Fractional exhaled nitric oxide distribution and its relevant factors in the general adult population and its healthy subpopulation

Background Measurement of fractional exhaled nitric oxide (Feno) has been used in the diagnosis and management of asthma. Understanding the distribution of Feno in a larger resident population and its “healthy” subpopulation would contribute to the interpretation of Feno in clinical practice. Objective This study aimed to investigate the distribution and its associated factors in the adult population and its healthy subpopulations. Methods We conducted a cross-sectional study of 8,638 men and 17,288 women aged 20 years or older living in Miyagi prefecture, Japan. We investigated the distribution of Feno and its associated factors in all subjects, a subpopulation with no history of upper and lower airway diseases (healthy subpopulation 1), and a subpopulation with no history of upper and lower airway diseases, normal lung function, and no positivity for other biomarkers of type 2 inflammation (healthy subpopulation 2). Results The distribution of Feno in healthy subpopulations, especially in healthy subpopulation 2 (median [interquartile range], 17 [12-23] with 95th percentile of 36 ppb) was lower than in all subjects (19 [13-26] ppb with 95th percentile of 47 ppb). In healthy subpopulation 1, 10.3% had elevated Feno (≥35 ppb), and elevated Feno was positively associated with factors including obstructive ventilatory defect, blood eosinophilia, house dust mite–specific IgE positivity, and history of hypertension. Male sex was associated with elevated Feno in all subjects and healthy subpopulations. Conclusion The distribution of Feno in the healthy subpopulation supports the validity of the criteria (≥35 ppb) currently used in Japan for the diagnosis of asthma.

7][8][9] The 2011 American Thoracic Society (ATS) guidelines set FENO cutoff values of 25 and 50 ppb, with interpretations according to eosinophilic airway inflammation and responsiveness to steroid therapy. 102][13] Previous reports also suggested that various factors such as smoking status, sinus disease, and allergic rhinitis may influence FENO. 11,12,14,15e sought to elucidate the distribution of FENO in a large adult resident population and its ''healthy'' subpopulation (that is, with no history of upper and lower airway diseases that could influence FENO, such as asthma, allergic rhinitis, or chronic sinusitis) so we could contribute to the interpretation of FENO in clinical practice.We were also interested in how many people had elevated FENO and the characteristics surrounding it, including lung function parameters, in an adult healthy subpopulation.Therefore, in the present study, we examined FENO in a population of about 25,000 Japanese adults in the Tohoku Medical Megabank Communitybased Cohort Study (TMM CommCohort Study). 16,17We then investigated the distribution of FENO and its associated factors in a general adult population.Further, to figure out the distribution of FENO in healthy adult population, we examined the distribution in 2 subpopulations: one with no history of upper and lower airway diseases, including asthma, allergic rhinitis, and chronic sinusitis, termed healthy subpopulation 1; and the other with no history of upper or lower airway disease, normal lung function, and no positivity for other biomarkers of type 2 inflammation, termed healthy subpopulation 2. We also examined associated factors, including respiratory function parameters, smoking status, medical history, and other biomarkers of type 2 inflammation, to elucidate the characteristics of subjects with elevated FENO levels in healthy subpopulations.

Study design and subjects
7][18] The study included men and women aged 20 years and older living in Miyagi prefecture, northeastern Japan.The study was approved by the institutional ethics committee of the Tohoku Medical Megabank Organization (approval 2023-4-076).
Detailed methods are provided in the Methods section in this article's Online Repository at www.jaci-global.org.Briefly, data from 29,383 people who underwent FENO measurement and spirometry at a follow-up examination were used in this study.We excluded data from those who withdrew from the study by January 17, 2022; those who did not return the self-report questionnaire; those who did not undergo lung function testing; and those with missing data for height, weight, smoking history, white blood cell count, peripheral blood eosinophil count, specific IgE (sIgE) to house dust mite (HDM; Dermatophagoides farinae), and sIgE to cedar pollen (n 5 3,457).Therefore, only data from 25,926 participants were analyzed.

