Distinct phenotype and risk factor analysis of persistent airflow limitation among asthmatic children: a case-control study

Background: Persistent airflow limitation (PAL) in childhood asthma is associated with a poor prognosis. The aim of this study was to categorize asthmatic children with PAL into distinct phenotypes and investigate the risk factors associated with each phenotype.

Methods: We conducted a case-control study with a total of 119 PAL patients and 120 non-PAL (NPAL) individuals. To classify the patients into appropriate clusters, unsupervised cluster analysis using K-means clustering was employed. The clusters were then compared to explore different PAL phenotypes. Univariate and multivariate logistic regression analyses were performed to identify risk factors for PAL and calculate odds ratios (ORs) with 95% confidence intervals (95%CIs).

Results: K-means clustering divided patients into three clusters: Cluster 0 included 120 NPAL patients, Cluster 1 characterized by elevated blood neutrophils included 66 PAL patients, and Cluster 2 exhibited elevated blood eosinophils and FeNO levels, containing 53 PAL patients. Independent risk factors for PAL included older age in both Cluster 1 (9 $\sim$ 11y: OR 12.67, 95%CI 3.30-55.74; $\ge$ 11y: OR 5.42, 95%CI 1.26-25.69) and Cluster 2 (9 $\sim$ 11y: OR 7.25, 95%CI 1.70-33.35; $\ge$ 11y: OR 11.28, 95%CI 2.79-51.89), as well as pneumonia history, with an OR of 6.41(95%CI 1.34-33.41) in Cluster 1 and an OR of 7.92(95%CI 1.83-37.44) in Cluster 2. Furthermore, specific factors associated with Cluster 1 included BMI above 22 kg/ $m^2$ (OR 12.28, 95%CI 2.68-70.45), asthma duration exceeding three years (OR 4.77, 95%CI 1.60-15.94), and a blood neutrophil percentage between 0.4 and 0.5 (OR 4.13, 95%CI 1.17-16.6). In Cluster 2, independent risk factors included a blood eosinophil percentage greater than 0.07 (OR 4.36, 95%CI 1.16-19.73) and a high FeNO level (OR 3.94, 95%CI 1.35-11.97).

Conclusion: Our study identified two phenotypes of PAL in asthmatic children: non-eosinophilic and eosinophilic inflammation. Older age and a history of pneumonia were independent risk factors for both phenotypes. For non-eosinophilic inflammation PAL, specific contributing factors included higher BMI, long duration of asthma, and a blood neutrophil percentage between 0.4 and 0.5. Elevated FeNO levels and blood eosinophilic percentage were independently associated with eosinophilic inflammation PAL.