiating children with any wheeze, preschool asthma and preschool asthma with AS and/ or AD at 3 years from non-atopic children are described in Figure S3 and Table S4.

| DISCUSS ION
In this study of 1233 non-selected children, we report serum EDN levels at 1 and 3 years.Levels of EDN were higher at 1 compared to 3 years, also observed in non-atopic children, and in boys compared to girls at both ages.Higher EDN levels at age 1 and 3 years were TA B L E 2 EDN levels in children with information at 1 (n = 787) and 3 (n = 857) years, at both ages (n = 411), as well as EDN levels at 1 and 3 years by atopic status at 3 years.Wilcoxon's signed rank test.
b Upper limit of normal (ULN) was defined as the 95th percentile.
c Non-atopic children were defined as children with no history of asthma, wheeze, rBO episodes, doctor diagnosed asthma, use of and/or prescribed asthma medications, atopic dermatitis or allergic sensitization between birth and 3 years.
d Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.
e Preschool asthma at 3 years was defined as having ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years.
g Atopic dermatitis at 3 years was defined as fulfilling the United Kingdom Working Party and/or Hanifin and Rajka criteria at 3 years.
h Preschool asthma with AS and/or AD at 3 years was defined as having allergic sensitization and/or atopic dermatitis at 3 years, in addition to meeting the criteria for preschool asthma at 3 years.
i Any allergic disease at 3 years was defined as children with either any wheeze, preschool asthma, allergic sensitization and/or atopic dermatitis at 3 years.associated with any wheeze, preschool asthma and preschool asthma with AS and/or AD at 3 years.
To our knowledge, this is the first report describing EDN levels in non-selected and in non-atopic children retrieved from a large paediatric population at two distinctive time points in early childhood.
Our finding of higher levels at 1 year compared to that at 3 years is supported by similar observations for eosinophil cationic protein (ECP) levels, with higher levels in infancy (0-23 months) compared to that at preschool age (24-41 months), among 245 children. 28On the other hand, two smaller studies found no age differences in ECP levels over a larger age spectrum in childhood (0-15 and 0-12 years, respectively). 29,30Also, blood eosinophil counts were higher in infancy compared to that at preschool-and school age, and in younger children compared to adolescents, in two large studies establishing reference values. 31,32Thus, a higher blood eosinophil count in infancy might hypothetically explain our novel observations.
We are unaware of other reports showing higher EDN levels in boys compared to girls at 1 and 3 years.In line with our findings, in adults, Granger et al. reported higher EDN levels in males compared to females. 8In support, male sex has also been associated with higher eosinophil counts in children. 33Of note, the sex differences were no longer apparent among non-atopic children, perhaps suggesting that more frequent allergic disease in boys may account for the higher levels observed in boys.
5][16][17][18]34,35 In already in infancy, in line with EDN being associated with AS, allergic rhinitis and AD in school children. 16Similarly, in the metaanalysis by Bao et al.,  a Mann-Whitney U-test.
b Upper limit of normal (ULN) was defined as the 95th percentile.
c Non-atopic children were defined as children with no history of asthma, wheeze, rBO episodes, doctor diagnosed asthma, use of and/or prescribed asthma medications, atopic dermatitis or allergic sensitization between birth and 3 years.
d Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.
e Preschool asthma at 3 years was defined as having ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years.
g Atopic dermatitis at 3 years was defined as fulfilling the United Kingdom Working Party and/or Hanifin and Rajka criteria at 3 years.
h Preschool asthma with AS and/or AD at 3 years was defined as having allergic sensitization and/or atopic dermatitis at 3 years, in addition to meeting the criteria for preschool asthma at 3 years.
i Any allergic disease at 3 years was defined as children with either any wheeze, preschool asthma, allergic sensitization and/or atopic dermatitis at 3 years.
