Abstract
A major goal of asthma management is maintaining optimal control. Current assessment is based on symptoms and lung function.
We evaluated whether domiciliary daily home exhaled nitric oxide fraction (FeNO) monitoring could be useful as an index of asthma control. 50 asthmatic subjects and 15 healthy volunteers with a range of asthma severity underwent asthma control questionnaire (ACQ), spirometry before and after salbutamol and sputum induction. FeNO and peak expiratory flow (PEF) were measured twice daily for 2 weeks. A record of exacerbations was obtained 3 months later.
Diurnal FeNO variation in uncontrolled asthmatics was significantly greater than in controlled asthmatics (p<0.01). PEF variation was not different. The daily variation of FeNO levels was also greater in uncontrolled asthmatics compared with controlled asthmatic and healthy subjects (p<0.01). 80% of uncontrolled asthmatics experienced at least one or more exacerbations over the 3 months after the enrolment. The combination of diurnal FeNO variation ≥16.6% and ACQ scores ≥1.8 was best at predicting uncontrolled asthma (area under curve 0.91, 95% CI 0.86–0.97; p<0.001).
Diurnal variation in FeNO can be used as a biomarker of asthma control and as a predictor of the risk of future exacerbation. Prospective studies are warranted.
Abstract
Diurnal variation in FeNO can be used as a biomarker of asthma control and a predictor of the risk of future exacerbation http://ow.ly/r2MCY
Introduction
Over the past decade, there has been a clearer definition of the concepts of asthma severity and control with important implications for the management of asthma. Asthma control is defined by the components of clinical control and future risk of exacerbation, and asthma severity by the requirement for high intensity treatment [1]. The Global Initiative for Asthma (GINA) has also emphasised that a major goal of asthma management is not only achieving and maintaining optimal control but also reducing future risks, particularly those of exacerbations [2]. Symptom questionnaires and spirometry have been recommended for managing asthma [2]. The guidelines have also recommended the need for assessing control over a period of time rather than just at one assessment. Hence, home monitoring can be important for this [3, 4]. Daily home monitoring of peak expiratory flow (PEF) also provides an additional tool for asthma management [5, 6].
Using the recommended criteria for determining control, between 20% and 50% of asthma patients have been reported to be uncontrolled [7–9]. Uncontrolled asthma is associated with a decreased quality of life, chronic airflow limitation, a higher risk of having an exacerbation and hospitalisation, associated with a greater probability of death, and an increased economic burden [10–13]. One of the potential reasons for a continuing poor level of control is that the tools currently available to evaluate control may not be adequate. For example, a marker of inflammation is not currently included in the assessment of control when the level of airway inflammation is generally accepted to contribute to symptoms and to the risk of exacerbations [2].
Measurement of exhaled nitric oxide fraction (FeNO) has been available as an indirect way of assessing the eosinophilic inflammation of asthma [14]. Longitudinal measurements of FeNO in asthmatics may be helpful to predict deterioration [15, 16]. These daily fluctuations have been proposed to inform about both severity and control of asthma [17]. With the arrival of portable FeNO monitors, it is now possible for patients to measure FeNO in the home environment on a daily basis [17, 18].
We evaluated whether domiciliary diurnal variations and fluctuations of FeNO levels provide useful information to assess asthma control and determined their value for predicting future risks, particularly of asthma exacerbations. In addition, we wanted to determine whether these also could provide information about asthma severity. Finally, we determined how comparable this was to the analysis of PEF diurnal measurements.
Methods
Subjects
50 asthmatic subjects (22 nonsevere and 28 severe asthmatics) were recruited from outpatient clinics at the Royal Brompton Hospital, London, UK, and 15 normal healthy volunteers through advertisements. The diagnosis of asthma was made by respiratory physicians according to a clinical history of characteristic symptoms (i.e. dry cough, wheezing, chest tightness and breathlessness), as well as a history of either forced expiratory volume in 1 s (FEV1) reversibility ≥12% or provocative concentration of methacholine causing a 20% fall in FEV1 (PC20) of <8 mg·mL−1. Asthmatic subjects were categorised as having controlled, partly controlled or uncontrolled asthma according to GINA guidelines [2].
