Effectiveness of FeNO‐guided treatment in adult asthma patients: A systematic review and meta‐analysis

Asthma control is generally monitored by assessing symptoms and lung function. However, optimal treatment is also dependent on the type and extent of airway inflammation. Fraction of exhaled Nitric Oxide (FeNO) is a noninvasive biomarker of type 2 airway inflammation, but its effectiveness in guiding asthma treatment remains disputed. We performed a systematic review and meta‐analysis to obtain summary estimates of the effectiveness of FeNO‐guided asthma treatment.


| INTRODUC TI ON
Asthma is a chronic airway disease, characterized by airway inflammation and variable expiratory airflow limitation. 1 The goal of asthma treatment is to minimize symptoms, optimize lung function, and prevent acute exacerbations. The cornerstone of treatment consists of corticosteroids, preferably in inhaled form (ICS), combined with bronchodilators. International guidelines recommend to treat asthma using a stepwise approach: treatment is increased (i.e., by adding a medicine or increasing dosage) if disease is insufficiently controlled, and treatment is maintained or decreased when disease is stable. 1,2 Asthma control is commonly monitored by assessing symptoms, sometimes in combination with lung function testing. However, increasing evidence is emerging that asthma is a heterogenous disease with different inflammatory endotypes. 3 In most patients, asthma is predominantly driven by type 2 airway inflammation, with high levels of eosinophils. In others, type 2 inflammation plays a smaller, or no role at all, and these tend to respond poorly to corticosteroid therapy. Ideally, asthma treatment is tailored in accordance with the type and extent of airway inflammation.
It has been shown that the frequency of asthma exacerbations is significantly lower in patients in whom the dose of ICS is guided by sputum eosinophil levels, as compared with those in whom management is based on usual methods of asthma monitoring. 4 Unfortunately, sputum induction requires experienced laboratory personnel, is time-consuming, does not provide immediate results, and is not feasible in every patient. 5,6 An alternative could be Fraction of exhaled Nitric Oxide (FeNO), which strongly correlates with sputum eosinophils, is noninvasive and quick. [7][8][9][10][11] However, a Cochrane systematic review by Petsky and colleagues from 2016 found only limited evidence in favour of FeNO-guided asthma treatment. 12 Since then, several new studies have appeared. In addition, that Cochrane systematic review focused on asthma control in the overall population, without looking at specific subgroups.
We performed a systematic review and meta-analysis to summarize the effectiveness of FeNO-guided asthma treatment compared to usual (symptom-guided) treatment in (specific subgroups of) adult asthma patients.

| Search and selection
The abovementioned Cochrane systematic review by Petsky and colleagues on the effectiveness of FeNO-guided treatment in adult asthma patients (which included seven randomized controlled trials (RCTs); searches were performed in June 2016) served as our starting point. 12 The exact same search was used to identify studies published since then, that is, between 1 January 2016 and 9 May 2023. We searched the Cochrane Airways Group Trials Register (composed of airway-related RCTs identified through systematic searches in MEDLINE, Embase, Cochrane CENTRAL, PsycINFO and CINAHL databases, and through handsearching of respiratory medicine journals and conference abstracts). The full search strategy is reported in Data S1-Supplementary Material 1.
Study selection was independently performed by two reviewers. First, titles and abstracts were screened and disagreements were discussed between the two reviewers. All abstracts deemed potentially relevant were assessed for inclusion on full texts.
Differences in the full-text assessment between the reviewers were discussed. In case of disagreement, a third reviewer made the final decision. Studies were included if they were RCTs (study design) comparing the effectiveness of FeNO-guided treatment (intervention) versus usual treatment (control) in adult asthma patients (study population), or specific subgroups thereof. FeNO-guided treatment could be tailored by FeNO results alone, or in combination with other measures of asthma control, such as symptoms, lung function or other biomarkers of type 2 airway inflammation, such as blood eosinophils or periostin. 9 Usual treatment could be tailored by clinical asthma symptoms alone, or in combination with other variables such as lung function. Studies were only included if they reported on any of the following outcomes: asthma exacerbations (≥1 exacerbation during the study period, or exacerbation rate), asthma control (assessed by Asthma Control Questionnaire (ACQ) or Asthma Control Test (ACT)), quality of life (assessed by Asthma Quality of Life Questionnaire (AQLQ)), lung function (Forced Expiratory Volume in 1 second (FEV1) % predicted), medication use (dosage of ICS) or FeNO. We excluded nonrandomized studies, such as observational studies and literature reviews. We also excluded studies in children and studies only reported as conference abstracts. We included studies written in English, Dutch, French, German or Spanish. In addition to database searching, we also scanned reference lists of included articles for RCTs potentially missed in our search and selection process. We did not search trial registers for ongoing studies. 13,14 K E Y W O R D S asthma, eosinophils, Fraction of exhaled Nitric Oxide, meta-analysis, systematic review

