Asthma is a chronic disorder, characterised by three main pathophysiological characteristics: airways inflammation, variable and reversible bronchial obstruction, and bronchial hyper-responsiveness (BHR) to different stimuli.1 In addition, upper airways inflammatory disorders, including allergic rhinitis (AR), are frequently associated with asthma.2 Hence, airways inflammation represents a common pathogenic pathway. In this regard, the accurate measurement and monitoring of airways inflammation should be a crucial topic in managing patients with asthma as inflammation induces bronchial obstruction.3 Unfortunately, a direct assessment of bronchial inflammation needs invasive methods (e.g. bronchial biopsy and/or lavage) or sputum analysis; both are complex and require adequate setting. However, it is actually possible to evaluate bronchial inflammation also by measuring a surrogate marker, such as the fractional concentration of exhaled nitric oxide (FeNO), simply during an office visit.4,5 There is evidence that FeNO levels correlate well with peripheral or bronchial eosinophilia,6,7 thus it can be considered a marker of eosinophilic asthmatic inflammation. Recently, the determinants of exhaled NO in atopic asthma have been defined.7 In particular, the role of inducible nitric oxide synthase isoform in the airway inflammatory pathway has been investigated and well documented.7–9 In addition, a previous study showed that endogenous airway acidification during asthma attack was able to induce increased FeNO levels10, which, in theory, might also influence disease control in asthmatic patients with gastro-oesophageal reflux disease (GERD).

Asthma control is the long-term goal of asthma management even though many patients have sub-optimal or poor control, despite the large diffusion of international guidelines and availability of effective medications.1,11 In asthmatics it might be appropriate to assess and to monitor airways inflammation by measuring non-invasive biomarkers in order to achieve asthma control. Studies investigating this issue are still conflicting,7,12,13 as most of these studies are conducted on small numbers of patients or in experimental settings. Therefore, the present study aimed to evaluate whether FeNO measurement may relate with clinical and functional parameters in a cohort of patients with asthma observed in a real life setting.

Baseline demographic and clinical characteristics are shown in Table 1; statistically significant results of any relationship between FeNO and other variables are reported in Table 2.

Statistically significant differences in FeNO were evidenced for the following clinical variables: “asthma control”, “wheezing”, “asthma symptoms”, “rhinitis diagnosis”, “rhinitis symptoms”, “GERD”, and “pneumonia”.

In particular, poorly controlled asthmatics showed the highest FeNO values (median: 42.90, 25th–75th: 19.63–77.15; p<0.001) as shown in Fig. 1, with a probability almost four times greater to have pathological values compared to controlled asthmatic patients (OR: 3.71, 95% CI: 1.74–7.89; p=0.002).

Figure 1.

FeNO values in patients with controlled, partly controlled, and poorly controlled asthma.

(0.06MB).

The highest FeNO levels were recorded in subjects with wheezing (p<0.001) (Fig. 2A), asthma symptoms (p=0.002) (Fig. 2B), rhinitis diagnosis (p=0.009) (Fig. 2C), rhinitis symptoms (p=0.006) (Fig. 2D), and in patients without gastro-oesophageal reflux disease (p=0.038) (Fig. 3A) and any history of pneumonia (p=0.027) (Fig. 3A). On the other hand, there was no relevant correlation between ACT scores and FeNO values (r=−0.16).

Figure 2.

(A) FeNO values in patients without or with wheezing; (B) FeNO values in patients without or with asthma symptoms; (C) FeNO values in patients without or with rhinitis diagnosis; (D) FeNO values in patients without or with rhinitis symptoms.

(0.2MB).

Figure 3.

(A) FeNO values in patients without or with GERD; (B) FeNO values in patients without or with Pneumonia.

(0.1MB).

A statistically significant difference in ICS dose (p=0.031) was found between patients with pathological FeNO (median: 250.00μg; 25th–75th: 200.00–425.00) and patients with normal FeNO (400.00μg; 25th–75th: 200.00–600.00) (Fig. 4A).

