Exhaled nitric oxide in 4-year-old children: relationship with asthma and atopy

Study design

The Prevention and Incidence of Asthma and Mite Allergy study involved 4,146 children. Children were recruited through pre-natal clinics in three regions of the Netherlands (western, middle and northern areas). During pregnancy, their mothers completed a validated screening questionnaire on asthma and inhalant allergies 17, from which their atopic status was determined. Based on the atopy of the mother, children were labelled as high-risk (atopic mother) and low-risk (nonatopic mother).

Data on demographic factors, respiratory symptoms and risk factors for asthma were collected by yearly questionnaires. At age 4 yrs, all high-risk children (n = 1173) and a random sample of the low-risk children (n = 635) were invited for a medical examination, including measurement of F_E,NO. A detailed description of the study design is given elsewhere 18. The study protocol was approved by the medical ethics committees of the participating institutes. All parents gave written informed consent.

Study subjects

From the 1,808 children invited, medical examinations were performed in 1,279. In 344 of 1,279 children, it was not possible to obtain F_E,NO values, mostly because NO analysers were not available for the total medical examination period in the middle and northern part of the Netherlands. Thus, an attempt to measure F_E,NO was made in 935 children. Exhaled air was successfully collected in 659 children. From these 659 children, 230 children were excluded from statistical analysis because of missing questionnaires (n = 22), medication use in the 12 h before measurement (n = 22), high ambient NO (>20 ppb) during measurement (n = 114), technical problems during analysis of NO balloons (n = 62) or large differences (>10 ppb) between duplicate NO measurements (n = 10). Finally, the data of 429 children were available for statistical analysis. From these 429 children, 35 used inhaled corticosteroids at aged 4 yrs. A detailed flow chart of the study population is given in figure 1⇓.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Flow chart of the number of children progressing through the medical examination at age 4 yrs. F_E,NO: nitric oxide fraction in exhaled air.

F_E,NO measurement

F_E,NO was measured offline by the balloon method, according to recent European Respiratory Society/American Thoracic Society guidelines 19. In short, children were asked to take a deep breath through a charcoal NO scrubber, and to exhale immediately into a collection device employing dynamic flow restriction, using a two-way valve. Exhalation flow was kept constant at 50 mL·s⁻¹ over a pressure range of 5–20 cm H₂O. Mouth pressure was monitored during measurement using a manometer. After discarding dead space air for 3–4 s, exhaled air was collected in a NO impermeable 150 mL Mylar balloon (Jurjen de Vries BV, Leeuwarden, the Netherlands) 3. For every child, duplicate exhaled air samples and a sample of ambient air were collected. Balloons were sealed, stored and analysed within 8 h with a chemiluminescence analyser. In the western and middle areas of the Netherlands, a Sievers NOA 280 analyser (Boulder, CO, USA; sensitivity 0.5 ppb, detection range 0.5–500.000 ppb, sample flow 200 mL·min⁻¹, sampling rate 4·s⁻¹ (middle), 20·s⁻¹ (western)), was used. In the northern area, an Ecophysics CLD 700 AL analyser (Ecophysics, Basel, Switzerland; sensitivity 1 ppb, detection range 0–50.000 ppb, sample flow rate 700 mL·min⁻¹, data sampling rate 0.5·s⁻¹) was used. A check for systematic differences between the different NO analysers was performed by measuring balloons filled with calibration gas on all three analysers on the same day. This analysis revealed a good agreement, with no significant systematic differences between the analysers.

Asthma and wheezing phenotypes

Information about doctor's diagnosed asthma was collected by the following question: “Did a doctor ever diagnose asthma in your child?”. Symptoms of wheeze were assessed by International Study of Asthma and Allergies in Childhood core questions 20. Based on the longitudinal questionnaire data on wheeze symptoms in the first 4 yrs of life, children were divided into four wheezing phenotypes 21: “never wheeze”, “early transient wheeze” (at least one episode of wheeze in the first 3 yrs of life), “late-onset wheeze” (at least one episode of wheeze in the 4th yr of life) and “persistent wheeze” (at least one episode of wheeze in the first 3 yrs of life, and at least one episode of wheeze in the 4th yr of life). In the analyses, late-onset and persistent wheezing phenotypes were combined into one category.

Specific IgE and blood eosinophils

Total and specific immunoglobulin (Ig)E to inhalant allergens (house dust mite (Dermatophagoides pteronyssinus), cat, dog, grass (Dactylis glomerata), birch (Betula verrucosa) and mould (Alternaria alternata)) were determined by Radio Allergo Sorbent Test. Atopy was defined as specific IgE concentration >0.35 IU·mL⁻¹ for at least one inhalant allergen. The numbers of eosinophil granulocytes in peripheral blood were determined by automated differential cell count after peroxidase staining. Blood eosinophilia was arbitrarily defined as eosinophil counts >400 per µL 22. From the 429 children for whom reliable F_E,NO values were obtained, specific IgE and eosinophil counts were available for 247 and 234 children, respectively. The remaining children either refused or venepuncture failed. As eosinophil count followed a right-skewed distribution, data were log₁₀-transformed.

