The aim of the study was to analyze results of FeNO levels and reported lack or occurrence of the respiratory/allergic symptoms measured at baseline and confronta-tion them with reported results of lack or occurrence of the respiratory/allergic symp-toms reported at follow-up in children after 8 years from the FeNO measurement. The results of this study suggest that children, aged 6–9 years, with asthma-like symptoms and FeNO values > 35 ppb have the highest chance of having respiratory disease in the future. The diagnostic odds ratio resulted in almost 47 score which according to the test value has an extremely good predictive meaning [18]. However, in this study, the other diagnostic indicators like AUC or sensitivity, specificity, and true positive values don’t support the value of the diagnostic odds ratio. The FeNO had the best accuracy in relation to asthma. The sensitivity was 40% with 95% confidence interval (95%CI) ranging from 0 to 83% for each analyzed FeNO cut-off (> 20 ppb, > 25 ppb and > 35 pbb). The specificity was 81.9% (95%CI 73%-90%), 94.4% (95%CI 89%-100%) and 98.6% (95%CI 96%-100%) for each FeNO cut-off. The true positive indicators for FeNO cut-offs were 13.3% (95%CI 9%-30%),33.3% (95%CI 0–71%) and 66.6% (95%CI 27%-100%), respectievely.The results of false negative indicators for FeNO cut-offs were 95.1% (95%CI 90%-100%), 95.7% (95%CI 91%-100%), 95.9% (95%CI 91%-100%), respectively. Such findings are consistent with the results reported in systematic re-view and meta-analyses from cross-sectional studies [19].
The other issue that cannot be omitted is the question, if whether the children with increased FeNO values were not yet diagnosed with asthma (underdiagnosis of asthma). From a total group of children (n = 447), there were 22 cases of asthma. In the group of children who participated in the follow-up part, at baseline there were 4 cases of asthma (mean FeNO: 33.0 ppb; range: 10–52 ppb) and 8 cases of asthma (mean FeNO: 47.2 ppb; range: 6-186 ppb) during follow-up. One child from baseline asthma had excluded asthma after 8 years so in total there were 5 new cases of asthma (mean FeNO: 53.4 ppb, range: 6-186 ppb). That probably suggests that children with elevated FeNO levels (> 34 ppb) have not been diagnosed by a physician. In the 5 new cases of asthma, 2 children had attacks of dyspnoea, 3 children had wheezy in the chest and in all 5 cases, there were symptoms of allergic rhinitis and atopic dermatitis. In the study performed by Caudri et. all the FeNO was used as a predictor of asthma in preschool children from PIAMA cohort. It shows FeNO as a significant predictor of asthma-like symptoms such as wheezing and steroid use in the future [20]. However, the study was performed in the clinical field and in children with any asthma-like symptoms, so the conditions of the study differed from the methodology of the presented study. Moreo-ver, the presented study was performed in children who were free of respiratory symptoms as well.
More studies focused on the accuracy of FeNO in the diagnosis of asthma rather than the prediction of FeNO in asthma diagnosis. Some studies showed that FeNO (> 15.8 ppb) diagnostic accuracy (area under the curve ROC = 0.53) was poor in children with symptoms suggesting asthma [21]. The different results were shown in a study performed by Malberg. Children with probable asthma had higher FeNO concentra-tions in comparison to healthy controls. Moreover, the same study showed the discri-minant accuracy of FeNO with a sensitivity of 86% and specificity of 92% [22]. That corresponds with the conclusions suggested by Pijnenburg. It seems that the measure-ment of FeNO in preschool children might be more accurate for asthma diagnosis in children and respiratory symptoms, particularly in atopic children [10].
Moreover, the in the author’s previous study, the conclusion considered FeNO measurement as not a good screening tool for pediatric asthma in a community setting [23]. However, the current results of this study show that children with asthma-like symptoms and elevated FeNO levels are at increased risk of asthma and need further investigation, such results have been presented in the cross-sectional screening study conducted by Prasad et. all [15].
The predictive meaning of FeNO needs to be controlled according to the determi-nants of FeNO. In the study performed by See and Christiani, ethnicity, height, self-reported rhinoconjunctivitis, and household smoke exposure were responsible for 10.3% of FeNO variability in children aged 6–11. [24]. According to their study, they based the symptoms from the respiratory system on the responses from the question-naire The lack of validated diagnosis probably decreased the explained variability of FeNO, however, their study reflected the study performed in the epidemiological con-ditions [23]. In the study performed by Garcia-Marcos, the variability of FeNO was ex-plained in 27%. The 20% of the variability was explained by age, rhinoconjunctivitis, positive skin prick test and removal of cat and/or dog from the house [25]. The internal variability (within the group) of FeNO seems to be stable according to current litera-ture [26, 27] and should not impact the differences between the status of children.
It seems that the predictive meaning of FeNO for asthma diagnosis is dependent on a couple of crucial factors. First, it is the level of FeNO which is influenced by a lot of determinants that need to be clearly explained [28, 29]. Moreover, the exact refer-ence FeNO values like for spirometry should be developed. Current literature results imply that baseline FeNO levels seem to be a good predictor of a greater risk of moder-ate to severe asthma exacerbations, especially in uncontrolled asthma. Such results in-crease the meaning of FeNO measurement for community-setting [30].
And finally, for the epidemiological purpose, it is necessary to accurately assess the respiratory status of children. The questionnaire is a very important tool but needs additional support to increase its accuracy since it seems that the level of it is not sati-factionary. In the study performed by Kim et. al. the ISAAC questionnaire sensitivity, specificity, and accuracy for allergic rhinitis were 39.8%, 76.9%, and 63.4, respectively [31].
Study limitations
The study has a couple of limitations. The most important is that children in the follow-up did not willing for the FeNO measurement. Measurement at follow-up could improve the results of the questionnaire and verify if children have elevated FeNO levels, especially in those who had asthma previously. Another limitation is related to the small number of new cases of asthma, more cases of asthma could improve the in-vestigation of current symptoms and FeNO levels in this specific group. A further problem is related to the nature of the study, the cross-sectional study on the popula-tional level should be marked with limited trust since all symptoms are declared by legal guardians. The study has a potential strength. According to the author’s knowledge, it is the first study that considers FeNO levels and confronted them with asthma-like symptoms during follow-up. Most studies, consider prospective methods but according to the assessment of the relation of FeNO to the treatment control of asthma.