Figure 4.

(A) ICS dose in patients with normal or pathological FeNO values; (B) ROC curve defining the FeNO cut-off level to predict poorly controlled patients.

(0.14MB).

The probability of pathological FeNO was more than three times higher in wheezing patients compared to patients with negative physical examination (OR: 3.50, 95% CI: 1.80–6.80; p<0.001).

Patients with asthma or rhinitis symptoms had almost double the probability of having a pathological FeNO (for asthma symptoms, OR: 1.79; 95% CI: 1.14–2.81; p=0.015; for rhinitis symptoms, OR: 1.89, 95% CI: 1.13–3.14; p=0.019) and the probability was more than twice if the patient had a diagnosis of rhinitis (OR: 2.11, 95% CI: 0.98–4.53; p=0.050).

In patients either without any gastro-oesophageal reflux complications (OR: 2.67, 95% CI: 1.19–5.96, p=0.024) or without any history of pneumonia (OR: 2.63, 95% CI: 0.98–7.02, p=0.048) the risk to have pathological FeNO values was more than two-fold higher compared to patients with any history of these diseases.

A receiving operator characteristic (ROC) curve was applied to assess how accurate FeNO values were in predicting poorly controlled asthmatic patients. FeNO values can be considered fair predictors for poorly controlled asthma: the area under the ROC curve was 0.70 (95% CI: 0.61–0.80; p<0.001), as reported in Fig. 4B. FeNO value of 29.95ppb (with a sensitivity of 65% and a specificity of 66.4%) was considered the cut-off value to define poorly controlled asthma.

FeNO has been considered a surrogate marker of airway inflammation, mainly in allergic patients. FeNO measurement is a simple tool to assess bronchial inflammation and this study aimed to investigate whether FeNO measurement may relate to clinical and/or functional parameters in asthmatics observed in a real life setting.

This study demonstrated that FeNO values were significantly higher in poorly controlled compared to both partly or controlled asthmatics. The highest FeNO values were associated with the presence of asthma signs at clinical examination and asthma symptoms in the past month, rhinitis diagnosis and nasal symptoms in the previous month, and low ICS dose, whereas GERD and pneumonia history were associated with the lowest FeNO values.

The logistic regression highlighted that the presence of impaired FeNO values (>25ppb) was significantly associated with poorly controlled asthma (OR 3.71) as well as asthma signs (OR 3.5) and recent symptoms (OR 1.79), comorbid rhinitis (OR 2.11), and current rhinitis symptoms (OR 1.89). On the other hand, GERD (OR 2.67) and pneumonia history (OR 2.63) were associated with low FeNO values. These findings were confirmed by the ROC curve with fair reliability (AUC 0.70). Therefore, high FeNO values may reflect an asthma state that is not adequately controlled, mainly concerning the presence of asthma signs and symptoms. In addition, the current study underlines the link between rhinitis and asthma: comorbid rhinitis is associated with bronchial inflammation and poorly controlled asthma.

Concerning GERD, this effect on FeNO levels might be explained by pH and airway acidification. Nitrite and nitrate levels in plasma, for example, can reflect the dietary intake rather than NO metabolism in vivo.18 Moreover, NO is also formed enzyme-independently from nitrite under acidic conditions.19 Hunt et al. showed that the airways pH drops dramatically during an acute asthma attack, which facilitates the conversion of nitrite to NO.10 Hence, increased NO concentrations in the exhaled air of asthmatic patients may reflect nitrite conversion rather than NOS activity. Those authors also speculated that therapies directed at normalising airway pH early in the course of an acute exacerbation of asthma will help to prevent the cascade of events that leads to airflow obstruction. In line with the latter observation, we may hypothesise that in our population of asthmatics with GERD the use of pump-proton inhibitors probably influences pH, determining its increase and, as a consequence, a lower production of FeNO. Therefore, our findings are consistent with previous reports underlining these pathophysiological aspects.20–22 Additionally, in patients with inflammatory bowel disease FeNO levels are slightly higher, correlating with disease activity, compared to controls suggesting a systemic involvement of the disease.23