Higher EDN levels at 1 and 3 years were associated with any wheeze, preschool asthma and preschool asthma with AS and/or AD at 3 years, in line with previous reports on wheeze 22 and different asthma phenotypes. 16To our knowledge, our study is the largest to investigate associations between infant EDN levels and preschool asthma, as well as EDN levels and concurrent asthma at preschool age.

| Strengths and limitations
The automated ImmunoCAP assay 9 may be beneficial over manual enzyme-linked immunosorbent assays. 37The distinct definitions of any wheeze, preschool asthma, AS, AD, preschool asthma with AS and/or AD, and any allergic disease is strength.a Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.
b Preschool asthma at 3 years was defined as having ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years. c Preschool asthma with AS and/or AD at 3 years was defined as having a positive SPT (≥3 mm) and/or specific IgE (>0.35 kU A /L) towards any allergen (egg, cow's milk, peanut, wheat, soy, cod, birch, grass, mugwort, dog, cat, horse, house dust mite and/or mould) and/or fulfilling the diagnostic criteria for AD through either United Kingdom Working Party and/or Hanifin and Rajka at 3 years, in addition to meeting the criteria for preschool asthma at 3 years.
d Logistic regression models adjusted for tobacco smoke in pregnancy, parental allergic disease (any), sex, GA at birth, caesarean section and the PreventADALL interventions.
e Cut-off levels for higher EDN levels by Youden's Index for any wheeze, preschool asthma and preschool asthma with AS and/or AD at 3 years equal to EDN levels (μg/L) ≥ 27.8, ≥26.7 and ≥ 27.8 at 1 year, and ≥ 20.6, ≥20.5 and ≥ 20.6 at 3 years, respectively.
With the objective to document EDN levels, the slightly higher education level and proportion of allergic diseases among the parents in the PreventADALL cohort than in the general population 24 limit the generalizability of our observations to similar populations, as disease burden, 38 and thereby likely EDN levels, may differ with socioeconomic status.Children sensitized to allergens such as tree nuts will not have been identified at 3 years, but the potential impact on observed associations is likely to be small.Nor can we account for release of EDN triggered by infections. 5As blood eosinophil counts are not available, we are unable to determine the relative usefulness of EDN versus eosinophil counts, to expand on previous findings of associations between eosinophil count and EDN, 8-10 age 31,32 or male sex. 33As EDN levels appear to correlate with eosinophil counts, [8][9][10] one may speculate that the declining trend of blood eosinophils from infancy to school age and adolescence also will be observed in EDN levels if monitored beyond preschool age. 31,32However, the burden of allergic diseases increases with age, suggesting EDN levels may increase, in line with the higher levels in older children reported by Rydell et al. 9

| Clinical implications for future research
Based on a large paediatric population, we document EDN levels at 1 and 3 years representing expected levels in young boys and girls, which may be considered as reference values in similar populations.
Future studies in older children are however necessary to establish clinical implications of potential sex differences in eosinophil activation.Though higher EDN levels were observed across children with any wheeze, preschool asthma, AS and AD, their levels largely overlapped between the levels of non-atopic children.Furthermore, the low sensitivity and specificity of EDN levels at 1 year to predict the three clinical outcomes, indicate limited diagnostic ability of EDN, in line with our observations that approximately one of 10 with allergic disease had EDN levels above the ULN.No reliable biomarker for paediatric allergic diseases has yet been identified, suggesting the need of a comprehensive approach using a combination of diagnostic markers together with clinical observations and medical history. 39limited number of biomarkers for a Th2-driven inflammation are available, of which all face different challenges. 40,41However, as an easily obtained and stable marker of eosinophil inflammation, EDN may by itself or in combination with other biomarkers be a useful complement in clinical practice.

