Prediction of asthma in preschool children is challenging and lacks objective indicators. The aim is to observe and analyze the variances between impulse oscillometry (IOS) and fractional expiratory nitric oxide (FeNO) in preschool children with wheezing, establish a joint prediction model, and explore the diagnostic value of combining IOS with FeNO in diagnosing asthma among preschool children.
Patients and methods:This study enrolled children aged 3-6 years with wheezing between June 2021 and June 2022. They were categorized as asthmatic (n=104) or non-asthmatic (n=109) after a 1-year follow-up. Clinical data, along with IOS and FeNO measurements from both groups, underwent univariate regression and multiple regression analyses to identify predictive factors and develop the most accurate model. The prediction model was built using the stepwise (stepAIC) method. The receiver operating characteristic curve (ROC), calibration curve, Hosmer-Lemeshow test, and decision curve analysis (DCA) were employed to validate and assess the model. Results: During univariate analysis, a history of allergic rhinitis, a history of eczema or atopic dermatitis, and measures including FeNO, R5, X5, R20, Fres, and R5-R20 were found to be associated with asthma diagnosis. Subsequent multivariate analysis revealed elevated FeNO, R5, and X5 as independent risk factors. The stepAIC method selected five factors (history of allergic rhinitis, history of eczema or atopic dermatitis, FeNO, R5, X5) and established a prediction model. The combined model achieved an AUROC of 0.94, with a sensitivity of 0.89 and specificity of 0.88, surpassing that of individual factors. Calibration plots and the HL test confirmed satisfactory accuracy.
Conclusion:This study has developed a prediction model based on five factors, potentially aiding clinicians in early identification of asthma risk among preschool children.