Key messages
• FeNO-guided asthma treatment probably reduces the number of asthma exacerbations.
• No effect was found on severe exacerbations, quality of life, FEV1 and treatment dosage.
• There were no indications that effectiveness is different in subgroups of asthma patients.

| Data extraction and quality assessment
For each included study, descriptive data were collected regarding patient characteristics, intervention (i.e., FeNO-guided asthma treatment), control (i.e., usual treatment) and outcomes. The methodological quality of each study was assessed using the Cochrane Risk of Bias tool for RCTs. 15 Data extraction and quality assessment were performed by two reviewers independently, where differences were discussed. If necessary, a third reviewer made the final decision. For the meta-analyses, data extraction and quality assessment were performed exclusively on newly identified RCTs; for the seven RCTs already included in the Cochrane systematic review by Petsky and colleagues, we used the results that were previously extracted and presented in their review report. 12

| Statistical analysis
If possible, results were pooled using inverse-variance randomeffects meta-analysis, accounting for differences between studies.
Studies for which insufficient data were presented and outcomes for which insufficient studies were available (two or fewer) were described qualitatively. The following predefined subgroups were evaluated: asthma severity (mild-moderate vs. severe), asthma control (controlled vs. uncontrolled), allergy/atopy (allergic/atopic vs. nonallergic/non-atopic asthma), pregnancy and obesity. Analyses were performed in Review Manager. 16

| Grading the evidence
For each outcome, two investigators independently assigned the certainty of the evidence using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology, whereby the scientific evidence is assessed using five established criteria (study design, risk of bias, imprecision, inconsistency and reporting bias). Based on this, a final 'level of certainty' is attributed to the evidence, which can be high, moderate, low or very low. 17

| Funding and registration
This systematic review was part of a wider literature investigation of the usefulness of FeNO in the diagnosis and treatment of specific groups of asthma patients, performed by Cochrane Netherlands commissioned by the Dutch National Health Care Institute, for which the full report was published online in Dutch in June 2020. 18 Updated results are reported in the current article. The review protocol was inspired by the abovementioned prior Cochrane systematic review by Petsky and colleagues, but was not identical (e.g., the same search strategy was used, but the subgroup analyses were newly added). Our protocol was finalized and submitted to the Dutch National Health Care Institute prior to initiation of the searches and study selection process, and can be accessed at https://osf.io/ycxt7/. The current review has not been published in the Cochrane Library, and the authors of the previous Cochrane systematic review by Petsky and colleagues were not involved in this updated systematic review, but were contacted for clarification of some of the reported data.

| Study selection
The Cochrane systematic review by Petsky and colleagues identified seven RCTs up to June 2016. 12,[19][20][21][22][23][24][25] In our update of the search, we identified 640 records ( Figure 1). Of these, 598 could be excluded after screening titles and abstracts. For the remaining 42 records, full texts were assessed, of which 36 were excluded (Data S1-Supplementary Material 2). Of the remaining studies, five were RCTs that fulfilled the inclusion criteria, [26][27][28][29][30] and one was a reanalysis of an RCT already included in the Cochrane systematic review. 31 So in total, we included 12 RCTs (and one reanalysis) on the effectiveness of FeNO-guided asthma treatment, covering 2,116 randomized asthma patients. No protocols of, or conference abstracts corresponding to, ongoing studies were identified in our literature searches.

