Exhaled Nitric Oxide Predicts Persistence of Wheezing, Exacerbations, and Decline in Lung Function in Wheezy Infants and Toddlers

Abstract

Background

There are limited data assessing the predictive value of fraction of exhaled nitric oxide (FE_NO) for persistence of wheezing, exacerbations, or lung function change over time in infants/toddlers with recurrent wheezing.

Objectives

In an ongoing longitudinal cohort of infants and toddlers with recurrent wheezing compare predictive values of single-breath FE_NO (SB-FE_NO), tidal-breathing mixed-expired FE_NO (tidal-FE_NO), bronchodilator responsiveness (BDR), and the Castro-Rodriquez asthma predictive index (API) for persistence of wheezing, exacerbations, and lung function change through age 3 yrs.

Methods

Enrollment forced expiratory flows and volumes (iPFTs) were measured in 44 infants/toddlers using the raised-volume rapid thoracoabdominal compression method. SB-FE_NO was measured at 50 mL/sec, and tidal-FE_NO was measured during awake tidal breathing. Clinical outcomes were assessed at age 3 yrs. in 42 infants. Follow-up iPFTs were completed between ages 2.5-3 yrs. in 32 subjects.

Results

An enrollment SB-FE_NO concentration ≥30 ppb predicted persistence of wheezing at age 3 years with a sensitivity of 77%, a specificity of 94%, and an area under the curve (AUC) of 0.86 (95% CI: 0.74 – 0.98). The sensitivity, specificity, positive predictive, and negative predictive values of SB-FE_NO for persistence of wheezing and exacerbations were superior to tidal-FE_NO, BDR, and the API. SB-FE_NO ≥30 ppb and tidal FE_NO ≥7 ppb measured at enrollment was associated with a decline in both FEV_0.5 and FEF_25-75 between enrollment and age 3 years.

Conclusions

In wheezy infants/toddlers SB-FE_NO was superior to tidal-FE_NO, BDR, and the API in predicting future exacerbations and persistence of wheezing at age 3 years. Both SB-FE_NO and tidal FE_NO were associated with lung function decline over time.

INTRODUCTION

More than half of school age children with persistent asthma were symptomatic during their preschool years, usually presenting with recurrent episodes of wheezing.[1] Recurrent wheezing affects up to 30% of infants and toddlers, yet resolves in at least 50% of these children by school age. An accurate predictor of asthma does not currently exist for infants/toddlers; hampering decision making as clinicians generally cannot differentiate infants/toddlers with transient wheezing from those with persistent asthma. NHLBI guidelines currently recommend daily use of inhaled corticosteroids (ICS) for children under age 5 yrs. with > 3 prior episodes of wheezing and epidemiologic risk factors for asthma.[2] These guidelines are based primarily on the Asthma Predictive Index (API) proposed by Castro-Rodriquez et al[3]. However, the API has not been assessed prospectively in unselected infants/toddlers with recurrent wheeze. When applied to the birth cohort from which it was developed[3], and to a large independent population-based cohort,[4] the API had a high negative predictive value yet a sensitivity of only 28-37% and positive predictive value of only 40-48% for asthma at age 6 yrs.

Many parents are reluctant to treat their children with inhaled steroids if the diagnosis of asthma is uncertain. Concern about treating infants/toddlers with inhaled or systemic steroids has further increased with recent trials suggesting that treatment of viral-induced wheezing with oral steroids in preschool children is ineffective [5], and that prevention of such attacks with inhaled steroids may be associated with reduced growth velocity.[6] A biomarker with a high positive predictive value for atopic asthma among infants and toddlers with recurrent wheezing would be useful to investigators testing potential asthma controller therapies in very young children by allowing for more targeted enrollment of young asthmatics in clinical trials, and helpful to clinicians as an adjunctive predictor of patients likely to benefit from current and future asthma controller regimens. The fractional concentration of exhaled nitric oxide (FE_NO) is a biomarker of airway inflammation repeatedly shown to be elevated in adults and children with allergic asthma[7], yet there is a paucity of data assessing FE_NO prospectively among wheezy infants/toddlers.[8-12] In adults and school-age children with atopic asthma FE_NO is correlated with sputum and bronchial eosinophils, peak flow variability, and bronchial reactivity[13-18], and decreases following systemic or inhaled steroid treatment.[19-21] In atopic infants at risk for asthma, FE_NO was associated with airway reactivity.[22] Furthermore, among adults with chronic respiratory symptoms FE_NO has been shown to be a superior predictor of steroid responsiveness than spirometry, BDR, or airway reactivity to methacholine or adenosine.[23] ATS FE_NO Clinical Practice Guidelines [24] recommend use of FE_NO to diagnose eosinophilic airway inflammation, determine likelihood of steroid responsiveness, and support the diagnosis of asthma when objective evidence is lacking. Data supporting these recommendations is most robust in steroidnaïve eosinophilic asthmatics. [24] Most wheezy infants/toddlers who go on to develop persistent childhood asthma can be characterized as having an atopic phenotype of asthma[25]; therefore FE_NO may be well suited as a biomarker to separate infants/toddlers with asthma from those with wheezing due to other etiologies.