Uncontrolled asthmatic subjects were defined as having three or more features of the following features over 4 weeks before recruitment: 1) daily symptoms more than twice a week; 2) any limitation of activities; 3) any nocturnal symptoms or awakening; 4) need for reliever/rescue treatment more than twice a week; and 5) spirometry (PEF or FEV1) <80% predicted or personal best if known [2]. Patients who had fewer than two of these features were assigned as having stable controlled asthma, which includes controlled and partly controlled asthma. Asthma subjects were also defined as having severe asthma according to the American Thoracic Society (ATS) definition of severe therapy-resistant asthma [19].
Subjects were excluded if they were current smokers, had other concomitant respiratory diseases other than asthma, or a respiratory tract infection within 6 weeks of study entry. Healthy volunteers were defined as having a smoking history of <5 pack-years, no respiratory diseases and a negative PC20 (>16 mg·mL−1).
All participants gave informed consent to the protocol approved by the Ethics Committee of Royal Brompton and Harefield NHS Trust/National Heart and Lung Institute, London, UK.
Study design
At study entry, all subjects performed spirometry before and after four puffs of β2-adrenergic agonist (salbutamol), sputum induction, asthma control questionnaire (ACQ) [20], asthma quality life questionnaire (AQLQ) [21] and asthma severity, as well as asthma control status in line with the GINA guidelines [2]. Next, a portable FeNO monitor (NObreath; Bedfont Scientific Ltd, Rochester, UK), a peak flow monitor (PIKO; nSpire Health Ltd, London, UK) and an asthma diary were provided to all subjects in order to measure FeNO and PEF levels twice a day over 2 weeks at home. At the second visit, when subjects brought the monitors back, the asthma diary was collected.
Spirometry and reversibility tests
Spirometry tests were performed with a dry wedge spirometer (Vitalograph, Buckingham, UK) on the first visit. Asthmatic subjects continued their usual medications except for salbutamol. Forced vital capacity (FVC) and FEV1 levels before and after inhalation of 400 μg salbutamol were recorded and the best of three acceptable manoeuvres reproducible to within 200 mL or 5% was retained [22].
FeNO and PEF measurements
At the first visit, all subjects were instructed on how to obtain measurements of FeNO and PEF, and to follow these measurements while taking their inhaled therapy. FeNO measurements were conducted at a constant flow of 50 mL·s−1 in line with the ATS/European Respiratory Society (ERS) recommendations using a portable handheld analyser (Nobreath; Bedfont Scientific Ltd) [23]. FeNO was performed twice a day for 2 weeks. Three successive measurements were performed on each occasion and levels were recorded in the diary. PEF measurements were also made three times after FeNO assessment. The best of the three readings was automatically recorded onto the monitor. All measurements over 2 weeks were transferred to a computer when subjects returned the monitor.
In order to avoid measurement bias, subjects were asked to measure FeNO and PEF at the same time and the same place twice a day, in the morning between 07:00 h and 10:00 h, and in the evening between 18:00 h and 21:00 h. Subjects were advised not to take their usual medications prior to making the FeNO or PEF measurements. FeNO was measured prior to PEF measurements.
Recording of asthma exacerbations
After the 2 weeks of measurements, patients were asked to keep a record of exacerbations over the subsequent 3 months, and were reviewed at 3 months. Severe exacerbations were defined as those needing treatment with oral prednisolone, or an increase in the maintenance dose of prednisolone for at least ≥3 days, or a hospitalisation or emergency room visit because of asthma, requiring systemic corticosteroids. Moderate exacerbations were defined as ≥2 days with having one or more of the following: deterioration in symptoms, deterioration in PEF levels and increased rescue bronchodilator use that is not severe enough to warrant systemic corticosteroid use [5].
