Clinical significance of mouth breathing as a marker of severity in pediatric obstructive sleep apnea

To investigate whether clinically assessed mouth-breathing severity is independently associated with polysomnography–defined obstructive sleep apnea severity in children with adenoidal hypertrophy and/or tonsillar hypertrophy, and to evaluate its potential as a non-invasive predictor of moderate-to-severe obstructive sleep apnea. We retrospectively analyzed pediatric inpatients admitted for planned adenotonsillectomy at Shanghai Sixth People’s Hospital. Based on a mouth-breathing score (range 1–7), patients were categorized into mild (scores 1–3) or severe (scores 4–7) mouth-breathing groups. Obstructive sleep apnea severity was classified as non-to-mild [obstructive apnea–hypopnea index ≤ 5 events/h] or moderate-to-severe (obstructive apnea–hypopnea index > 5 events/h). Demographic, clinical, and polysomnography parameters were compared. Correlation and logistic regression analyses were performed to identify independent predictors of moderate-to-severe condition, and the diagnostic performance of the predictive model was evaluated. Compared with mild mouth-breathing group, the severe group demonstrated significantly higher apnea–hypopnea index, obstructive apnea–hypopnea index, percentage of total sleep time with oxygen saturation < 90%, and lower mean oxygen saturation (all P < 0.05). Mouth-breathing scores positively correlated with apnea–hypopnea index and obstructive apnea–hypopnea index (rho = 0.2173, 0.2102, P < 0.0001). Multivariate analysis identified severe mouth-breathing scores (OR = 2.357; 95% CI, 1.354–4.101; P = 0.002), female sex (OR = 0.602; 95% CI, 0.377–0.961; P = 0.034), and tonsillar grade > II (OR = 1.820; 95% CI, 1.147–2.886; P = 0.011) as independent predictors of moderate-to-severe diseases. The multivariable model yielded an AUC of 0.658. At the optimal cutoff defined by the Youden index (0.228), it demonstrated a sensitivity of 73.1%, a specificity of 49.6%, a positive predictive value (PPV) of 35.7%, and a negative predictive value (NPV) of 82.8%. In comparison, using a mouth-breathing score ≥ 4 as a standalone diagnostic criterion produced an AUC of 0.636, with corresponding sensitivity, specificity, PPV, and NPV of 82.4%, 35.5%, 32.8%, and 84.0%, respectively. Mouth-breathing score was independently associated with the severity of obstructive sleep apnea in children with adenoidal hypertrophy and/or tonsillar hypertrophy. These findings indicate that structured mouth-breathing assessment may serve as a simple, non-invasive adjunct to clinical risk stratification for pediatric obstructive sleep apnea.