Statistical analysis

Because F_E,NO values followed a right-skewed distribution, data were log₁₀-transformed. Crude and adjusted geometric mean F_E,NO values were estimated by (multiple) linear regression analysis, with log₁₀ (F_E,NO) as the dependent variable and asthma, wheezing phenotype, atopy or eosinophilia as independent variables. Due to the stratified study design, all analyses were performed for the total study population as well as separately for high-risk and low-risk children. Recent symptoms of cold might influence F_E,NO and might be associated with asthma, therefore, these were considered as a potential confounder. To adjust for possible differences between study regions, region was taken into account. Atopy of the mother, sex, education of the mother, exposure to environmental tobacco smoke, exposure to pets, contact with other children and lower respiratory tract infections were also considered as potential confounders. The effect of inhaled corticosteroids was studied separately. To study the association of F_E,NO with eosinophil count without dividing the latter into categories, Pearson correlation coefficients were calculated. In the analyses, p-values <0.05 were considered to be statistically significant.

Selection of the study population

In total, 1,279 of the 1,808 children who were invited participated in the medical examination at 4 yrs of age. Children who did not participate (n = 529) more often had a mother with low education (32% versus 20%). Due to the relatively high percentage of children with missing questionnaires at 4 yrs of age (29% (154/529) versus 3% for the children who did participate), it was difficult to compare the percentages of children with respiratory symptoms between those who did and did not participate. However, with respect to respiratory symptoms at the ages of 1–3 yrs, there seemed to be no major differences between the two groups.

There were no differences in the percentage of young males, children with an atopic mother, children with low educated mothers, asthma, wheeze, atopy or blood eosinophilia between the children who participated in the medical examination, but were not included in the final study population (n = 850), and the final study population (n = 429). Within the study population, there were no differences between the children from whom blood samples were or were not available.

General characteristics of the study population

The study population consisted of 274 high-risk and 155 low-risk children (table 1⇓). High-risk children were more frequently living in the western area and less often in the middle area of the Netherlands and they less often had siblings than low-risk children. The percentage of low educated mothers was similar for high-risk and low-risk children. As compared with low-risk children, high-risk children more frequently reported doctor's diagnosed asthma (11% versus 6%; p = 0.08) and more frequently had specific IgE to inhalant allergens (25% versus 14%; p = 0.06) at the age of 4 yrs. These differences were of borderline statistical significance. The percentages of children with late-onset/persistent wheeze or peripheral blood eosinophilia were not different for high- and low-risk children (15% versus 14% and 30% versus 27%, respectively).

F_E,NO values in the total study population

The overall geometric mean F_E,NO value (95% confidence interval (CI)) was 8.0 ppb (7.7–8.3) (n = 429). In the children who did not have doctor's diagnosed asthma, never wheezed, did not have specific IgE to inhalant allergens or blood eosinophilia and did not use inhaled corticosteroids at 4 yrs of age (asymptomatic, nonatopic group), the geometric mean F_E,NO value (95% CI) was 7.9 ppb (7.1–8.8) (n = 64). F_E,NO values were slightly higher in children with doctor's diagnosed asthma and in children with specific IgE to at least one inhalant allergen as compared with those without (table 2⇓). Mean F_E,NO values were not associated with wheezing phenotype or peripheral blood eosinophilia. F_E,NO values were highest in children with both elevated specific IgE and doctor's diagnosed asthma, late-onset/persistent wheeze or peripheral blood eosinophilia. Because results were similar for the combinations of IgE with asthma, wheezing phenotype or eosinophilia, only the combination of IgE with asthma is shown in table 2⇓. There were no major differences between the results of the crude and adjusted analyses. Further adjustment for potential confounders did not alter the results of the analyses. There were also no major differences between the analyses in which children using inhaled corticosteroids at 4 yrs of age were included or excluded. When analyses were repeated for children for whom ambient NO levels were <10 ppb at the time of measurement, similar results were obtained.

The correlation between F_E,NO and eosinophil count (children using inhaled corticosteroids excluded) was 0.12 (r (Pearson), p = 0.07). The correlation was stronger for children with specific IgE (r = 0.34, p = 0.03) as compared with those without (r = 0.05, p = 0.52).

F_E,NO values in high-risk and low-risk children

In high-risk and low-risk children respectively, the overall geometric mean F_E,NO values (95% CI) were 8.0 ppb (7.6–8.5) and 7.9 (7.4–8.4). As was seen in the total study population, F_E,NO values were slightly higher in children with doctor's diagnosed asthma and in children with specific IgE to at least one inhalant allergen as compared with those without (table 3⇓, fig. 2⇓). In high-risk children, F_E,NO values were highest in children with both elevated specific IgE and doctor's diagnosed asthma, late-onset/persistent wheeze or blood eosinophilia. In low-risk children, high F_E,NO values were especially observed in children with elevated specific IgE, independent of asthma, wheezing phenotype or blood eosinophilia. However, the analyses of the combinations of specific IgE and asthma, wheezing phenotype or peripheral blood eosinophilia were limited by the low numbers of symptomatic children.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Distribution of nitric oxide levels in exhaled air (F_E,NO) in high-risk (○) and low-risk (•) children with (+) and without (−) elevated specific immunoglobulin (Ig)E to inhalant allergens. Children using inhaled corticosteroids were excluded. Bars represent geometric means.

The correlation between F_E,NO and eosinophil count was 0.09 (r Pearson, p = 0.29) for high-risk children and 0.20 (r Pearson, p = 0.08) for low-risk children. As was seen in the total study population, correlations were stronger for children with specific IgE (high-risk: r (Pearson) = 0.29, p = 0.13; low-risk: r = 0.64, p = 0.047) as compared with those without (high-risk: r = 0.03, p = 0.77; low-risk: r = 0.09, p = 0.48) (children using inhaled corticosteroids excluded).