Data analysis
Data are shown as mean (95% confidence interval). The daily morning and evening FeNO levels were calculated as the mean of two or three measurements. Diurnal variations of FeNO and PEF were expressed as the amplitude (highest–lowest). In addition, we used other ways of expressing the daily variation of these parameters in relation to their predictive value for uncontrolled asthma (table 1). The diurnal variations of FeNO and PEF were compared between groups using Kruskall–Wallis and Mann–Whitney U-test with Bonferroni correction. Receiver operating characteristic (ROC) curves were constructed and the area under the curves (AUCs) were determined for the detection of uncontrolled asthma. Multivariate logistic regression analysis was performed to these three markers simultaneously in order to determine the independent predictors of uncontrolled asthma. Finally, the Chi-squared test was used to evaluate the exacerbation rates between stable controlled and uncontrolled asthmatic subjects. A two-tailed p-value of <0.05 was considered significant.
Results
Characteristics of subjects
22 subjects had stable controlled asthma and 28 had uncontrolled asthma (table 2). Sputum eosinophils (%) in both stable and uncontrolled asthmatic groups were significantly higher compared to those in the healthy group (p<0.05). There was a significant positive correlation between FeNO levels at the study entry and reversibility to salbutamol (r=0.663, p<0.001) and sputum eosinophil % (r=0.370, p=0.014). There was a negative correlation between FeNO levels at entry, FEV1 % pred (r= -0.393, p=0.001) and PEF % pred (r= -0.379, p=0.002). There was no difference in the incidence of allergic rhinitis between stable controlled and uncontrolled asthmatic subjects.
Diurnal variation FeNO and PEF
Examples of individual diurnal variation in FeNO and PEF are shown in figure 1. There was a significant diurnal FeNO variation (ΔFeNO(diurnal)) in uncontrolled asthmatic subjects (mean 15.6 ppb, 95% CI 12.5–18.7 ppb) compared with stable controlled asthmatics (mean 8.18 ppb, 95% CI 6.69–9.67 ppb) and healthy subjects (mean 6.05 ppb, 95% CI 5.19–6.90 ppb) (p<0.001). Conversely, no significant diurnal variation in PEF levels (ΔPEF (diurnal)) was observed in uncontrolled and stable controlled asthmatic subjects (mean 39.4 L·min−1, 95% CI 30.3–48.5 L·min−1 versus mean 34.5 L·min−1, 95% CI 29.1–39.9 L·min−1) (fig. 2a and b). When the asthmatic groups were divided according to asthma severity, no significant differences in ΔFeNO (diurnal) and ΔPEF (diurnal) could be found between nonsevere and severe groups, but there was a greater variation in these parameters between asthmatic groups and healthy subjects (fig. 2c and d). In addition, being on asthma medications such as oral prednisolone or theophylline did not influence the diurnal variation of FeNO levels.
Fluctuations of daily average FeNO and PEF levels over 2 weeks
The daily average FeNO levels in uncontrolled asthmatic subjects over 2 weeks were significantly higher than those in stable controlled asthmatic subjects (mean 48.6 ppb, 95% CI 38.9–58.4 ppb versus mean 34.5 ppb, 95% CI 27.6–41.4 ppb; p=0.03) and healthy subjects (mean 48.6 ppb, 95% CI 38.9–58.4 ppb versus mean 18.1 ppb, 95% CI 14.6–21.7 ppb; p<0.001). In addition, the fluctuation pattern in uncontrolled asthmatic subjects was significantly different from those in stable controlled asthmatics and healthy subjects according to repeated ANOVA analysis (p<0.001) (fig. 3a), whereas the fluctuation pattern between stable controlled and healthy subjects showed no difference. When asthmatic subjects were divided into asthma severities, there was no significant difference in the daily average FeNO levels as well as the fluctuation pattern between nonsevere and severe asthmatic group (fig. 3b).There was no significant difference in the fluctuation pattern of FeNO levels between morning and evening (data not shown).
