Data-driven identification of IL-17A inflammatory phenotypes in relation to paediatric obstructive sleep apnoea

21 Nov 2026 11:48 12:00
Ming Yang Presenter Hong Kong
Background: Paediatric obstructive sleep apnoea (OSA) is a prevalent disorder characterised by recurrent upper airway obstruction, resulting in hypoxia and sleep fragmentation. Epidemiological evidence links Interleukin (IL)-17A to OSA. However, few studies have examined both local and systemic IL 17A profiles to identify meaningful inflammatory phenotypes, and the relationship between these phenotypes and polysomnographic indices remains unclear. Methods: Children aged 5-12 years underwent clinical assessment and polysomnography. Plasma and nasal epithelial lining fluid (NELF) IL-17A were measured by cytometric bead array, with K-means clustering of standardised values identifying inflammatory phenotypes. Across clusters, differences in obstructive apnoea-hypopnoea index (OAHI), oxygen desaturation index (ODI), and arousal index (ARI) were evaluated using Kruskal-Wallis tests with Dunn's post-hoc tests and confounder-adjusted generalised linear models (GLMs). Results: One hundred and six children were recruited, with plasma and NELF collected. Three distinct IL-17A inflammatory phenotypes were identified as high-IL-17A (n=59), low-IL-17A (n=22), and discordant-IL-17A profiles (n=25) (Figure 1). OAHI (p=0.014) and ARI (p=0.002) differed significantly across clusters, whereas ODI (p=0.837) showed no significant differences. Post-hoc test showed significantly higher OAHI (median (Q1-Q3): 3.23 (0.60-17.12) vs. 0.63 (0.22-2.06), adjusted p=0.022) and ARI (13.70 (9.21-25.63) vs. 8.79 (7.15-10.76), adjusted p=0.003) in the high-IL-17A cluster than those in the low-IL-17A cluster. GLMs showed that both low IL 17A and discordant-IL-17A clusters had significantly lower OAHI than the high IL 17A cluster (low: β=-0.692, 95%CI(-1.280,-0.105), p=0.023; discordant: β=-0.664, 95%CI(-1.223,-0.104), p=0.022), with the same pattern for ARI (low: β=-0.338, 95%CI(-0.604,-0.071), p=0.015; discordant: β=-0.294, 95%CI(-0.548,-0.041), p=0.025), while no significant differences for ODI. Conclusions: Clustering IL-17A from plasma and NELF revealed a distinct inflammatory phenotype with more respiratory events and fragmented sleep. These findings support IL-17A as a biomarker for identifying children with a distinct pattern of specific OSA-related features.