Assessment of FENO
FENO was measured using a NIOX VERO device (Circassia, Uppsala, Sweden).Measurements were performed according to the measurement guidelines jointly proposed by the ATS and the European Respiratory Society (ERS) 19 and the official statement of the Japanese Respiratory Society for FENO measurement and interpretation. 13 examined the distribution of FENO in (1) all participants, (2) healthy subpopulation 1 (participants without a history of airway diseases, including asthma, COPD, chronic bronchitis, other respiratory diseases, allergic rhinitis, and chronic sinusitis), and (3) healthy subpopulation 2 (participants without a history of airway diseases, restrictive or obstructive ventilatory defects, or positivity for type 2 biomarkers, including sIgE to HDM and cedar pollen and elevated blood eosinophil count [> _150 mL 21 ]).
FENO was categorized into the following 4 groups according to ATS/ERS 10 and Japanese Respiratory Society 13 guidelines: <25 ppb, 25-34 ppb, 35-49 ppb, and > _50 ppb.According to the cutoff value for the ancillary diagnosis of asthma, suggested by previous articles in Japan, 11,13 FENO > _35 ppb was considered elevated.

Assessment of lung function
To calculate the percentage predicted FEV 1 or vital capacity (% FEV 1 or %VC), which are values of the lung function measures adjusted for age, sex, and height, we used the reference spirometry values for Japanese adults calculated using the LMS (lambda, mu, and sigma) method. 20A restrictive ventilatory defect was defined as %VC of less than 80%.An obstructive ventilatory defect was defined as ratio of FEV 1 to forced VC (FVC) of less than 0.7.

Statistical analysis
Data are presented as means (standard deviations [SDs]) or medians (interquartile ranges [IQRs]) for continuous variables, and as numbers (%) for categorical variables.For the characteristics of the 4 subgroups classified according to FENO level, a trend test was performed for continuous variables using a simple linear model to assess the linear association.We also performed a chisquare test to compare the characteristics of categorical variables between FENO subgroups.To analyze the association between elevated FENO (> _35 ppb) and potentially associated factors, multivariate logistic analysis was performed to calculate odds ratios (ORs) and 95% confidence intervals (CIs).Receiver operating characteristic (ROC) analysis was performed to investigate the cutoff value of FENO that discriminates self-reported asthmatic subjects from healthy subjects (healthy subpopulation 2).The cutoff value was determined by the point on the ROC curve closest to the top left.P < .05 was considered significant.All analyses were performed by R 4.1.2software (R Project; www.r-project.org).

Characteristics of participants
Table I shows the characteristics of the study participants.A total of 8,638 men and 17,288 women met the inclusion criteria.The mean (SD) age of the study participants was 63.3 (11.7) years.Overall, 4.3% of participants had a lung function test with %VC less than 80 and 4.0% had an FEV 1 /FVC of less than 70%; 65.6% were never smokers, 25.8% were ex-smokers, and 8.6% were current smokers.The median [IQR] peripheral blood eosinophil count was 110 [66-180] mL 21 .The sIgE positivity to HDM was 17.9% and to cedar pollen was 40.5%.According to the self-report questionnaire regarding current medical history, 2,759 participants (5.3%) were diagnosed with asthma, 4,049 (15.6%) with allergic rhinitis, and 1,101 (4.2%) with chronic sinusitis.