In one recent study of early school-age and preschool children, FE_NO measured during tidal breathing was modestly associated with a greater risk of respiratory illness over one year.[26] However, the mean enrollment age was >3 years, and most children < 3yrs. could not perform acceptable tidal breathing FE_NO measurements, indicating that unsedated tidal-breathing FE_NO may be technically challenging to measure in infants/toddlers.[26] Additional limitations of tidal breathing-FE_NO in infants include: potential nasal NO contamination as nasal NO production far exceeds lower airway production, and variable expiratory flow which is problematic due the highly flow dependent nature of FE_NO[27, 28]. Despite these potential limitations, it is easier to measure FE_NO in infants/toddlers during tidal breathing than to measure FE_NO using a flow-regulated exhalation method.

Previously we reported that single-breath flow-regulated FE_NO (SB-FE_NO) was predictive of subsequent wheezing, exacerbations, and lung function decline in a cohort of wheezy infants and toddlers (mean enrollment age of 15.8 mos.) over a 6 month follow-up period.[29] We now report follow-up of this cohort through age 3 years. Our objective was to compare enrollment SB-FE_NO, mixed expired FE_NO measured during tidal breathing, and the Asthma Predictive Index (API)[3] as predictors of persistence of wheezing at age 3 years, exacerbations of wheezing during follow-up, and changes in lung function. We hypothesized that FE_NO would be a superior to the Asthma Predictive Index (API)[3] as a predictor of these outcomes. Some of these results have been reported in abstract form.[30]

METHODS

Subjects

Children 6-24 months of age with ≥ 3 episodes of physician-diagnosed wheezing treated with bronchodilators or corticosteroids were recruited for a single-center longitudinal study. Subjects with birth < 36 weeks gestation, congenital heart disease, dysphagia, severe gastroesophageal reflux, or upper airway obstruction were excluded. Systemic or inhaled corticosteroid treatment was not permitted for at least 3 weeks prior to baseline lung function and FE_NO measurement. Thereafter, corticosteroid treatment was at the discretion of the subject’s primary care or emergency department provider who did not have access to FE_NO data. Written consent was obtained from parents of subjects. The study was approved by the Seattle Children’s Hospital Institutional Review Board.

Study Visits

Study visits occurred a minimum of 3 weeks following resolution of an upper or lower respiratory infection, or acute exacerbation of wheezing. At enrollment medical history was reviewed, length and weight were measured with a calibrated stadiometer and digital scale, SB-FE_NO and tidal-FE_NO was measured, and lung function testing performed. Between enrollment and follow-up visits parents completed monthly symptom and medication diaries that recorded: wheezing episodes lasting ≥ 3 days treated with at least a bronchodilator, use of systemic (oral or intravenous) corticosteroids, and use of inhaled corticosteroids. Diaries were returned to study personnel on a monthly basis. Interval medical history and medication use was reviewed and lung function tests repeated at a follow-up visit 6 months after enrollment and at age 3 years.

Lung Function Measurements

Forced expiratory volumes and flows (FVC, FEV_0.5, and FEF_25-75) were measured using the RVRTC technique according to ATS/ERS guidelines for raised volume forced expirations in infants[31] via the InSpire® Infant Pulmonary Lab (IPL) at enrollment, 6 months post enrollment, and age 3 years. Lung function tests were performed at least 3 weeks after resolution of acute exacerbations of wheezing or respiratory illnesses. Lung function parameters were analyzed as z-scores calculated from published normative data. [32] Measurement of forced expiratory flows and volumes were repeated 10 minutes after administration of 4 puffs of albuterol. Bronchodilator responsiveness (BDR) was defined as ≥ 12% improvement in FEV_0.5, or ≥ 25% improvement in FEF_25-75.

Exhaled Nitric Oxide Measurements

At each testing session, at least 3 flow-regulated SB-FE_NO measurements were performed in each subject immediately prior to lung function testing using a method we have previously described.[29] Mixed expired FE_NO measurements during tidal breathing (tidal-FE_NO) were obtained prior to administration of chloral hydrate while subjects were awake and seated in parent’s lap watching a video for distraction. Mixed expired collections were obtained over 30 seconds of quiet regular breathing with a full facemask over the mouth and nose following 5 breaths to washout circuit dead space. Nitric oxide free air (<2 ppb) was provided during inspiration via a filter attached to the collection circuit with a one-way valve. Nitric oxide was measured using a Sievers® NOA 280 chemiluminescence analyzer (GE Analytical Instruments; Boulder, CO, USA).