| CON CLUS ION
This study documented serum EDN levels, showing higher EDN levels in general at 1 year compared to at 3 years, also observed in non-atopic children, and in boys compared to girls.Compared to non-atopic children, higher EDN levels were observed at both ages in children with any wheeze, preschool asthma and preschool asthma with AS and/or AD at 3 years.Higher EDN levels at 1 and EDN levels at 1 year positively correlated to EDN levels at 3 years (n = 411), β-coefficient (95% CI) 0.62 (0.44, 0.81), R 2 = 0.23, (p < .001)and r (95% CI) 0.49 (0.40, 0.55) (p < .001),see Figure 3.
levels at 1 (n = 787) and 3 (A) (n = 857) years, and in children with information on EDN at both ages (B) (n = 411).From the bottom of the boxes, horizontal lines represent the second quartile, median and third quartile.Vertical lines below and above are the first and fourth quartile, respectively.The narrow tips represent EDN levels above the fourth quartile.† EDN levels at 1 year among all children with information on EDN at 1 year.‡ EDN levels at 3 years among all children with information on EDN at 3 years.§ EDN levels at 1 year among children with information at both 1 and 3 years.¶ EDN levels at 3 years among children with information at both 1 and 3 years.EDN, eosinophil-derived neurotoxin; max, maximum; μg/L, micrograms per litre; min, minimum.3 years; adjusted OR (95% CI) 2.80 (1.49, 5.28), 2.20 (1.09, 4.41) and 4.61 (1.60, 13.3), respectively (Figure S2A, Table Rydell et al.'s study of 213 web-recruited children, 9 median EDN levels were higher in children with concurrent asthma (74.1 μg/L) compared to the children with preschool asthma in our study (25.1 μg/L).Differences in age, populations and asthma definitions may explain these discrepancies.Several novel findings were observed in relation to EDN at 1 year and allergic diseases by 3 years of age; higher EDN levels at 1 year in children who at 3 years had asthma-like symptoms (any wheeze and preschool asthma), as well as children with AS and AD, compared to children who remained non-atopic by 3 years.Interestingly, sensitized children had among the highest EDN levels observed at both 1 and 3 years, pointing to the involvement of eosinophil inflammation.In relation to children with preschool asthma, children with preschool asthma and comorbidities of AS and/or AD had among the highest EDN levels at 1 year, which largely may be driven by sensitization, indicating a phenotype in which eosinophil inflammation may be established F I G U R E 3 Bland-Altman plot, displaying the difference of the paired EDN measurements at 1 and 3 years plotted against the mean of the two measurements (n = 411).The red dashed lines in the plot represents ± 2 SD.EDN, eosinophil-derived neurotoxin; μg/L, micrograms per litre.

F I G U R E 4
EDN levels at 1 (A) and 3 (B) years by atopic status at 3 years.From the bottom of the boxes, horizontal lines represent the second quartile, median and third quartile.Vertical lines below and above are the first and fourth quartile, respectively.The dots represent EDN levels above the fourth quartile.† Any wheeze at 3 years was defined as children with any episode of wheeze between 2 and 3 years, not fulfilling the preschool asthma criteria at 3 years.‡ Preschool asthma at 3 years was defined as having ≥3 episodes of recurrent bronchial obstruction (rBO) between 2 and 3 years, and fulfilling minimum one of the following criteria; (1) doctor diagnosed asthma between 0 and 3 years, and (2) any use of and/or prescribed asthma medications (bronchodilators, inhaled corticosteroids or leukotriene-antagonist) between 9 months and 3 years.§ Allergic sensitization at 3 years was defined as having positive SPT (≥3 mm) and/or specific IgE level (>0.35 kU A /L) towards any allergen (egg, cow's milk, peanut, wheat, soy, cod, birch, grass, mugwort, dog, cat, horse, house dust mite and/or mould) at 3 years.¶ Atopic dermatitis at 3 years was defined as fulfilling the United Kingdom Working Party and/or Hanifin and Rajka criteria at 3 years.† † Non-atopic children were defined as children with no history of asthma, wheeze, rBO episodes, doctor diagnosed asthma, use of and/or prescribed asthma medications, atopic dermatitis or allergic sensitization between birth and 3 years.EDN, eosinophil-derived neurotoxin; IgE, immunoglobulin E; max, maximum; min, minimum; SPT, skin prick test; μg/L, micrograms per litre.TA B L E 4Crude and adjusted logistic regression models of EDN at 1 and 3 years and any wheeze, preschool asthma, and preschool asthma with allergic sensitization and/or atopic dermatitis at 3 years.