With regard to the relation between eosinophils and pneumonia, it is known that eosinophils may protect from invading pathogens in the lung through mechanisms (e.g. bromination) able to kill bacteria.24,25 In particular, the univalent reduction of oxygen to super oxide anion is the first step in the formation of reactive oxygen species. These compounds can either spontaneously or enzymatically dis-mutate to hydrogen peroxide. Granulocytes contain peroxidases (myeloperoxidase and eosinophil peroxidase) that are able to catalyse the reaction of hydrogen peroxide with halides leading to the formation of hypohalides, e.g., hypochlorous acid.24 Recent gas chromatography–mass spectrometry (GC–MS) studies demonstrated that eosinophils use eosinophil peroxidase (EPO) to generate oxidants in allergen-triggered asthma.25 BrY (3-bromotyrosine), a specific marker of protein modification by reactive brominating species, was markedly enriched in broncho-alveolar lavage (BAL) proteins recovered from asthmatic subjects following exposure to segmental allergen challenge.25 Thus, one chemical pathway used by eosinophils to promote oxidative modification of proteins during asthma is through EPO-generated reactive brominating species. Another potential pathway for oxidative modification of tissues in asthma by eosinophils may involve formation of reactive nitrogen species (RNS). NO production is increased in asthma and NO is a short-lived radical that can be converted into more potent RNS. Thus, we may speculate that high FeNO levels are related to high eosinophilia and, as a consequence, to greater protection against pathogens.

The present findings highlight the close link between bronchial inflammation and asthma control level, as a FeNO value>29.9ppb is associated with poorly controlled asthma. In this regard, several studies provided evidence that FeNO might be a useful biomarker in monitoring of asthma treatment efficacy. Jones et al. reported that FeNO measurement was as useful as induced sputum analysis and BHR in assessing airway inflammation, with the advantage that they are easy to perform, mainly concerning the FeNO usefulness for diagnosing and predicting loss of control in asthmatic patients after corticosteroid withdrawal.26 A longitudinal study in unselected asthmatics showed that a FeNO decrease <40% or increase <30% precluded asthma control optimisation or deterioration, respectively.27 In addition, in ICS-naïve patients, FeNO>35ppb predicted asthma control improvement in response to ICS treatment. However, the overall FeNO ability to reflect asthma control was reduced in patients using high doses of inhaled corticosteroids, whereas FEV1 had little additional value in assessing asthma control.27 The same group reported in a real-world setting study that not only high but also intermediate FeNO levels that are associated with a significant improvement in asthma control after starting ICS treatment.28 Another study showed that the use of FENO to guide anti-inflammatory treatment within primary care significantly reduced the exacerbation rate and improved asthma symptom control without increasing overall inhaled corticosteroid use.29 In a non-blind, pragmatic, cluster-randomised trial in primary care enrolling adults with a doctor's asthma diagnosis, it was concluded that a symptom- plus FeNO-driven strategy reduced asthma medication use and sustained asthma control and quality of life, suggesting that this strategy is preferred for adult asthmatics in primary care.30 Another real life study documented that FeNO was a reliable marker of asthma control in asthmatic patients treated with budesonide/formoterol as acute reliever.31 A retrospective analysis of steroid-naive school-age children with atopic and non-atopic asthma evidenced that FeNO may be used to predict BHR to adenosine monophosphate (AMP) in atopic but not in non-atopic steroid naive asthmatic children.32 In an elderly population the reference ranges have to be changed as the optimal cut-off, for the prediction of asthma, were 30.5ppb for males and 20.5ppb for females.33 Recently, a new algorithm has been developed to compute FeNO at 50ml/s from tidal breathing measurements opening the way to standardised FeNO measurements in preschool children and uncooperative patients.34 Therefore, all these recent studies were consistent with the present report confirming the significant association between FeNO and asthma control grade.