Asthma Predictive Index (API)

Using data collected at enrollment a positive or negative API was determined for each subject using criteria modified slightly from criteria proposed by Castro-Rodriquez et al. [3] A positive API was defined as meeting 1 of 2 major, or 2 minor criteria. Major Criteria included a history of parental physician-diagnosed asthma or a history of physician-diagnosed eczema in the subject. Minor Criteria included a history of physician-diagnosed allergic rhinitis or a history of wheezing apart from colds.

Statistical Analysis

Forty subjects were estimated to provide 80% power to identify SB-FE_NO as a significant predictor of FEV_0.5 if SB-FE accounted for ≥15% of the total variance in FEV_0.5 (i.e. r² NO =0.15). SB-FE_NO and lung function data were determined to be normally distributed by the D’Agostino and Pearson as well as Kolmogorov-Smirnov tests. Summary statistics, Pearson’s correlation coefficients and independent t-tests were used to describe and differentiate subjects by their wheeze and exacerbation status at 3 years of age.

Multivariate logistic regression was used to assess associations between enrollment FE_NO measures (SB-FE_NO and tidal-FE_NO) and: 1) acute exacerbations of wheezing (defined as ≥ 1 episode of wheezing lasting ≥ 3 days treated with both albuterol and systemic corticosteroids) during the 6 months prior to the age 3 year follow-up visit, and 2) persistence of wheezing at age 3 years (defined as ≥1 episode of any wheezing treated with albuterol during the 6 months prior to age 3 years). Persistence of wheezing was further characterized by parental report into multi-triggered persistent wheezing if wheezing was triggered by more than viral respiratory tract infections (e.g. vigorous play/exercise, food, pets, or other inhalant exposures per parent observation), or URI-triggered persistent wheezing if triggered exclusively by viral upper respiratory tract infections. Enrollment age, gender, family history of asthma, eczema, sustained inhaled corticosteroid treatment, and tobacco smoke exposure were included as covariates in the logistic models. Receiver operating characteristics (ROC) analyses were performed to determine the enrollment SB-FE_NO and tidal-FE_NO values that best predicted persistence of wheezing at age 3 years as well as exacerbations of wheezing. Using the best cutoff points for SB-FE_NO and tidal-FE_NO, the sensitivity, specificity, positive predictive value, and negative predictive value of enrollment SB-FE_NO and tidal-FE_NO, bronchodilator responsiveness at enrollment, a positive API, and SB-FE_NO combined with a positive API, were calculated as predictors of persistence of wheezing and exacerbations of wheezing.

Generalized estimating equations (GEE) regression models were used to assess associations between enrollment SB-FE_NO and tidal FE_NO, baseline subject characteristics, and changes in lung function during follow-up utilizing RVRTC lung function data from each visit at which it was obtained (enrollment, 6 month follow-up, and age 3 years). An ROC analysis was performed to determine the enrollment SB-FE_NO and tidal-FE_NO concentrations that best predicted a decline in FEV_0.5 or FEF_25-75. All reported p-values are two sided. Data analyses were conducted using GraphPad® Prism 4.0 (La Jolla, CA, USA), Intercooled Stata for Windows® version 10.1 (College Station, Texas, USA), and SAS version 9.2 (Cary, NC, USA).

RESULTS

Forty-seven infants and toddlers were recruited with a mean age of 15.6 (SD ± 5.2) months at enrollment. At enrollment technically acceptable RVRTC lung function tests were performed in 44 subjects, and FE_NO measurements obtained in 45 subjects. Baseline characteristics of the cohort are presented in Table 1. Forty subjects completed clinical follow-up 6 months following enrollment, of which 32 completed RVRTC lung function testing. At age 3 years, 42 subjects completed clinical follow-up (95% of subjects who completed enrollment lung function testing), and 34 subjects completed RVRTC lung function measurements (77% of subjects who completed enrollment testing). The mean duration between subject enrollment visits and follow-up at age 3 years was 19 months.

TABLE 1.

Baseline Characteristics of Subjects Completing Enrollment FENO Measurement

Subjects (n)	45
Gender (male)	30 (67%)
Race (Caucasian)	33 (73%)
Mean age at enrollment (months)*	15.7 ± 5.2
Eczema (physician-diagnosed and steroid treated)	16 (36%)
Family history asthma (1st degree relative)	24 (53%)
Enviromental tobacco smoke exposure	12 (27%)
Prior treatment with systemic steroids for wheezing	34 (76%)
History of ED visit for wheezing	34 (76%)
History of hospitalization for wheezing	16 (36%)
History of wheezing apart from viral infections	16 (36%)
Enrollment SB-FENO (ppb)*	28.7 ppb ± 16.7
Enrollment tidal-FENO (ppb)*	7.8 ppb ± 4.9
Enrollment FVC (z-score)**	−0.04 (−0.42 to 0.35)
Enrollment FEV0.5 (z-score)**	−0.42 (−0.78 to −0.07)
Enrollment FEF25-75 (z-score)**	−0.67 (−1.10 to −0.25)

Definition of Abbreviations: ED = emergency department; SB-FE_NO = single-breath exhaled nitric oxide; Tidal-FE_NO = mixed expired exhaled nitric oxide during tidal breathing; FVC = forced vital capacity; FEV_0.5 = forced expiratory volume in 0.5 seconds; FEF_25-75 = forced expiratory flow between 25% and 75% of forced expiratory maneuver.

