The value of small airway function parameters and fractional exhaled nitric oxide for predicting positive methacholine challenge test in asthmatics of different ages with FEV1 ≥ 80% predicted

Abstract Background Small airway function parameters (SAFPs) combined with fractional exhaled nitric oxide (FeNO) can predict a positive methacholine challenge test (MCT) for asthma diagnosis. However, their predictive utility in patients with forced expiratory volume in one second (FEV1) ≥80% predicted within different age ranges remains unclear. This study aimed to assess the utility of SAFPs, alone or combined with FeNO, to predict a positive MCT in patients in two age groups (<55 and ≥55 years) with asthma‐suggestive symptoms and FEV1 ≥80% predicted. Methods We enrolled 846 Chinese patients with suspected asthma and standard spirometry, FeNO, and MCT findings. Using the area under the curves (AUCs), the utility of SAFPs, alone or combined with FeNO, for predicting a positive MCT was analyzed in a discovery (n = 534) and validation cohort (n = 312) in both age groups with FEV1 ≥80% predicted. Results In the discovery cohort, the optimal cut‐off values for predicting a positive MCT in patients aged <55 years (74.2% and 74.9% for forced expiratory flow (FEF)50% and FEF25%–75%, respectively) were higher than those in patients aged ≥55 years (65.0% and 62.9% for FEF50%, FEF25%–75%, respectively). However, the optimal FeNO value in patients aged <55 years (43 ppb) was lower than that in patients aged ≥55 years (48 ppb). FeNO combined with SAFPs (FEF50%, FEF25%–75%) significantly increased the AUCs in both groups (≥55 years [0.851 for FEF50% and 0.844 for FEF25%–75%]; <55 years [0.865 for FEF50% and 0.883 for FEF25%–75%]) compared with a single parameter (p < 0.05). These findings were confirmed in the validation cohort. Compared with patients ≥55 years, those aged <55 years had higher and lower optimal cut‐off values for SAFPs and FeNO, respectively. The AUCs of FeNO combined with SAFPs for predicting a positive MCT for asthma diagnosis were significantly higher than those of the individual parameters (p < 0.05) in both age groups. Conclusions There were age‐group differences in the utility of SAFPs combined with FeNO for predicting a positive MCT. Patients with an asthma‐suggestive history and a normal FEV1 should be stratified by age when using SAFPs combined with FeNO to predict a positive MCT.


| BACKGROUND
Asthma is a common disorder caused by chronic inflammation of the lower respiratory tract. Its diagnosis is based on typical symptoms of wheeze, recurrent breath shortness, chest tightness, and cough, as well as evidence of variable expiratory airflow limitation based on objective pulmonary function tests. 1 Both a bronchodilating test with salbutamol and a methacholine challenge test (MCT) are used for asthma diagnosis in clinical practice. 2,3 A positive bronchodilating test, which is indicated by significant FEV 1 reversibility (a threshold of 12% and 200 ml reversibility to 400 μg salbutamol) combined with typical respiratory symptoms, is recommended by the Global Initiative for Asthma as proof of asthma. Airway hyper-responsiveness (AHR) to MCT is reflective of another aspect of airway lability. A positive MCT with a provocative concentration of inhaled methacholine <16 mg/ml or ≤0.48 mg is highly sensitive for identifying AHR presence to allow asthma diagnosis in patients with typical symptoms suggestive of asthma and preserved baseline pulmonary function (FEV 1 >70% pred). [3][4][5] However, in most hospitals in China, performing an MCT costs >$73 and at least half an hour, which is expensive and timeconsuming. Moreover, it involves 11 steps, 6 which is inconvenient for patients, technicians, and clinicians. Notably, MCT can cause unpleasant feelings in patients and can potentially induce severe bronchospasm. 6 Therefore, the use of a pulmonary function test or a test combination to predict AHR presence or a positive MCT could provide great value to clinicians for asthma diagnosis.
Small airways, which are defined by an inner diameter <2 mm, represent a "quiet zone" and provide minimal resistance to airflow in normal adult lungs 7 ; however, they are vulnerable to obstruction in asthma. 8 In the diseased state, pulmonary function changes are not detected by standard spirometry until approximately 75% of the small airways are obstructed. 9,10 Therefore, it is crucial for predicting AHR presence in patients with a normal FEV 1 for early stage asthma diagnosis based on small airway function parameters (SAFPs). 11 Additionally, eosinophilic inflammation identified through fractional exhaled nitric oxide (FeNO) is also associated with the pathophysiological process of asthma. 12 A previous study reported that the area under the curve (AUC) of the combination of forced expiratory flow (FEF) between 25% and 75% (FEF 25%-75% ) and FeNO for predicting AHR presence in patients with cough-variant asthma was significantly higher than that of the parameters being applied separately (all p < 0.05); furthermore, FeNO >43 ppb and FEF 25%-75% <78.5% were the optimal cutoff values. 11 Therefore, SAFPs combined with FeNO can improve the predictive value for AHR presence in asthma diagnosis. 11 Further, compared with MCT, SAFP, and FeNO measurement are more accessible, safer, cheaper ($5-10 lower), and time-saving (15-20 min).
Both SAFPs and FeNO are age dependent 13,14 with older age being an important small-airway dysfunction (SAD) predictor in patients with asthma 9,15 and being associated with a greater daily dose of inhaled corticosteroid 9 and a worse clinical asthma expression. 16 Moreover, FeNO has been reported to decrease with aging in patients with asthma. 17 Under-or overdiagnosis using these objective tests may occur when age is not considered. 18 Therefore, age should be considered when using these parameters to predict a positive MCT for asthma diagnosis. A previous study defined older age as age ≥50 years. 9 In our recent unpublished study on a Chinese population, 55 years was used as the optimal cut-off value for determining the impact of age on small airway function (Lili Hou, unpublished paper).
This study aimed to assess the impact of age on the predictive value of SAFPs, alone or combined with FeNO, for a positive MCT in patients with asthma-suggestive symptoms and FEV 1 ≥80% predicted.
This could contribute to reduced over-or underprediction of a positive MCT for asthma diagnosis, as well as the elucidation of the treatment effect assessed by SAFPs and FeNO in patients with asthma with different age ranges.  [30]). All the participants provided informed written consent for study participation.

