Predicting treatment success in pediatric mild-to-moderate OSA: real-world evidence from a model based on polysomnographic parameters.

Abstract

Background: Pediatric mild-to-moderate obstructive sleep apnea (OSA) is often treated with intranasal corticosteroids (INCS), but response rates vary. Identifying early predictors of treatment success may facilitate individualized therapy.

Methods: We conducted a single-center, two-phase observational study to develop and validate a predictive model for INCS response in children aged 3-12 years with mild-to-moderate OSA, defined by a baseline apnea-hypopnea index (AHI) of 1.0-10.0 events/hour. The derivation cohort (n = 175) was retrospectively enrolled between 2019 and 2023. A prospective validation cohort (n = 60) was recruited between 2024 and 2025 using identical diagnostic and treatment protocols. All patients received standardized INCS therapy. Treatment response was defined as a ≥ 50% reduction in AHI along with improvement in clinical symptoms at 6-9 months follow-up.

Results: Candidate predictors were extracted from baseline clinical and polysomnographic (PSG) parameters. Multivariable logistic regression was used to identify independent predictors. A predictive nomogram model was constructed based on these variables and externally validated in the prospective cohort. The model demonstrated good calibration and discrimination.

Conclusion: This study presents a validated nomogram model based on PSG and clinical parameters to predict treatment response to INCS in pediatric OSA, supporting early decision-making and personalized treatment strategies.