^*Data are presented as the mean ± SD

^**Data are presented as the mean ± 95% CI; n=44 subjects

The mean FEV_0.5 and FEF_25-75 for the cohort at enrollment were significantly less than zero (a z-score equal to zero represents the mean value of published normative data). [29, 32] Enrollment of subjects was evenly distributed across seasons.[29] Eight subjects used inhaled corticosteroids on a sustained basis between the enrollment and 6 month follow-up visits, and 9 of the 34 subjects who completed lung function testing at age 3 years used inhaled steroids on a sustained basis during the 6 months prior to the visit at age 3 years. We previously reported that SB-FE_NO measured at enrollment was associated with both age and a history of eczema, but was not associated with gender, family history of asthma, environmental tobacco smoke exposure, wheezing apart from viral respiratory infections, enrollment lung function measures, hospitalization for wheezing, or oral steroid treatment prior to enrollment.[29] At enrollment the mean SB-FE_NO for the cohort was 28.7 ppb (s.d. 16.7) and mean awake tidal-FE_NO was 7.8 ppb (s.d. 4.9). SB-FE_NO and awake tidal-FE_NO measurements at enrollment were weakly correlated (r=0.33, p=0.04; Figure 1, panel A.), however, at the final visit during which SB-FE_NO was measured prior to age 3 years, there was no longer a significant correlation between SB-FE_NO and tidal-FE_NO measurements (r=0.3, p=0.1; Figure 1, panel B.).

FIGURE 1.

Panel A: correlation between enrollment single-breath exhaled nitric oxide (SB-FE_NO) concentrations and mixed-expired exhaled nitric oxide concentrations collected during tidal breathing (tidal-FE_NO) (r=0.33, p=0.04). Panel B: correlation between SB-FE_NO and tidal-FE_NO concentrations collected at age 3 years (r=0.3, p=0.1). Correlations determined using the Pearson correlation coefficient.

Twenty-five subjects had persistence of wheezing at age 3 years, and during the 6 months prior to age 3 years, 19 subjects experienced an acute exacerbation of wheezing. Enrollment SB-FE_NO concentrations were significantly higher among subjects who experienced persistence of wheezing at age 3 years than among those subjects who did not (mean 35.1 ppb vs. 18.6 ppb, p < .001; Figure 2, panel A.). The odds of persistence of wheezing was three times higher (OR 3.1; 95% CI 1.5 – 6.4) for every 10 ppb increase in enrollment SB-FE_NO (Table II). Enrollment SB-FE_NO concentrations were also significantly higher among subjects who experienced acute exacerbations of wheezing than among those subjects who did not (37.4 ppb vs. 22.1 ppb, p = .001; Figure 3, panel A.). The odds of an acute exacerbation were 2.6 times higher (OR 2.6; 95% CI 1.4 – 5.2) for every 10ppb increase in enrollment SB-FE_NO (Table 2.). Enrollment tidal-FE_NO concentrations were not significantly different between children with persistence of wheezing and among those whose wheezing resolved (8.4 ppb vs. 7.1 ppb, p = 0.4: Figure 2, panel B.), or between those who experienced an acute exacerbation and those that did not (8.4 ppb vs. 7.2 ppb, p=0.4; Figure 3, panel B.).

FIGURE 2.

Panel A: enrollment single-breath exhaled nitric oxide (SB-FE_NO) concentrations were significantly higher among subjects who experienced persistence of wheezing (closed triangles) at age 3 years than among those subjects who experience resolution (open circles) of wheezing (35.1 ppb vs. 18.6 ppb, p < .001). Analysis conducted using an unpaired t test. Panel B: enrollment tidal-FE_NO concentrations were not significantly different between subjects with persistence of wheezing (closed triangles) and among those whose wheezing resolved (open circles) (8.4 ppb vs. 7.1 ppb, p = 0.4). Analysis conducted using the Mann-Whitney test.

TABLE 2.