| Description of included RCTs
An overview of characteristics of included RCTs is provided in Table 1. The majority was conducted in Europe (n = 6) or Asia (n = 3), sample size ranged from 72 to 392 patients (median 162), and mean age ranged from 28 to 58 years. Study duration varied from 18 weeks to 1 year. Seven RCTs included participants with asthma regardless of control or severity, whereas the others included patients with mild-moderate asthma (n = 2), moderate-severe asthma (n = 1) or severe asthma (n = 2). Two other RCTs were performed in primary care, where it can be expected that only patients with mildmoderate asthma were included. One RCT only included pregnant asthma patients. In four RCTs, the treatment decision was based on the FeNO value alone in the FeNO-guided asthma treatment group, while in the remaining studies FeNO was combined with other characteristics or tools, such as ACQ (n = 5), GINA (Global Initiative for Asthma) guidelines (n = 2) or other biomarkers of type 2 airway inflammation (n = 1). The cut-off values of FeNO used to adjust treatment also differed between RCTs. Two RCTs used a single cut-off value, that is, either 15 or 25 ppb. If the FeNO value was above this cut-off value, increasing treatment was considered, whereas treatment was adjusted downwards if FeNO was below this value. Eight RCTs used two or even three cut-off values; if the FeNO value was between these cut-off values, treatment was not adjusted. Finally, two RCTs used the relative change in FeNO from the previous measurement to adjust treatment. Usual treatment in the control group was in most cases guided by ACQ (n = 5) or according to the GINA guidelines (n = 4). Five RCTs reported support from a FeNO manufacturer . 20,21,24,25,28 An overview of the risk of bias assessment is provided in Table 2.
Risk of bias was high in at least one domain for five RCTs (only one of them being among the five newly identified RCTs); for the remaining seven RCTs, risk of bias was unclear in at least one domain. Nine RCTs had an unclear risk of bias in the randomization procedure (i.e., unclear random sequence generation and/or allocation concealment). Four RCTs had a high risk of bias because staff and participants were not blinded, and this was unclear in three. Three RCTs had a high risk of bias because outcome assessors were not blinded, and this was unclear in six. In one RCT, risk of bias due to incomplete outcome data was unclear. In one RCT, there was high risk of selective reporting, and this was unclear in five. GRADE low level of evidence; Figure 2C). In the fourth RCT, the total number of OCS-requiring exacerbations throughout the study course was similar (n = 4 for FeNO-guided treatment vs. n = 6 for usual treatment; no p-value reported). In the fifth RCT, the total number of severe asthma exacerbations (requiring OCS treatment and/or hospitalization) was not significantly different between FeNO-guided treatment versus usual treatment (OR 0.64 (95%CI 0.27-1.56)). 21 Additionally, six RCTs reported on the number of patients with exacerbations requiring hospitalization during the study period. Three of these could be included in meta-analysis, but in two of these, the outcome was not observed at all, and in one of these, the outcome was also infrequent and no significant difference was found (OR = 0.14 (95%CI 0.01 to 2.67); GRADE very low level of evidence; Figure 2D). In the fourth F I G U R E 1 Study selection. RCT, no significant difference was found in the annual hospitalization rate (RR = 0.77 (95%CI 0.32 to 1.84)). 28 In the fifth RCT, no significant difference was found in median days of hospitalization per patient (absolute difference = 0; p = .25). 20 In the sixth RCT, the number of hospitalizations was similar (n = 3 for FeNO-guided treatment vs. n = 2 for usual care; no p-value reported). 21

| Asthma control
Nine RCTs reported on asthma control measured by ACQ (where lower scores indicate better asthma control) at final visit, of which six could be included in meta-analysis. Although none of these found a significant difference, in meta-analysis a statistically significantly lower ACQ was found for FeNO-guided treatment (MD = −0.10 (95%CI -0.18 to −0.02); GRADE low level of evidence; Figure 2E). In the seventh RCT, no significant difference in mean ACQ change from baseline was observed (MD = 0.14 (95%CI −0.14 to 0.42); p = .37). 20 In the eighth RCT, no significant difference was found in mean ACQ (MD = −0.05 (95%CI −0.15 to 0.06)). 21 In the ninth RCT, no significant difference was found in median ACQ score (FeNO group: 0.8 (IQR 0.4 to 1.8); control group: 0.8 (IQR 0.4 to 2); p = .7). 26 In addition, one RCT reported on asthma control measured by ACT (where higher scores indicate better asthma control) at final visit, and found no significant difference in mean scores (MD = −1 (95%CI −2.63 to 0.63)). 29

| Quality of life
Seven RCTs reported on quality of life measured by AQLQ (where lower scores indicate higher quality of life) at final visit, of which three could be included in meta-analysis, but no significant difference was found (MD = 0.02 (95%CI −0.10 to 0.14); GRADE low level of evidence; Figure 2F). In the fourth RCT, no difference was found in mean AQLQ change from baseline to final visit between groups (FeNO: −0.03 (SE 0.10); control −0.14 (SE 0.13); p = .30). 20 In the fifth RCT, no difference was found in median score on MiniAQLQ at final visit (FeNO: 6.2 (IQR 5.3 to 6.6), control: 6.2 (IQR 5.3 to 6.6); p = .5). 26 In the sixth RCT, the median overall MiniAQLQ score did not significantly improve more in

| Lung function
Eleven RCTs reported FEV1% predicted at final visit, of which eight could be included in meta-analysis, but no significant difference was found (MD = 0.14; 95%CI −0.98 to 1.26; GRADE low level of evidence; Figure 2G). In the remaining three RCTs, it was reported that there was no difference in FEV1 between groups over the duration of the study, 23 that no significant differences in FEV1 were observed between groups, 19 and that changes in FEV1 were not significantly different between the two groups, 20 without additional data reported.

