From Heterogeneity to Precision: Endotypic Traits in Pediatric Obstructive Sleep Apnea

Pediatric obstructive sleep apnea (pOSA) is a heterogeneous sleep-related breathing disorder with varying degrees of severity and clinical manifestations. These differences are apparent among children of different races even with similar degrees of risk factors, such as adenoidal and tonsillar size and body mass index. Specifically, data suggest that children identified as Black or Hispanic had a higher prevalence of OSA together with an increase in reported OSA symptoms (1). The association was also supported by the Cleveland Family Study (2), which demonstrated a strong association between Black/African American ethnicity and increased apnea–hypopnea index (AHI). However, the pathophysiological underpinnings of pOSA in different racial and ethnic groups are not well described or well understood.

The CHAT study (Childhood Adenotonsillectomy Trial) was a randomized controlled trial investigating the impact of adenotonsillectomy on neurocognitive and behavioral outcomes in children aged 5–9 years with OSA (obstructive AHI, 2–30 events/h). A total of 1,244 children underwent baseline polysomnography (PSG) and were screened for enrollment, following which 464 were randomized to receive adenotonsillectomy (AT) or watchful waiting (WW) (3). In this secondary analysis, which is published in this issue of AnnalsATS, Tsou and colleagues (pp. 1931–1941) used noninvasive techniques to explore the role of pathophysiologic endotypes associated with the presence and severity of pOSA across racial and ethnic groups (4). Specifically, the authors estimated four different endotypes: increased Pharyngeal collapsibility, altered Arousal threshold, elevated Loop gain, and reduced Muscle responsiveness/compensation (PALM) (5) in pOSA. Although these endotypes are well described in the adult literature (6), PALM endotypes in pOSA and changes in endotypic traits after interventions require further research.

Endotypic traits were evaluated using advanced noninvasive automated computational modeling in 1,117 baseline and 360 postintervention PSG data. Endotypes were estimated separately for rapid eye movement (REM) and non–rapid eye movement (NREM) sleep from an established protocol previously validated in adult NREM sleep. Data were further stratified according to caregiver-reported race (Black/African American, n = 529; White/Caucasian, n = 434; Asian/Indigenous, n = 22; or other, n = 105) and ethnicity (Hispanic/non-Hispanic). The main findings of the study highlighted that pOSA is a physiologically heterogeneous disorder whereby, beyond airway anatomy, pharyngeal collapsibility and reduced muscle compensation play key roles in pOSA with variability observed between races. Specifically, there was a higher REM-AHI observed in Black children (mean ± SD, 15.26 ± 23.33 events/h) and Asian/Indigenous (16.09 ± 33.52 events/h) children than in White children (5.80 ± 12.77 events/h). For the whole cohort combined, AHI severity during REM sleep was associated with greater upper airway collapsibility, decreased compensation, and higher arousal threshold but was not associated with loop gain. When the data were stratified according to race/ethnicity, Black children had worse collapsibility and higher arousal threshold, whereas children of Asian or Hispanic background had poorer compensation. The combination of collapsibility and compensation endotypes was estimated to account for 19% (95% confidence interval [CI], −1, 40%) of the association between Black race and increased REM-AHI, which increased to 25% (95% CI, 4, 47%) of the association when the contributions of all four endotypes were combined. Together, these endotypes may predispose patients to more severe and prolonged obstructive events in REM sleep (7). Similar endotypic relationships were not observed during NREM sleep.

The authors also highlight that the increased REM-AHI in Asian/Indigenous children (n = 22) was associated with collapsibility and compensation endotypes and was estimated to explain up to 80% (95% CI, 2, 186%) of the association. Although the point estimate of 80% suggests a potentially large mediation effect, the wide CI indicates that the true proportion mediated by the two endotypes could be small or even nonexistent, limiting interpretation of these data.

What is of increasing interest in pOSA are the changes in endotype after treatment, particularly the role of AT and the efficacy of more personalized and targeted treatments. In this study, the effects of AT on endotypes were analyzed as group data only, and it was observed that the reduction in AHI after AT was associated with an improvement in REM and NREM airway collapsibility (but not compensation), supporting the idea that adenotonsillar hypertrophy is a modifiable risk factor for pOSA. Not surprisingly, improvements in AHI and airway collapsibility were less pronounced in the WW group, even among those who experienced spontaneous resolution of OSA.

Important questions arise from these baseline data, specifically regarding how consistent these endotypes are with changes in sleep position on any given night and the inherent night-to-night variability of AHI that may result from sleeping position. These factors are crucial to our understanding, because evidence suggests that transitioning to nonsupine sleep position can lead to a substantial improvement in airway collapsibility, improve muscle compensation, and reduce loop gain, especially in patients with supine-related OSA (8, 9). Indeed, an acknowledged limitation of this study was the lack of data to estimate position-specific endotypes, because position data were recorded only as the percentage of supine versus nonsupine sleep for each participant.

Furthermore, thoughtful and critical consideration needs to be applied to the practice of stratifying and analyzing data by race and/or ethnicity. Although the racial and ethnic categories employed in this study are consistent with those commonly used in population health research, they encompass substantial heterogeneity. For instance, the “Black/African American” category includes individuals with diverse ancestral, cultural, and environmental backgrounds, such as Caribbean, African, multiracial, and African American populations. Similarly, aggregating Asian and Indigenous participants into a single analytical category is conceptually problematic, because these groups are socially, culturally, and biologically distinct. The use of such broad classification risks obscuring important within-group differences and may lead to misleading interpretations. Furthermore, stratifying and reporting data by “race” can inadvertently perpetuate the misconception that biological differences are inherently tied to race, when in fact race is largely a social construct. Rather, race as operationalized in this context may serve as a proxy for multiple interrelated determinants of OSA risk, such as craniofacial structure, genetic predisposition (10), socioeconomic disadvantage (11), and environmental exposures such as secondhand smoke, air pollution, and substandard housing (12, 13), all of which may contribute to airway inflammation. Future research should explore the contribution of these individual factors to endotypic traits.

In summary, the study presents interesting findings that endotypic traits can be derived from routine clinical PSG data, highlighting a feasible option to determine endotypic traits in clinical settings with the potential to guide personalized treatment algorithms (14). Future research should aim to establish normative values for these endotypes, determine age- and puberty-related changes, and evaluate whether trait-directed interventions improve outcomes beyond conventional therapy. Integrating social determinants of health with physiologic profiling will be essential to disentangle the intertwined biological and social as well as environmental risks for pOSA.