Associations between Enrollment SB-FENO and Persistence of Wheezing and Exacerbations of Wheezing

	Persistence of Wheezing		Exacerbations of Wheezing
	OR	95% CI	OR	95% CI
Enrollment SB-FENO (10 ppb increments)	3.1	1.5 – 6.4	2.6	1.4 – 5.2
Age at enrollment (mos.)	1.0	0.8 – 1.1	1.0	0.8 – 1.2
Gender (female)	0.9	0.2 – 4.8	1.5	0.3 – 7.5
FH Asthma (yes)	3.7	0.6 – 21.5	2.4	0.5 – 12.2
Eczema (yes)	0.4	0.1 – 2.0	0.5	0.1 – 2.8
Sustained ICS use during follow-up (yes)	5.3	0.2 - 140	9.3	0.5 - 186
Tobacco Smoke Exposure (yes)	2.3	0.3 – 16.8	2.3	0.3 – 16.8

Definition of Abbreviations: SB-FE_NO = single-breath exhaled nitric oxide; ICS = inhaled corticosteroids; FH = family history; OR = odds ratio; CI = confidence interval. N=42 subjects. Bold type denotes a statistically significant association.

FIGURE 3.

Panel A: enrollment single-breath exhaled nitric oxide (SB-FE_NO) concentrations were significantly higher among subjects who experienced an acute exacerbation of wheezing (closed triangles) between ages 2.5 and 3 years than among those subjects who did not experience an exacerbation (open circles) of wheezing (37.4 ppb vs. 22.1 ppb, p = .001). Analysis conducted using an unpaired t test. Panel B: enrollment tidal-FE_NO concentrations were not significantly different between subjects who experienced an acute exacerbation (closed triangles) and those that did not (open circles) (8.4 ppb vs. 7.2 ppb, p=0.4). Analysis conducted using the Mann-Whitney test.

An ROC analysis showed that an enrollment SB-FE_NO concentration > 30 ppb predicted persistence of wheezing (regardless of trigger) with a sensitivity of 77%, a specificity of 94%, and an area under the curve (AUC) of 0.86 (95% CI: 0.74 – 0.98) (Figure 4; Table 3). A SB-FE_NO concentration > 30 ppb predicted the prevalence of multi-trigger persistent wheezing prior to age 3 years (n=15 subjects) with a sensitivity of 93%, a specificity of 75%, a positive predictive value of 67%, and negative predictive value of 95%. In contrast, a SB-FE_NO concentration > 30 ppb predicted the prevalence of URI-triggered persistent wheezing (n=12 subjects) with a sensitivity of 58%, a specificity of 59%, a positive predictive value of 37%, and negative predictive value of 77%. A SB-FE_NO concentration > 30 ppb predicted acute exacerbations of wheezing with a sensitivity of 84%, a specificity of 78%, and an area under the curve (AUC) of 0.83 (95% CI: 0.7 – 0.96) (Figure 4). The sensitivity, specificity, positive predictive value, and negative predictive values of enrollment SB-FE_NO for persistence of wheezing and acute exacerbations were superior to enrollment tidal-FE_NO, bronchodilator responsiveness, or a positive API (Table 3.). The combination of a positive API and a SB-FE_NO concentration > 30 ppb yielded a modestly improved specificity for predicting persistence of wheezing and exacerbations, however, the sensitivity was much lower than for SB-FE_NO or the API alone (Table 3).

FIGURE 4.

Single-breath exhaled nitric oxide (SB-FE_NO) concentrations measured at age 3 years were significantly greater among subjects with persistence of wheezing at age 3 years (closed triangles) as compared to subjects with resolution of wheezing (open circles), both among all subjects (panel A.; 30.8 ppb vs. 20.7 ppb, p=0.048) and when subjects using inhaled corticosteroids on a sustained basis were excluded (panel B.; 33.5 ppb vs. 20.7 ppb, p=0.036). SB-FE_NO concentrations were also significantly higher among subjects who experienced exacerbations prior to age 3 years (closed triangles) then among subjects who did not experience exacerbations prior to age 3 years (open circles), both among all subjects (panel C.; 32.9 ppb vs. 22.6 ppb, p=0.04) and when subjects using inhaled corticosteroids on a sustained basis were excluded (panel D.; 36.9 ppb vs. 22.2 ppb, p=0.036). All analyses conducted using unpaired t tests.

TABLE 3.