| Inclusion criteria
The inclusion criteria were as follows: age 18-80 years; history of chronic dyspnea, cough, and wheeze for ≥2 months; normal chest Xray or computed tomography results; predicted FEV 1 of ≥80% with spirometric measurement, and having undergone MCT.

| Exclusion criteria
The exclusion criteria were as follows: having fever or acute respiratory tract infections within the previous 8 weeks; having taken montelukast, long-acting β 2 -agonists, theophylline, anticholinergic agents, and inhaled or oral corticosteroids within the previous 4 weeks; or having comorbid severe systemic diseases, including chronic obstructive pulmonary diseases.

| FeNO measurements
FeNO was measured using NIOX MINO (Aerocrine AB) at a standard flow rate of 50 ml/s following the American Thoracic Society (ATS)/ European Respiratory Society (ERS) recommendations. 11,19 FeNO measurements were performed before the spirometric assessments and MCT since the involved breathing maneuvers could distort FeNO results. 11

| Spirometric measurements
Spirometry tests were performed between 8 and 11 a.m. using an MS-PFT spirometer (Jaeger or SensorMedics). Spirometry tests followed the standards and recommendations of the ATS/ERS. 6,20 Expected values for lung function parameters were based on the prediction equation for patients in East China.

| Methacholine challenge test
MCT was performed using the Jaeger APS Pro system using a Medic-Aid sidestream nebulizer with doubling methacholine doses (0.0725-0.48 mg) following the ATS/ERS recommendations. 6

| Outcomes
The primary outcome was the predictive value of SAFPs, alone or combined with FeNO, for the presence of a positive MCT in asthma diagnosis of the discovery cohort enrolled from two centers located in Shanghai, China. The secondary outcome was the confirmation of the primary outcome in a validation cohort enrolled from a third center located in Hangzhou, China.

| Statistical analysis
Data were analyzed for normality of distribution using Kolmogorov-Smirov test. Normally distributed data were presented as mean ± SD.
Nonnormally distributed data were expressed as median and interquartile range. Independent samples were compared using Student's t test (two-tailed) or Mann-Whitney U test. Count data were presented as the percentage and between-group comparisons were performed using the χ 2 . Between-parameter correlations were analyzed using Spearman's analysis.
For patients with suspected asthma who presented small airway dysfunction, MCT was considered the gold standard for defining AHR. Logistic regression was applied to determine the impact of continuous test variables with dichotomous state variables. Univariate logistic regression was used to determine the impact of the variables of interest; subsequently, multiple logistic regression was used to check whether the model could be improved. Predictive values of single or combined measurements were calculated by constructing receiver-operating characteristic (ROC) curves and measuring AUCs. 11 In the ROC plot, sensitivity was plotted against 100-specificity. We determined the cut-off value based on Youden's Index.
HOU ET AL.

| Optimal cut-off values for positive MCT prediction
The optimal cut-off values were calculated based on Youden's Index.

