Mediating effects of postoperative complications in children with high-risk obstructive apnea following adenotonsillectomy.

BACKGROUND

Pediatric patients with high-risk obstructive sleep apnea (OSA) undergoing adenotonsillectomy (AT) are at increased risk for postoperative complications. While prior studies have identified risk factors using univariate analysis and logistic regression, the mediating variables that explain how OSA severity leads to postoperative complications have not been studied. Identifying these mediating factors can guide risk stratification and targeted interventions for this vulnerable patient population.

METHODS

307 pediatric patients (<18 years) with high-risk OSA who underwent AT and polysomnography (PSG) from 2019 to 2021 were included. Causal-mediation analysis was conducted. Primary outcomes were intensive care unit (ICU) intervention, major respiratory intervention, and prolonged hospitalization (>48 h). Twenty-five different patient characteristics and PSG parameters were analyzed as mediators.

RESULTS

The relationship between high-risk OSA and all three postoperative complications is mediated by O2 nadir (P < 0.001), CO2 max (P<0.002), and time spent with oxygen saturation below 90% (TST O2 <90%) (P < 0.001). Sensitivity analyses revealed that the effects are sensitive to moderate levels of unmeasured confounding.

CONCLUSION

Prior univariate analysis identified O2 nadir and CO2 max as significantly associated with postoperative complications. In this study, TST O2 <90% emerged as a novel and consistent mediator, suggesting it may be an important marker for identifying high-risk patients that simple OSA severity scores may miss. While causal mediation analysis clarified which variables acted as candidate mediators, sensitivity analysis revealed that they are sensitive to moderate levels of unmeasured confounding and should be interpreted cautiously.