FENO distribution in general population and healthy subpopulation
The median [IQR] value of FENO for all participants studied was 19 [13-26] ppb, with a 95th percentile of 47 ppb (Fig 1, A). FENO was elevated in the participants with a history of asthma compared to those without a diagnosis of asthma, regardless of smoking status, whereas FENO decreased in current smokers compared to never smokers or ex-smokers, regardless of asthma diagnosis (Table I), as previously reported. 11,12,15e stratified the subjects into 4 subgroups according to FENO level and compared characteristics, including physical and laboratory characteristics, smoking history, and diagnostic history, between subgroups (Table II).A total of 70.1% of the participants had FENO levels <25 ppb; 17.4% had 25.0-34.9ppb; 8.0% had 35.0-49.9ppb; and 4.5% had > _50 ppb.The subgroup with elevated FENO levels had more male subjects; lower %VC, %FEV 1 , and FEV 1 / FVC; a higher percentage of subjects with FEV 1 /FVC of less than 70%; fewer never smokers and current smokers; and more ex-smokers.Laboratory values in the subgroup with elevated FENO showed higher eosinophil counts and higher sIgE positivity to HDM and cedar pollen.Regarding medical history, asthma, allergic rhinitis, and chronic sinusitis were highly prevalent.
all participants studied (Table III).The results showed that factors associated with elevated FENO were male sex (OR [95% CI] 1.48  per day, 0.27 [0.20-0.37])was significantly negatively associated with having elevated FENO.The association between underweight and overweight was analyzed separately from that between height and weight, which could not be analyzed simultaneously because body mass index (BMI) is calculated from height and weight and cannot be independent of them.Nevertheless, being overweight was significant (1.15 [1.05-1.25]),and it correlated with elevated FENO levels (Table IV).
To determine the distribution of FENO in our healthy subjects, we then examined the distribution of FENO in healthy subpopulation 1 (n 5 20,137).Their median [IQR] FENO level was 18 [13-26] ppb, with a 95th percentile of 42 ppb (Fig 1 , B, and Table V).A total of 72.2% of healthy subpopulation 1 participants had FENO levels <25 ppb, 17.4% had 25.0-34.9ppb, 7.3% had 35.0-49.9ppb, and 3.0% had > _50 ppb.In healthy subpopulation 1, the subgroup with elevated FENO levels had more male subjects, lower % FEV 1 and FEV 1 /FVC, a higher percentage of subjects with FEV 1 / FVC less than 70%, fewer never smokers and current smokers, more ex-smokers, higher eosinophil counts, and higher sIgE positivity to HDM and cedar pollen (Table V).Multivariate logistic analysis showed that factors associated with elevated FENO were similar in all participants studied, including male sex (OR [95% CI] 1.66 [1.41  Values are expressed as mean (SD) or median [IQR] for continuous variables or as nos.(%) for categorical variables.Underweight was defined as BMI <18.5 kg/m 2 ; normal weight, BMI 18.5-24.9kg/m 2 ; and overweight/obesity, BMI > _25.0 kg/m 2 .
We further investigated the distribution of the subpopulation without a history of airway disease, a restrictive ventilatory defect (%VC < 80), an obstructive ventilatory defect (FEV 1 /FVC less than 0.7), and positivity of type 2 biomarkers (sIgE to the HDM and cedar pollen, and elevated blood eosinophil count [> _150 mL 21 ])-that is, healthy subpopulation 2-because we thought that the participants, who had no history of airway diseases but had a defect found on lung function testing and/or positivity of other type 2 biomarkers might have an undiagnosed airway disease that could influence FENO levels.We excluded these participants from the study population and examined the distribution of healthy subpopulation 2 (n 5 8,321;   III and IV).
Because both sex and smoking status were independently associated with elevated FENO in all subjects and in healthy subpopulations, we also examined the distribution in subgroups divided by sex and smoking status (Table VII).In healthy subpopulation 2, the FENO levels were higher in men than in women, irrespective of smoking status.
To investigate the cutoff value of FENO that discriminates those with self-reported asthma from healthy subjects, we performed ROC analysis using data from the 8321 individuals who were assumed to be healthy (healthy subpopulation 2; Table VI) and the 1386 individuals who self-reported asthma (Table I).ROC analysis of data from all individuals revealed that the cutoff value of FENO that discriminated between these 2 groups was 58 ppb (area under the receiver operating characteristic curve 5 0.650, sensitivity 5 0.56, specificity 5 0.65; see Fig E1 and Table E1 in the Online Repository at www.jaci-global.org).We also performed ROC analyses in the subgroups classified by sex and smoking history.The cutoff value in each subgroup is as follows: male never smokers, 68 ppb; male ex-smokers, 67 ppb; male current smokers, 61 ppb; female never smokers, 53 ppb; female ex-smokers, 37 ppb; and female current smokers, 31 ppb (Fig E1 and Table E1).

