摘要
Sleep-disordered breathing (SDB) is highly prevalent in children and is associated with significant cognitive, metabolic, and quality-of-life (QOL) impairments. While obesity is a primary risk factor, the role of specific dietary behaviors as independent, modifiable contributors to SDB symptom severity remains underexplored.
We aimed to (1) examine the association between dietary intake and SDB symptom severity in children, and (2) develop and validate a dietary-based prediction model for moderate-to-severe SDB (msSDB).
This cross-sectional study analyzed baseline data from 406 children (aged 5.0-12.9 years) with and without SDB symptoms. Caregivers completed a validated 25-item Short Food Frequency Questionnaire to assess dietary intake. SDB symptom severity was measured using the obstructive sleep apnea (OSA)-18 questionnaire, with a total score ≥ 60 defining msSDB. Multivariable regression models were used to identify dietary associations. A dietary prediction score was developed and its performance assessed using receiver operating characteristic (ROC) analysis, followed by external validation in a separate cohort (n = 24).
Among 406 participants (mean [SD] age, 8.0 [2.1] years; 174 girls [42.9%]), 80 children (19.7%) had msSDB. Higher intake of orange juice, sweets, chocolate, rice, and processed meats was significantly associated with greater total OSA-18 scores and worse subscale scores across domains. Conversely, fish for dinner and savory snacks were inversely associated with symptom burden. A dietary prediction model demonstrated acceptable discrimination for msSDB (area under the ROC curve [AUC] = 0.78; 95% CI, 0.72-0.84). The model showed 79.6% accuracy, 56.3% sensitivity, and 85.3% specificity, with comparable performance in the validation cohort.
Specific dietary behaviors are associated with SDB symptom severity in children. A simple dietary-based score can help identify children at risk for clinically significant SDB, supporting the investigation of dietary interventions in pediatric SDB management.