On the contrary, there are also conflicting studies that should suggest caution in interpreting FeNO findings in particular conditions. One study, conducted on 118 patients with a primary care diagnosis of asthma, tested the hypothesis that titrating corticosteroid dose, using the concentration of FeNO, could result in fewer asthma exacerbations (primary endpoint) and more efficient use of corticosteroids, when compared with traditional management.35 The results showed that an asthma treatment strategy based on FeNO measurement did not result in a large reduction in asthma exacerbations or in the total amount of inhaled corticosteroid therapy used over 12 months, when compared with current asthma guidelines.35 A paediatric study reported that common steps in asthma treatment, i.e. inhalation instruction and increasing ICS dose, were both ineffective in reducing FeNO in atopic asthmatic children with elevated FeNO values despite treatment with ICS.36 Thus, the authors concluded that FENO cannot simply be incorporated in current treatment guidelines. Another study performed in young patients with persistent asthma reported that conventional asthma management resulted in good control of symptoms in most participants, whereas FeNO addition as an indicator of asthma control resulted in higher doses of inhaled corticosteroids, without clinically important improvements in symptomatic asthma control.37 A paediatric study assessed daily FeNO 0.05 and/or symptoms tele-monitoring in the management of childhood asthma38 revealing that daily FeNO 0.05 monitoring compared with daily symptom monitoring alone did not added clinical value. Furthermore, a systematic review and meta-analysis reported that tailoring of asthma treatment based on FeNO levels has not been shown to be effective in improving asthma outcomes in children and adults.39 The authors concluded that there was insufficient justification to advocate the routine use of either sputum analysis (due to technical expertise required) or FeNO in everyday clinical practice. A real-life study also reported that in a broad spectrum of asthmatics encountered in clinical practice, sputum eosinophilia and methacholine bronchial hyperresponsiveness, but not FeNO, were associated with uncontrolled asthma.40

However, Turner very recently pointed out that the past studies contributed to the rise and the apparent fall of FeNO as biomarker for asthma, but recent evidence would suggest that FeNO measurement may play a role in the management of childhood asthma, and in particular preventing exacerbations.41 However, it should also consider that in addition to FeNO measurement there are also other alternative non-invasive methods for assessing inflammation in respiratory tract, for example the analysis of exhaled breath condensates with various immunological and biochemical markers, such leukotrienes, cytokines, metallo-proteinases, etc.42,43

The relevance of our study is supported by the consistent number of patients enrolled and by the real life setting. However, possible limitations of this study are the cross-sectional design, the high prevalence of allergic patients, and the absence of comparisons with other inflammation mediators. Another relevant limitation is the lack of checking for the presence of chronic rhinosinusitis (CRS) with or without nasal polyposis, which may mimic classical rhinitis. These disorders have been described as being important determinants of asthma control and FeNO as their presence, in particular nasal polyps, is strictly related to worse asthma control and very high levels of FeNO, independently of concomitant asthma.44 These findings are also consistent with the current ones, as the current study reported the absence of a relationship between ACT scores and FeNO. Nasal polyps, but not FeNO, have been reported to be the only significant determinant of poor asthma control.45 In addition, nasal polyps and asthma have been shown to be independent determinants of high FeNO values in patients with chronic rhinosinusitis. Moreover, there is now evidence that one of the most severe phenotypes of severe asthma is the refractory eosinophilic one, which is described to be very often associated with nasal polyposis, increasing the evidence that nasal polyps are important determinants of a more severe asthma. Therefore, further studies should be conducted to consider this relevant issue.

In conclusion, the current study provides evidence that high FeNO is related to poorly controlled asthma and that relevant clinical features are associated with impaired FeNO values. These outcomes may suggest assessing FeNO in all naïve asthmatics in order to identify those with high levels (phenotyping), in whom the monitoring of FeNO may be useful to obtain clinically-related information for practical asthma management.

All authors contributed to the paper and none of them has any conflict of interest.