Sensitivity, Specificity, PPV, and NPV of Enrollment SB-FE_NO, Bronchodilator Responsiveness, and the Asthma Predictive Index for Persistence of Wheezing and Acute Exacerbations of Wheezing

Persistence of Wheezing at Age 3 Years
Predictor	Sensitivity	Specificity	PPV	NPV
SB-FENO (>30ppb)	77	94	95	73
Tidal-FENO (>6.2ppb)	48	56	61	43
BDR	32	91	85	43
API	46	63	67	42
SB-eNO (>30ppb) + positive API	33	100	100	45
Acute Exacerbations of Wheezing between Ages 2.5 and 3 Years
Predictor	Sensitivity	Specificity	PPV	NPV
SB-FENO (>30ppb)	84	78	76	86
Tidal-FENO (>6.2ppb)	53	61	52	60
BDR	38	88	71	65
API	47	61	50	58
SB-eNO (>30ppb) + positive API	37	91	78	64

Definition of Abbreviations: PPV = positive predictive value; NPV = negative predictive value; SB-FE_NO = single-breath exhaled nitric oxide; Tidal-FE_NO = mixed expired exhaled nitric oxide during tidal breathing; BDR = bronchodilator responsive; API = Asthma Predictive Index (modified) defined as 1 of 2 major, or 2 minor criteria (major criteria = parental physician diagnosed asthma, or physician-diagnosed and treated eczema; minor criteria = physician diagnosed atopic dermatitis, or wheezing apart from colds).

Increased SB-FE_NO concentration at enrollment was correlated with a decline in both FEV_0.5 (r = −0.67, p = 0.01) and FEF_25-75 (r = −0.4, p = 0.02) between enrollment and age 3 years. An ROC analysis showed that an enrollment SB-FE_NO concentration ≥ 30 ppb predicted a decline in FEV_0.5 between enrollment and age 3 years with a sensitivity of 74% and a specificity of 87%, with an area under the curve (AUC) of 0.86 (95% CI: 0.77 , 0.99). Using GEE modeling to analyze lung function data from all three study time points and adjust for age at enrollment, gender, family history of asthma, and sustained use of inhaled corticosteroids during follow-up, we found that an enrollment SB-FE_NO ≥ 30ppb was associated with a −1.55 z-score decline in FEV_0.5 from baseline to age 3 (95% CI: −2.18, −0.93), and a −1.01 z-score decline in FEF_25-75 (95% CI: −1.74, −0.28) compared to no change or increase in those with enrollment SB-FE_NO <30 ppb (Table 4.). There was a significant interaction effect of baseline SB-FE_NO by FEV₀₅ change (p = 0.013) and FEF_25-75 change at 3 years (p = 0.032). Sustained use of inhaled corticosteroids was associated with an overall higher FEV_0.5 (0.52 z-score, 95% CI: −0.02, 1.06) and FEF_25-75 z-score (0.61, 95% CI: 0.15, 1.08) whereas female gender was associated with a lower lung function (FEV₀₅: −0.68, 95% CI: −1.16, −0.22 and FEF_25-75: −0.98, 95% CI: −1.51, −0.46). Tidal-FE_NO ≥ 7 ppb was associated with a −1.44 z-score decline in FEV₀₅ (95% CI: −2.08, −0.81) and a −0.91 decline in FEF_25-75 z-score (95% CI: −1.69, −0.12); again the change in FEV₀₅ (p = 0.014) and FEF_25-75 was significantly different for the two levels of baseline tidal FE_NO. While not statistically significant, baseline lung function z-scores were slightly higher in those with high SB-FE_NO or tidal FE_NO than the others; an analyses adjusting for baseline FEV₀₅ and FEF_25-75 did not alter the conclusions or substantially alter the change estimates. At the final visit during which SB-FE_NO was measured prior to age 3 years, SB-FE_NO was significantly higher among subjects with BDR than among those without BDR (mean 39.2 ppb +/− s.d. 14.6 vs. 27.9 ppb +/− s.d. 13; p = 0.05). In contrast, tidal-FE_NO was similar among subjects with and without BDR (mean 11.9 ppb +/− s.d. 9.5 vs. 8.5 ppb +/− s.d. 9.6; p = 0.4).

Table 4.

Associations between Enrollment SB-FENO and tidal FENO and Change in LungFunction through Age 3 Years, Adjusted for Gender, Age, FH Asthma, and Sustained ICS use in Follow-up.

	FEV0.5		FEF25-75
	ß Coefficient (Δ FEV0.5 z-scores)	P value	ß Coefficient (Δ FEF25-75 z-score)	P value
Enrollment SB-FENO (<30 ppb)	−0.02	0.94	0.95	.009
Enrollment SB-FENO (≥30 ppb)	− 1.55	<0.0001	− 1.01	0.006
Enrollment tidal FENO (<7 ppb)	0.02	0.9	1.13	0.0007
Enrollment tidal FENO (≥7 ppb)	− 1.44	<0.0001	− 0.91	0.02

Definition of Abbreviations: SB-FE_NO = single-breath exhaled nitric oxide; ICS = inhaled corticosteroids; FH = family history; Δ = change between enrollment and age 3 years, FEV_0.5 = forced expiratory volume in 0.5 seconds; FEF_25-75 = forced expiratory flow between 25% and 75% of forced expiratory maneuver. N=34 subjects. Associations determined using generalized estimating equation (GEE) regression modeling. P value for test of parameter estimate (β coefficient) differing from zero. Bold type denotes a statistically significant association.

