Domiciliary exhaled nitric oxide and eosinophilic airway inflammation in adults with asthma

To the Editor:

Sputum eosinophils are the gold standard measure of eosinophilic airway inflammation and there is strong evidence that titrating corticosteroid therapy to sputum eosinophils reduces the rate of asthma exacerbations [1]. However, this measurement is time consuming and requires specific technical expertise. Exhaled nitric oxide (F_eNO) is a noninvasive surrogate marker of eosinophilic airway inflammation that may be measured using small portable devices. Studies of using F_eNO to titrate therapy have yielded mixed results [2, 3]. However, it is possible that one-off clinic readings may not be the best method of utilising F_eNO, and the home monitoring potential of F_eNO has not yet been fully explored.

We aimed i) to determine the reliability and feasibility of F_eNO home monitoring in adults with moderate-to-severe asthma and ii) to investigate the relationship between serial F_eNO measurements and recognised biomarkers of eosinophilic airways disease in asthma, specifically sputum and blood eosinophil counts. We hypothesised that serial F_eNO measurements would correlate more closely with other biomarkers of eosinophilic airways disease than one-off readings, and that there may be a temporal relationship between F_eNO and other eosinophil markers.

In this pilot study we recruited 10 adults (age >18 years) with moderate-to-severe asthma (Global Initiative for Asthma treatment steps 3–5 [4]), diagnosed by a specialist asthma physician in a secondary care setting, according to British Thoracic Society guidelines [5]. The study was approved by the National Research Ethics Committee – East Midlands Leicester, and all participants gave their written informed consent. The participant group comprised five men and five women with a mean±sd age of 64±5 years. All participants were treated with inhaled corticosteroids (800–2000 μg per day beclometasone dipropionate equivalent) and long-acting β₂-agonists. Six patients received maintenance low-dose prednisolone (5–15 mg per day). Mean±sd post-bronchodilator forced expiratory volume in 1 s (FEV₁) (% pred) was 71.4±24.2 with mean±sd bronchodilator reversibility of 9.3±11.2%. A previous history of atopy was documented in three patients. Geometric mean (95% confidence interval) sputum eosinophil count was 6.1% (1.7–22.2%) and blood eosinophil count was 0.18×10⁹ cells·L⁻¹ (0.07–0.46×10⁹ cells·L⁻¹).

Participants were recruited during the period from December 2013 – April 2014. They were each provided with a NObreath F_eNO monitor (Bedfont Scientific, Maidstone, UK) and carefully instructed in its use. Participants were provided with a paper diary and asked to record three morning F_eNO readings on a daily basis over a 56-day period. Participants attended study visits twice-weekly, at which time daily diaries were retrieved, a blood sample was drawn for measurement of eosinophil count, and sputum induction was performed (or a spontaneous sample obtained) [6]. Sputum cell counts were performed by an experienced technician.

Eight out of ten participants completed the study and good quality data were obtained in each case. The median missing data rate for home-monitored F_eNO was 7.1% (range 2.7–14.3%). Home-monitored F_eNO was repeatable, with a median intraclass correlation coefficient of triplicate measurements of 0.83 (range 0.78–0.92). Further analyses presented below utilise the average of the three daily F_eNO measurements. Sputum samples were obtained and analysed successfully for 39 out of 40 study visits, with 36 of these being spontaneous samples.

In order to assess the relationship between F_eNO and recognised biomarkers of eosinophilic airways disease, we constructed a linear mixed model (SPSS 20, IBM Corporation, Somers, New York, NY, USA) in which blood or sputum eosinophil count was the dependent variable and F_eNO was specified as a fixed effect in the model. We specified an autoregressive covariance structure for the repeated measures of blood or sputum eosinophil count within each individual, since we observed that repeated measurements became less closely correlated with the passage of time. We performed our analysis using a range of lag periods and windows to determine 1) how changes in F_eNO related temporally to changes in blood and sputum eosinophil counts, and 2) whether taking a moving average (window) of 3, 5, 7 or 9 consecutive days' F_eNO readings could improve the correlation with blood and sputum eosinophil counts. Figure 1a–b show example data from a single individual, demonstrating the relationship between F_eNO and sputum eosinophil count. Increasing the window width results in smoothing of the data, masking day-to-day variability in F_eNO and allowing the underlying trends to be better seen.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Relationship between domiciliary exhaled nitric oxide and eosinophilic inflammation. Overlaid time series of exhaled nitric oxide (F_eNO) and sputum eosinophil count in one participant, with a F_eNO window of a) 1 day and b) 5 days. Colour maps indicating the effect size of F_eNO as a predictor of c) sputum eosinophil count and d) blood eosinophil count at a range of lag times and window sizes. A negative lag indicates that F_eNO in the present correlates with past blood or sputum eosinophil counts. A positive lag indicates that F_eNO in the present correlates with future blood or sputum eosinophil counts.

Figure 1c–d demonstrates colour maps indicating the effect size of F_eNO within the model to predict either sputum or blood eosinophil count at different lag periods and window sizes. A higher effect size indicates a stronger association. In these figures, a negative lag indicates that F_eNO in the present correlates with past blood or sputum eosinophil counts. A positive lag indicates that F_eNO in the present correlates with future blood or sputum eosinophil counts. The association between F_eNO and sputum eosinophils was strongest at a lag of -1 and a window of 5, indicating that changes in sputum eosinophils tended to precede changes in F_eNO by 1 day. The association between F_eNO and blood eosinophils was strongest at a lag of -5 and a window of 7, indicating that changes in blood eosinophils tended to precede changes in F_eNO by 5 days.

This is the first study to show that F_eNO is a reliable and repeatable method of measuring eosinophilic airway inflammation at home in patients with asthma. Our results indicate that maximal changes in F_eNO appear to lag behind changes in sputum and blood eosinophils by 1 and 5 days respectively. However, F_eNO can still give a good insight into the current level of sputum eosinophils, particularly if a moving average of 5–7 consecutive days' measurements is used. It should be noted that most patients in this study were able to provide spontaneous sputum samples and did not require formal induction with nebulised hypertonic saline. While spontaneous and induced sputum are not exactly equivalent, previous studies have indicated that the cellular content does not differ significantly between the two methods of expectoration [7].

The home monitoring approach has the advantage of patient convenience, as well as the potential for day-to-day adjustment of therapy or the detection and preventative treatment of incipient asthma exacerbations. Hashimoto et al. [8] recently performed a pragmatic randomised controlled trial in which an internet-based approach to oral corticosteroid tapering, incorporating domiciliary F_eNO, was compared to a conventional approach. The internet algorithm was found to allow more rapid tapering of oral corticosteroids without any reduction in asthma control or quality of life. Honkoop et al. [9] found that adjustment of inhaled corticosteroid therapy in primary care using a symptom plus F_eNO-driven strategy reduced asthma medication use while sustaining asthma control and quality of life. Malerba et al. [10] found that titrating inhaled corticosteroids using a combination of F_eNO and sputum eosinophils reduced exacerbation rates and asthma symptoms compared to a conventional strategy, without additional increases in inhaled corticosteroid doses.

In conclusion, serial F_eNO measurements may identify airway inflammation within close temporal proximity to changes in conventional biomarkers of eosinophilic disease such as sputum eosinophil count, with the added value of being of point-of-care. Further prospective trials are required to assess if this approach can successfully predict exacerbations or be used to titrate therapy in adults with asthma.