DISCUSSION
In this study, we investigated FENO in a large Japanese adult population and analyzed the distribution of FENO levels and factors associated with FENO levels.In addition, to estimate Underweight was defined as BMI <18.5 kg/m 2 ; normal weight, BMI 18.5-24.9kg/m 2 ; and overweight/obesity, BMI > _25.0 kg/m 2 .
the distribution of FENO levels in a healthy adult population, we analyzed the subpopulation with no history of upper or lower respiratory disease (healthy subpopulation 1), and the subpopulation with no history of upper or lower respiratory disease, normal lung function, and no positivity of other type 2 biomarkers (healthy subpopulation 2).The results showed that FENO levels were lower in healthy subpopulations 1 and 2 than in all subjects.In particular, the distribution of FENO in healthy subpopulation 2 was almost identical to that reported in a previous and relatively small healthy Japanese population. 12Our study suggests that the currently used indicator, FENO > _35 ppb for elevated FENO in Japan according to previous studies, 11,12 is reasonable for a diagnostic aid for asthma and asthma-COPD overlap.Of all the subjects in this study, 12.5% had elevated FENO (> _35 ppb), although only 12.9% of this population with elevated FENO had been diagnosed with asthma by a physician.To characterize the population with elevated FENO in a healthy adult population, we then performed an analysis in healthy subpopulations.The results showed that 10.3% of healthy subpopulation 1 had elevated FENO (> _35 ppb).In this population, male sex, BMI, obstructive ventilatory defect, peripheral blood eosinophilia, HDM sIgE positivity, cedar pollen sIgE positivity, and history of hypertension were significantly positively associated with elevated FENO, whereas current smoking was negatively associated.The population with elevated FENO in healthy subpopulation 1 is a symptom-free population with no history of upper or lower respiratory tract disease, albeit self-reported.However, there were more subjects with an obstructive ventilatory defect than those in the low FENO population, as well as more subjects with peripheral blood eosinophilia, as well as more HDM sIgE-positive subjects and cedar pollen sIgE-positive subjects.These facts may suggest that the subjects with elevated FENO in healthy subpopulation 1 may be partially confounded by those with undiagnosed upper and lower airway diseases that can cause elevated FENO, including asthma.To clarify the characteristics of these elevated subjects, we plan to follow them for development and changes in lung function over time.Values are expressed as mean (SD) or median [IQR] for continuous variables or as nos.(%) for categorical variables.Underweight was defined as BMI <18.5 kg/m 2 ; normal weight, BMI 18.5-24.9kg/m 2 ; and overweight/obesity, BMI > _25.0 kg/m 2 .
In the meantime, to estimate the distribution of FENO and its related factors in a healthy population under more stringent conditions for the reasons mentioned above, we performed analysis in a healthy subpopulation 2 and found that 6.5% had elevated FENO levels (> _35 ppb).In this population, male sex and some weight ranges were significantly positively associated with elevated FENO levels, and current smoking was negatively associated with elevated FENO levels.On the one hand, there are consistent reports of a negative association with current smoking, and a mechanism has been suggested. 21,22On the other hand, there are conflicting reports regarding sex, with some reporting an association 15 and others reporting no association. 12,14In the present analysis of a relatively large population, male sex was independently and positively associated with elevated FENO levels, even after adjustment for smoking status and body size, which differed between men and women.The results of this study regarding sex differences are not consistent with the biological studies showing estrogen action on inducible nitric oxide synthase activity [23][24][25] and that an increase in androgen receptors is associated with a decrease in FENO. 26Although the differences in their distribution indices, such as median, were small and unlikely to have a significant impact on actual clinical practice decisions,  further investigation is needed to determine whether male sex is mechanistically associated with higher FENO.Obesity was positively associated with elevated FENO in our analysis of a population including healthy individuals, in contrast to results in the asthmatic population, [27][28][29] especially in late-onset asthma. 27This association between obesity and elevated FENO was no longer significant in healthy subpopulation 2, but the results suggest an associated trend (Table IV).Similar to our study, previous reports in nonasthmatic children 30 and adults 31 have reported a positive association between obesity and FENO.This may suggest that in healthy subjects without type 2 inflammation, unlike in people with asthma, obesity might increase FENO without mediating the mechanism of type 2 inflammation.However, further biochemical analysis is needed to elucidate the exact mechanism.
Healthy subpopulation 2 had 6.5% with high FENO levels (> _35 ppb), indicating that FENO levels are still distributed in the high range despite this population's being assumed to be healthy and without type 2 inflammation.This observation suggests that FENO levels are influenced not only by the presence of type 2 inflammation but also by other factors, 32 including the microbiome, [33][34][35] airway S-nitrosothiol concentration, 36 airway pH, [37][38][39] and genetic background. 18,40These facts, including our findings, again remind us that FENO is only an adjunctive diagnostic tool in the diagnosis of asthma.
In the present study, we found a weak but independent association between elevated FENO and hypertension in all subjects and in healthy subpopulation 1, although there was no association in healthy subpopulation 2. Like a previous study that reported an association between allergic rhinitis and hypertension risk, 41 our results may suggest an unexplored pathomechanism between type 2 inflammation and hypertension.
We performed an ROC analysis of data from individuals who were considered healthy (healthy subpopulation 2) and individuals who self-reported asthma.The cutoff value of FENO was high, with lower area under the receiver operating characteristic curve, sensitivity, and specificity compared to a previous report. 11This is probably because those self-reporting asthma in our study were likely to include patients receiving treatment, including inhaled corticosteroid therapy, and both the median and range of FENO levels were lower (Table I) than those previously reported in untreated asthma. 11his study has several strengths.First, it is the largest ever cohort of FENO and spirometry in a voluntary population of community residents, including healthy subjects.Second, it is a valuable data set in a generally understudied ethnic population.However, there are several limitations.First, the information on participants' medical history is based on self-report through questionnaires, which may not reflect reality.However, to account for this possibility, this study attempted to select healthy participants using spirometry, eosinophil count, and sIgE in addition to selfreported medical history.The distribution of FENO results was almost identical to that previously reported in a small but healthy Japanese population screened by medical examination. 12Second, the study subjects were adults aged 20 years or older living in Miyagi prefecture, and only those who agreed to participate in the study were included, so their demographic composition differed from the actual resident population, 42 with more women and a higher mean age.The values for physical indicators, laboratory values, and disease prevalence may also differ from those of the actual resident population.To account for these biases, we performed a multivariable analysis to search for factors associated with high FENO levels after adjusting for explanatory factors.However, we cannot exclude the possibility of bias that cannot be fully adjusted for in the multivariate analysis.In addition, the data on the distribution of FENO may differ from the actual resident population.
In conclusion, we analyzed the distribution of FENO levels in a larger Japanese adult population than previously reported.We also analyzed their distribution in the subpopulation without upper and lower airway disease (healthy subpopulation 1) and in the subpopulation with no upper and lower airway disease, normal lung function, and negative biomarkers of type 2 inflammation (healthy subpopulation 2) to estimate the distribution of FENO levels in healthy adults.The distribution of FENO levels in healthy subpopulation 2, using more stringent criteria, was consistent with previous results in the healthy subpopulation, supporting the validity of the criteria for high FENO levels currently used in Japan.While high FENO levels were correlated with factors such as asthma, eosinophilia, and sIgE positivity, consistent with previous reports, male sex was independently and positively correlated with elevated FENO levels in all subjects and in the healthy subpopulations.