| Medication use
Nine RCTs reported on the dosage of ICS (budesonide variant) at final visit, of which seven could be included in meta-analysis, but no significant difference was found (MD = −57 μg/day; 95%CI −135 to 20; GRADE low level of evidence; Figure 2H). In the eighth RCT, no significant difference was found in the median dose of ICS (FeNO:  100 μg/day (IQR 0-600); control: 0 μg/day (IQR 0-800); p = .8). 26 In the ninth RCT, the median change in ICS dose showed no change in the FeNO-guided group (range 0 to 250 μg/day), and was 125 μg/day (range − 250 to 250 μg/day) in the control group (p < .01). 20

| FeNO values
Nine RCTs reported on FeNO values at final visit, of which six could be included in the meta-analysis, but no significant difference was found (SMD = −0.07; 95%CI −0.23 to 0.08; GRADE low quality of evidence; Figure 2I). In the seventh RCT, a significant difference in median FeNO values was found (FeNO group: 15 (IQR 12 to 18); control group: 21 (IQR 14 to 29); p = .03). 26 In the eighth RCT, no significant difference was found (ratio of geometric means = 1.02; 95%CI 0.87 to 1.19). 28 In the ninth RCT, it was reported that the increase in FeNO was significantly greater in the control group than in the FeNO-guided group (p = .007). 19

| Subgroups based on asthma severity, asthma control, atopic asthma, pregnancy and obesity
Results on subgroup analyses are reported in Data S1- The FeNO-guided treatment protocol differed considerably across the included studies, both with regard to cut-offs used, as well as to whether or not additional measures of asthma control (e.g., ACQ) were incorporated. Despite this heterogeneity, RCT results were generally consistent for most outcomes. Still, our systematic review does not answer the question which protocol is optimal.
Given the fact that it is unlikely that there is one single cut-off above which patients will, and below which patients will not, respond to treatment, a two-cut-off protocol seems rational. In the in-between group, it may be appropriate to keep medication as it is, and to monitor the patient more closely. Alternatively, the relative change in FeNO from the previous measurement could be used, which has the advantage of taking into account inter-person differences in FeNO levels due to intrinsic and extrinsic factors not related to type 2 airway inflammation. 7 For the same reason, we believe that FeNO should ideally not be used as a standalone test to make treatment decisions, but should be combined with other simple measures of asthma control (e.g., ACQ), as most studies did. Whether addition of other noninvasive markers of asthma control that are mechanistically complimentary, such as blood eosinophils, can further improve outcomes is mostly unclear. 28,34,35 Over the past years, treatment with biologicals affecting the airway inflammatory pathways involved in asthma has rapidly emerged.
These biologicals are now being used in selected patients with severe asthma that is uncontrolled under conventional treatment. 36,37 FeNO plays an important role in selecting the optimal biological in a given patient. 7 In addition, large numbers of studies have evaluated a potential role of FeNO in asthma care, not only for treatment selection, but also for, for example, diagnosis of (eosinophilic) asthma, prediction of asthma outcomes and assessing adherence to treatm ent. [7][8][9][10]34,35,[38][39][40] Although FeNO has been implemented in clinical practice in many healthcare centres worldwide, discussion regarding the added value for most of these indications remains. 7,8 Our findings provide new evidence to this discussion.
Compared to the Cochrane systematic review by Petsky and colleagues from 2016, 12 we were able to almost double the number of included RCTs. Our findings confirm that FeNO can have a role in the treatment of adult asthma patients, but the added value in the general asthma population is likely to be limited. This is in line with the recommendations from most clinical guidelines and consensus documents, which generally advice against routinely using FeNO to monitor disease control in asthma patients, although most acknowledge that this can be considered in selected patients. 1,2,7,8 FeNOguided treatment led to a reduction of asthma exacerbations, but a (clinically relevant) effect on other outcomes could not be demonstrated. Therefore, especially asthma patients with type 2 airway inflammation and frequent exacerbations may benefit, and future studies could focus specifically on the prevention of exacerbations in this subgroup of patients. Future studies on FeNO-guided treatment in the general asthma population seem futile, considering the large amount of data already available.

AUTH O R CO NTR I B UTI O N S
All authors participated in the development of the protocol and interpretation of the data; RS performed the literature searches; DAK, JAD and PH did study selection, data extraction and data analysis; DAK drafted the first version of the manuscript; and all authors participated in refining the manuscript and approved with the final version.

ACK N O WLE D G E M ENTS
We thank Kevin Jenniskens, Isabelle Kamm, Steven W. Nijman and Robin W. Vernooij for their contribution to the study selection and data extraction.

FU N D I N G I N FO R M ATI O N
Commissioned by the Dutch National Health Care Institute.

Ilonka H. van Veen received payment from AstraZeneca and Sanofi
Genzyme. None of the other authors have a conflict of interest to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.