After excluding subjects who used inhaled corticosteroids on a sustained basis during the final 6 months prior to age 3 years (n=9), there was a significant direct correlation between SB-FE_NO concentrations measured at enrollment and age 3 years (r = 0.6, P = 0.004). SB-FE_NO concentrations at enrollment were not significantly different from concentrations at age 3 years (mean 28.7 vs. 31.2 ppb, p = 0.2).

DISCUSSION

In a cohort of infants and toddlers (mean age 15.6 months) with a history of recurrent episodes of physician-diagnosed and treated wheezing we found that SB-FE_NO measured at enrollment was associated with the persistence of wheezing at age 3 years. Furthermore, higher SB-FE_NO concentrations measured at study entry were associated with an increased risk of exacerbations prior to age 3 years. SB-FE_NO concentration at enrollment was superior to both the Asthma Predictive Index and the presence of BDR as predictors of the persistence of wheezing at age 3 years and acute exacerbations during the 6 months prior to age 3 years. Both SB-FE_NO and tidal-FE_NO measurements at enrollment were associated with a decline in lung function between enrollment and age 3 years. However, in contrast to flow-controlled SB-FE_NO measurements, we found no association between tidal-FE_NO measurements at enrollment and persistence of wheezing at age 3 years or exacerbations of wheezing prior to age 3 years. Finally, SB-FE_NO, but not tidal-FE_NO, measured prior to age 3 years was associated with BDR.

Our SB-FE_NO results are consistent with a recent report from a prospective cohort of infants with atopic dermatitis but without prior wheezing wherein SB-FE_NO measured at enrollment (mean age 10.7 months) was predictive of asthma as well as airway reactivity at age 5 years.[33] However, to our knowledge, our study is the first to perform SB-FE_NO measurements and obtain longitudinal lung function measurements in a cohort of infants and toddlers with recurrent physician-diagnosed wheezing. Although our finding that SB-FE_NO is associated with persistence of wheezing at age 3 years is consistent with our hypothesis, follow-up of this cohort into the school-age years will be essential to determine if SB-FE_NO robustly predicts which wheezy infants will develop atopic asthma and which will experience resolution of wheezing.

Our observed lack of an association between tidal-FE_NO measured at enrollment and persistence of wheezing at age 3 years or exacerbations of wheezing is consistent with other recent studies of FE_NO measured during tidal breathing in young children with recurrent wheezing.[34] [35] There are several potential explanations for the poor predictive value of FE_NO measured during tidal breathing. First, there is potential for contamination of sampled lower airway NO by nasal NO when measurements are performed during tidal breathing given that the soft palate is not closed and nasal NO concentrations are frequently at least 100-fold greater than lower airway NO concentrations.[36] Contamination of measurements by nasal NO therefore would tend to obscure inter-individual differences in lower airway FE_NO and bias towards a lack of association between tidal-FE_NO and our clinical outcomes. Secondly, expiratory flow is often variable during tidal breathing measurements conducted in awake young children. Variable expiratory flow is problematic due the highly flow dependent nature of FE_NO.[27, 28]

Although the prevalence of recurrent wheezing among infants/toddlers is high, this group of children is heterogeneous. The majority of wheezy infants/toddlers do not have asthma, but rather wheeze for a variety of anatomic and/or pathophysiologic reasons, including but not limited to post-viral airway inflammation and injury, small airway caliber, central airway malacia, or dysphagia. For most of these young children episodic wheezing resolves by school age. However, discriminating between infants/toddlers with transient wheezing and young children with evolving asthma is a major challenge for clinicians. Castro-Rodriquez et al proposed the Asthma Predictive Index, derived from epidemiologic asthma risk factors, to assist in predicting asthma in wheezy infants/toddlers.[3] While the API has a high negative predictive value for asthma when applied to recurrently wheezy infants/toddlers, the clinical utility of the API is limited by its poor positive predictive value.[4] Our data demonstrating that the predictive value of SB-FE_NO is markedly superior to the API through age 3 years suggests that this biomarker holds promise as an adjunctive diagnostic test for asthma in wheezy infants and toddlers. Diagnosing asthma at an early age using an objective test such as FE_NO could greatly facilitate more targeted use of asthma therapies, prevent under- and over-treatment with corticosteroids, and reduce preventable emergency department visits and hospitalization for asthma exacerbations. However, the lack of an association between tidal-FE_NO and persistence of wheezing or acute exacerbations dampens enthusiasm for FE_NO as practical biomarker that could be used clinically on a large scale given that SB-FE_NO measurements require that children undergo sedation and measurements require specialized equipment as well as technical expertise to perform. Alternatively, SB-FE_NO measurements may have utility in future clinical trials involving infants and toddlers with recurrent wheezing.