Figure 4a shows the daily average levels of PEF based on asthma control status. The daily average PEF levels in uncontrolled asthmatic subjects were significantly lower than those in stable controlled asthmatic subjects (mean 311 L·min−1, 95% CI 277–346 L·min−1 versus mean 415 L·min−1, 95% CI 377–452 L·min−1; p<0.001) and healthy subjects (mean 311 L·min−1, 95% CI 277–346 L·min−1 versus mean 513 L·min−1, 95% CI 482–544 L·min−1; p<0.001). In addition, there were no significant fluctuated patterns in the daily average PEF levels over 2 weeks between healthy, stable controlled and uncontrolled asthmatic subjects (p>0.05). Similar results could be obtained when subjects were divided into healthy, nonsevere asthmatic and severe asthmatic subjects (fig. 4b). There was no difference in the fluctuation pattern of PEF levels between morning and evening (data not shown).
Asthma exacerbations
In uncontrolled asthmatic subjects group, 22 (78.6%) subjects experienced at least one or more asthma exacerbations, while only two (9%) subjects in stable controlled asthmatic subjects group suffered from exacerbation over the next 3 months after completed measurements (p<0.001) (table 3). 16 (57.1%) out of 24 asthmatic subjects with exacerbation had severe and eight (36.3%) had nonsevere asthma (table 4).
FeNO and PEF as predictor of control
ROC curves were constructed for predicting uncontrolled asthma using different patterns of calculations in FeNO and PEF levels (table 5). In terms of FeNO levels, ΔFeNO (diurnal) was the best predictor for evaluating uncontrolled asthmatic subjects (AUC 0.803, sensitivity of 64.3%, specificity of 95.5%, positive predictive value (PPV) of 94.7% and negative predictive value (NPV) of 67.7%; p<0.001). With regard to PEF levels, the best parameter to estimate uncontrolled asthmatic subjects was minimum PEF expressed as a percentage of maximum PEF (AUC 0.741, sensitivity of 60.7%, specificity of 81.8%, PPV of 81.0% and NPV of 62.1%; p<0.001) and %PEF (week) (AUC 0.741, sensitivity of 60.7%, specificity of 81.8%, PPV of 81.0% and NPV of 62.1%; p<0.001). ACQ score (AUC 0.772, sensitivity of 71.4%, specificity of 86.4%, PPV of 87.0% and NPV of 70.4%; p<0.001) was noted to predict uncontrolled asthmatic subjects.
Multiple logistic regression and ROC curves for predicting control
According to the multiple logistic regression analysis, ACQ scores, ΔFeNO (diurnal) and minimum PEF expressed as a percentage of maximum PEF were found to be independent parameters for predicting uncontrolled asthmatic subjects (table 6). Neither marker was superior according to the AUC curves, but the combination of ΔFeNO (diurnal) ≥16.6% and ACQ scores ≥1.8 was the best marker to detect uncontrolled asthmatic subjects with sensitivity of 85.7%, specificity of 86.4%, PPV of 88.9% and NPV of 82.6% (AUC 0.914, 95% CI 0.857–0.971; p<0.001) (figure 5).
Discussion
In 50 patients with asthma subdivided according to asthma control, we found significant diurnal FeNO variations in uncontrolled asthmatic subjects compared with stable controlled asthmatic and healthy subjects. In addition, the daily average FeNO levels, but not PEF, in uncontrolled asthmatic subjects over 2 weeks showed a greater fluctuation than those in stable controlled asthmatic and healthy subjects. Severity of asthma was indicated by the raised level of baseline FeNO but not by the disturbance in diurnal variation. Finally, the combination of diurnal FeNO variation and ACQ scores, rather than the use of each parameter separately, provides better information to predict uncontrolled asthma and future risks of asthma exacerbation.