3
years were associated with all three clinical outcomes at 3 years, but had limited diagnostic potential.AUTH O R CO NTR I B UTI O N SMartin Färdig contributed to conception and design of the study, data collection, data curation, analysis and interpretation of data, manuscript writing and editing.Anine Lie contributed to data collection, data curation and critically revised the manuscript.Magnus P. Borres contributed to conception of the study, and critically revised the manuscript.Tina Ekenkrantz contributed to data curation, and critically revised the manuscript.Berit Granum contributed to interpretation of data, and critically revised the manuscript.Guttorm Haugen contributed to interpretation of data, and critically revised the manuscript.Christine M. Jonassen contributed to interpretation of data, and critically revised the manuscript.Robert Movérare contributed to conception of the study, and critically revised the manuscript.Eva Maria Rehbinder contributed to data collection, interpretation of data and critically revised the manuscript.Håvard O. Skjerven contributed to conception and design of the study, interpretation of data and critically revised the manuscript.Anne Cathrine Staff contributed to conception and design of the Pre-ventADALL study, interpretation of data and critically revised the manuscript.Riyas Vettukattil contributed to data curation, and critically revised the manuscript.Karin C. Lødrup Carlsen contributed to conception and design of the study, interpretation of data, and critically revised the manuscript.Cilla Söderhäll contributed to conception and design of the study, interpretation of data and critically revised the manuscript.Björn Nordlund contributed to conception and design of the study, interpretation of data, and critically revised the manuscript.All listed authors approved the final version of the manuscript before submission and agreed to be accountable for all aspects of the work.ACK N O WLE D G E M ENTSWe sincerely thank all families participating in the Preventing Atopic Dermatitis and ALLergies (PreventADALL) study and the study personnel contributing to enrolling and managing the study: Hilde Aaneland, Anna Asarnoj, Ann Berglind, Jessica Björk, Oda C. Lødrup Carlsen, Åshild Wik Despriée, Kim M. A. Endre, Thea Aspelund Fatnes, Peder A. Granlund, Berit Granum, Malén Gudbrandsgard, Sandra Götberg, Gunilla Hedlin, Mari Rønning Kjendsli, Ina Kreyberg, Linn Landro, Caroline-Aleksi Olsson Mägi, Nora Nilsson, Monika Nordenbrand, Carina M. Saunders, Kajsa Sedergren, Natasha Sedergren, Päivi Söderman, Liv Julie Sørdal, Sandra Ganrud Tedner, Ellen Tegnerud, Magdalena R. Vaernesbranden and Johanna Wiik and in memoriam Kai-Håkon Carlsen.FU N D I N G I N FO R M ATI O NThe PreventADALL study has received funding from the following sources: South-Eastern Norway Regional Health Authority, The Norwegian Research Council, Oslo University Hospital, The University of Oslo, Health and Rehabilitation Norway, The Foundation for Healthcare and Allergy Research in Sweden -Vårdalstiftelsen,
EDN levels by sex in children with information at 1 (n = 787) and 3 (n = 857) years, as well as by atopic status at 3 years.
36 relation to tobacco smoke exposure, sex, heredity for asthma and wheezing, children with AS and AD had the highest likelihood of developing asthma by preschool-or school age.36However,EDNlevels were overall somewhat higher in children with any wheeze or preschool asthma compared to non-atopic children, suggesting eosinophil inflammation may also be involved in children with early asthma-like symptoms in the absence of AS and/or AD.TA B L E 3Abbreviations: EDN, eosinophil-derived neurotoxin; IgE, immunoglobulin E; na, not applicable; no., number; SPT, skin prick test; μg/L, micrograms per litre.