There are several limitations to this study. First, although we obtained clinical follow-up through age 3 years in 42 of the 45 subjects (93%) who completed enrollment FE_NO measurements, we were unable to obtain infant lung function data at age 3 years in 28% of subjects. Although, a higher percentage of subjects completing follow-up measurements would be preferable, the failure rate of sedation with chloral hydrate increases with age.[37] Chloral hydrate is used to sedate infants for lung function testing by nearly all centers using the RVRTC method, and our overall success rate is comparable to other centers.[38] [39] Second, we measured SB-FE_NO at an expiratory flow of 50 mL/sec. to be consistent with ATS recommendations for FE_NO measurements in older children and adults.[36] A positive association between age and FE_NO has been reported in several large studies of healthy school age children.[40] [41] [42] Possible mechanisms for the age dependency of FE_NO include a higher relative flow in young subjects with smaller airways compared to older subjects with larger airways, changes in NO diffusion coefficients, changes in nitric oxide synthase expression with age, or changes in S-nitrosoglutathione (GSNO) kinetics. [43] Measurement at 50 mL/sec may not be the ideal flow to maximize the predictive value of this biomarker, therefore future studies in this age group should assess FE_NO at several different flows. Variable patterns of inhaled corticosteroid use during follow-up limited our ability to assess associations between change in FE_NO over time and longitudinal lung function and clinical outcomes. Finally, the normative data used to calculate lung function z-scores was obtained from children ranging in age from 1 month to 2.8 years, with a mean age of 1 year [32], while the final lung function measurements in our cohort were obtained near age 3 years. Although the normative data used in our analysis could potentially over- or under-estimate normal flows for children near age 3 years, given that the same reference data was applied to all subjects any potential effect on our results should be non-differential.

An important limitation of this study is that we did not perform skin or radioallergosorbent testing of subjects to food or inhalant allergens at enrollment. Kulig et demonstrated in a birth cohort that development of allergic sensitization to inhalant allergens occurs mostly beyond age 2 years[44], however, ideally we would have objective data characterizing the sensitization status of our subjects at enrollment. Future studies evaluating the predictive value of FE_NO must include prospective serial measures of allergic sensitization.

Subjects were not receiving any corticosteroid treatment when enrollment SB-FE_NO concentrations were measured. Although by study design corticosteroids did not bias enrollment SB-FE_NO concentrations, 9 subjects in our study did use inhaled steroids on a sustained basis during follow-up, and therefore it is possible that inhaled steroid use prior to age 3 years could decrease the incidence of wheezing exacerbations or persistence of wheezing. This effect could have biased our results toward decreased specificity and positive predictive values of enrollment SB-FE_NO. Another limitation of this study is an insufficient sample size to allow adequate statistical power to determine the sensitivity, specificity and predictive values of FE_NO for persistence of wheezing independently among the subjects who used inhaled steroids on a sustained basis during follow-up. However, only nine subjects used inhaled steroids on a sustained basis prior to age 3 years. Among those nine subjects, all had persistence of wheezing at age 3 years, and enrollment SB-FE_NO values were greater than 30ppb in eight of the nine subjects.

In conclusion, we have demonstrated that SB-FE_NO measurements in infants and toddlers with a history of recurrent episodes of physician-diagnosed wheezing is associated with the persistence of wheezing and risk of acute exacerbations of wheezing through age 3 years, and a decline in lung function between age 15 months and age 3 years. Furthermore, SB-FE_NO was a superior predictor of these clinical outcomes than the API or documented BDR during lung function testing. In contrast, tidal-FE_NO, although associated with decline in lung function, was a poor predictor of future clinical outcomes in wheezy infants/toddlers. A larger multi-center study with follow-up into the school-age years will be required to definitely evaluate the predictive value of SB-FE_NO for asthma in wheezy infants and toddlers.

FIGURE 5.

Receiver Operator Curves (ROC) of enrollment single-breath exhaled nitric oxide (SB-FE_NO) concentrations as predictors of persistence of wheezing at age 3 yrs. (panel A.) and acute exacerbations of wheezing (panel B.) between ages 2.5 and 3 years.

Abbreviations

FE_NO

fractional concentration of exhaled nitric oxide

SB-FENO

single-breath exhaled nitric oxide

Tidal-FE_NO

mixed expired exhaled nitric oxide during tidal breathing

BDR

bronchodilator responsive

API

Asthma Predictive Index

FVC

forced vital capacity

FEV_0.5

forced expiratory volume in 0.5 seconds

FEF_25-75

forced expiratory flow 25-75% of expiration

RVRTC

raised volume rapid thoracic compression

ROC

receiver operating characteristics

AUC

area under the curve

NHLBI

National Heart Lung Blood Institute

ATS

American Thoracic Society

ERS

European Respiratory Society