| Predictive values of single and combined variables for positive AHR in the validation cohort
Before age stratification, we found that FeNO >43 ppb has a sensitivity, specificity, PPV, and NPV of 61.23%, 84.11%, 80.50%, and 67.0%, respectively, for predicting a positive MCT in patients with suspected asthma. The FeNO value has been reported to be affected by age. 13,14 Specifically, FeNO >25.5 ppb had a high predictive value (AUC = 0.905) with a sensitivity and specificity of 82.2% and 90.0%, respectively, for CVA diagnosis in Chinese children with an average age of 8 years, 24 which differed from that in the aforementioned Chinese adults with CVA. 11 Consistent with this trend, our study showed that the FeNO value was higher in patients with positive AHR aged <55 years than in those aged ≥55 years (p < 0.05), which indicated that airway eosinophilic inflammation might be more severe in younger patients with asthma who may be more sensitive to corticosteroids. This is consistent with previous findings that older patients with asthma had lower FeNO level, which indirectly indicates predominant neutrophilic bronchial inflammation in elderly patients that contributes to greater airflow limitation. 17 However, our findings are inconsistent with a previous report that FeNO level increased with aging. 27 However, this previous study was conducted in a healthy population without a smoking history. 27 Compared with aging, bronchial inflammation might be more crucially involved in affecting the FeNO level in the present study. Before age stratification, the cut-off value of FeNO for predicting AHR was 43 ppb; however, for patients aged ≥55 years, this value was 48 ppb, which indicated an overdiagnosis in older participants. In patients aged ≥55 years, FeNO >48 ppb had a high specificity (88.70%), but a relative low sensitivity (50.78%), for identifying patients with asthma, which indicated that FeNO is limited as a sole diagnostic test for these patients. In patients aged <55 years, FeNO >43 ppb had a relative low sensitivity 68.92% and NPV 72.60%, but a higher specificity 85.31% and PPV 82.90%, for predicting AHR in patients with typical asthma-like symptoms and an asthma-suggestive history.
Finally, we validate the findings from the discovery cohort using a validation cohort. This indicates that our findings could be generalized in patients with asthma in other care centers in China. This study has several limitations. First, SAD included airway wall thickening, airway narrowing, and air trapping, and so forth. 28 However, we did not further confirm these changes using computed tomography or tissues. Further studies should elucidate the associations among changes of pulmonary function and imaging and histological structures of peripheral airways. Second, FeNO was specifically correlated with airway eosinophilic, but not neutrophilic, inflammation. In previous studies, the diagnostic accuracy of FeNO increased upon exclusion of patients with neutrophilic inflammation. 29,30 There is a need for sputum eosinophil and neutrophil measurements to confirm asthma phenotypes. Third, comorbid diseases, including rhinitis, gastroesophageal reflux disease, and so forth, may affect result interpretation, which should have been considered in this study. However, given the retrospective and crosssectional design of this study, there was limited information regarding these comorbid diseases. Furthermore, the predictive value Abbreviations: AUC, area under the curve; constant coef., constant coefficient of logistic regression; FEF 50% , forced expiratory flow at 50% of forced vital capacity; FEF 25%-75% , Forced expiratory flow between 25% and 75%; FeNO, fractional exhaled nitric oxide; FeNO coef. (p), FeNO coefficient of logistic regression and it's t test for the p value; +LR, positive likelihood ratios; −LR, negative likelihood ratios; MCT, methacholine challenge test; NPV, negative predictive values; PPV, positive predictive values; variable coef. (p), another characteristic variables coefficient of logistic regression and it's t test for the p value. a Compared with FEF 50% or FEF 25%-75% alone. of SAFPs, alone or combined with FeNO, for the presence of a positive MCT was only assessed in two age groups (≤54 years and ≥55 years). However, asthma can occur at any age 31,32 and exhibits a similar rate among adults. 33 Moreover, small airway function has been shown to decline with aging 22,23,34 and contribute to asthma emergence and exacerbation. 23,35 Unfortunately, there is limited information regarding the starting age of decline for small airway function in Chinese healthy individuals and patients with asthma.
Therefore, the trend of the value for predicting a positive MCT in these patients should be evaluated in more different age groups.
Unfortunately, some groups stratified at 10-year intervals in this study had few patients with asthma; therefore, future studies should recruit a larger study group to better represent the population.

| CONCLUSIONS
Patients with asthma with SAD, which was affected by aging, were more prone to have a positive MCT. Therefore, Chinese patients with FEV 1 ≥ 80% and an asthma-suggestive history should be stratified by age (55 years) when SAFPs (FEF 50% , FEF 25%-75% ), alone or combined with FeNO, are used to predict the presence of a positive MCT in asthma diagnosis.