DISCLOSURE STATEMENT
Supported by grants from the Japan Society for the Promotion of Science (JSPS; Grant-in-Aid for Science Research, 19K10637 to T.N., 23K07594 to M.Y., and 20H03684 to S.H.); Tohoku Medical Megabank Project from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT); the Japan Agency for Medical Research and Development (AMED; JP21tm0124005); and JST SPRING (grant JPMJSP2114 to M.T.).This research used the supercomputer system provided by the Tohoku Medical Megabank Project founded by AMED (grant JP21tm0424601).
Disclosure of potential conflict of interest: The authors declare that they have no relevant conflicts of interest.

TABLE I .
Characteristics of participants according to asthma and smoking status

TABLE II .
Comparisons of subgroup characteristics according to FENO level Table VI).The median [IQR] value of FENO for healthy subpopulation 2 was 17 [12-23] ppb, with a 95th percentile of 36 ppb (Fig 1, C, and Table VI).A total of 78.1% of healthy subpopulation

TABLE III .
Logistic multivariable regression analysis for factors (including height and weight) associated with elevated FENO levels (> _35 ppb) by population

TABLE IV .
Logistic multivariable regression analysis for factors (including BMI) associated with elevated FENO levels (> _35 ppb) by population

TABLE V .
Comparisons of characteristics between subgroups classified by FENO in healthy subpopulation 1

TABLE VI .
Comparison of characteristics between subgroups classified by FENO in healthy subpopulation 2

TABLE VII .
Distribution of FENO classified by sex and smoking status by population