0778 A Novel Endotype of Pediatric Obstructive Sleep Apnea Syndrome Revealed Using K-Means Clustering

Abstract Introduction With the recent emergence of precision medicine, endotype / phenotype research for sleep disorders has been increasing. However, there were still few related studies on pediatric obstructive sleep apnea syndrome (OSAS). The authors conducted this study to cluster pediatric OSAS patients using polysomnography (PSG) parameters and find new endotypes. Methods We retrospectively investigated the medical records of children and adolescents who had been diagnosed with OSAS (apnea-hypopnea index, AHI 1/h on PSG) at the single sleep disorders center. We used age, total sleep time (TST), sleep latency (SL), REM sleep latency (RL), sleep efficiency (SE), wake time after sleep onset (WASO), and the percentage of stage N1/N2/N3/R (%N1/%N2/%N3/%R) for clustering. Prior to clustering, the dimensions of variables were reduced through uniform manifold approximation projection (UMAP). And then K-means clustering (KM) were used for clustering. We used average silhouette index (ASI) as evaluation metrics. After clustering, the differences among clusters were compared using one-way ANOVA. Results Total 306 subjects were enrolled (mean age 8.2 ± 4.1 years old, boy : girl = 194:112). There were statistical significant differences in pediatric daytime sleepiness scale (PDSS), children's depression inventory (CDI), and age (p-value 0.01, respectively) among clusters. Sleep architectures, such as TST, SL, RL, WASO, and %N1/%N2/%N3/%R also showed a significance difference among clusters. There was no statistical significance in total AHI, SE, and sex. KM showed ASI 0.42 and revealed 4 clusters. Cluster 3 (daytime sleepiness with depressive mood, n=72) showed the higher PDSS (mean 13.9 ± 7.2) and CDI (mean 12.0 ± 9.6) than other clusters. Cluster 3 also showed the older age (mean 12.3 ± 3.1) than other clusters. Conclusion The authors revealed the novel endotype of pediatric OSAS accompanied by excessive daytime sleepiness and depressive mood regardless of AHI and SE, which were previously used as indicators of sleep quality. Further research is needed to identify new endotypes of pediatric OSAS that cannot be identified by the classic indicators. Support (if any) A grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health &Welfare, Republic of Korea (grant number : RS-2023-00267049)