FeNO has been shown to mirror eosinophilic inflammation in the airways of patients with asthma [14]. Studies have also supported its role in reflecting asthma control [15, 24, 25] and as a marker of response to inhaled corticosteroid therapy [26, 27]. However, several studies have shown its limitation in monitoring asthma therapy [28–30]. Support for a greater usefulness of daily FeNO measurements has recently been obtained [17, 18, 31, 32]. In the current study, a single measurement of FeNO correlated well with FEV1 % pred and with percentage of sputum eosinophils, indicating that FeNO reflected both the inflammatory component as well as airflow obstruction. A recent study showed that the level of asthma control was associated with inflammatory markers including FeNO and airway mucosal eosinophil numbers [33]. Our new findings indicate that the daily fluctuation and the diurnal variation of FeNO provided more information regarding the control of asthma and the future risk of exacerbation. Our data is in agreement with recent studies reporting that monitoring of FeNO revealed changes in FeNO prior to the onset of moderate exacerbations [32], and that analysis of the fluctuation of daily FeNO provided useful information on asthma control [17].
In our study, ΔFeNO (diurnal), but not ΔPEF (diurnal), in uncontrolled asthmatic subjects was significantly higher than in stable controlled asthma. There are reports indicating good relationships between diurnal PEF variations, asthma control and asthma exacerbation [6, 34, 35], while others report only limited usefulness as a measure of asthma control [36–38]. However, there are reports of significant diurnal variation in FeNO measurements in the mornings compared with evenings, as we have found [18, 39]. In our study, we measured the fluctuation as the absolute variability irrespective of whether the level was higher or lower in any one direction. Frey and colleagues reported that the exponent of long-range fluctuation (α value) of FeNO and PEF predicted individual future risks of asthma exacerbation and reflected the current state of asthma control [6, 17, 35, 40]. Our short-term analysis over only 2 weeks indicated that the diurnal fluctuation in FeNO was more related to asthma control than PEF was. ΔFeNO (diurnal) appears to be a better marker than FeNO (daily), ΔFeNO (week) or %FeNO (week) for predicting asthma control and future risks of asthma exacerbation.
When ΔFeNO (diurnal) and ACQ scores are combined, the highest AUC could be obtained, whereas neither parameter seemed superior when used alone. In addition, these parameters are independent of each other from the multiple logistic regression analysis. There is no gold standard to estimate asthma control. What is clear is that no single parameter is adequate to assess asthma control and to predict future risks of exacerbation [5, 41]. However, our results showed that most uncontrolled asthmatic subjects experienced asthma exacerbation over the following 3 months after measurements had been made, suggesting that asthma control could strongly influence the future risk of exacerbation, as reported previously [10–12].
One potential limitation of the study is that we did not assess adherence to treatment and because FeNO has been previously used to assess patient adherence [42], our diurnal variation of FeNO may be a reflection of poor adherence. However, it is unlikely that there was poor adherence in the group of patients we studied. This was a short 2-week observational study and all participants provided all the measurements of FeNO and PEF on a twice daily basis. As they had to take their inhaler treatments after taking these measurements, we presumed that this happened as this would be their usual treatment. One potential issue with the portable FeNO machine is its expense, which will preclude its widespread use as a home monitoring device for an individual patient. However, if a 2-week monitoring period is sufficient to predict future risks of exacerbations and help to prevent these, then the value of the FeNO monitor would be increased.
In conclusion, diurnal variation, as well as day-to-day variation in FeNO, can be used as a surrogate biomarker of asthma control and as a predictor of the risk of future exacerbation. Further prospective studies are needed to clarify which markers could be combined to provide the best prediction for poor asthma control and future risks of exacerbation. Whether the daily and diurnal fluctuation of FeNO could be used to optimise asthma management should also be studied.
Acknowledgments
We thank Bedfont Scientific Ltd (Rochester, UK) for allowing us to use the portable analyser (NObreath) to record FeNO.
Footnotes
Support statement: This project was supported by the National Institute for Health Research (NIHR) Respiratory Disease Biomedical Research Unit at the Royal Brompton NHS Foundation Trust and Imperial College London (both London, UK). K.F. Chung is a senior investigator of NIHR. J. Saito was supported by a grant from the Fukushima Medical University, Fukushima, Japan.
Conflict of interest: Disclosures can be found alongside the online version of this article at www.erj.ersjournals.com
- Received March 18, 2013.
- Accepted July 30, 2013.
- ©ERS 2014