Predicting obstructive sleep apnea severity in children referred for polysomnography: use of the Pediatric Sleep Questionnaire and Subscales.

PURPOSE

This study evaluated the role of the Pediatric Sleep Questionnaire (PSQ) and associated subscales in predicting the severity of obstructive sleep apnea (OSA) in children referred for attended polysomnography (PSG).

METHODS

This is a retrospective study of children (0-18 years) who completed PSQs the night of their initial diagnostic PSG (2019-2020). We excluded children with previous PSG, positive airway pressure titrations, or underlying genetic or craniofacial syndromes. Area under the receiver operating characteristic curve (AUC [95%CIs]) were estimated for prediction of varying severities of obstructive apnea-hypopnea index (oAHI > 2, 5, 10, and 25/h) by the PSQ's sleep-related breathing disorders (SRBD) scale and subscales.

RESULTS

Of 477 children, median (IQR) age at PSG was 5.7 (4.3); 60% of children were male, 21% were obese, and 4% had oAHI > 25/h. SRBD score did not improve discrimination of OSA cases at any oAHI threshold, with AUC CI that crossed 50% at all severities. Snoring subscale scores were predictive at oAHI > 2/h (AUC = 64.5% [59.5-69.5%]), oAHI > 5/h (AUC = 64.3% [59.6-69.0%]), and oAHI > 10 (AUC = 67.2% [62.0-72.4%]) thresholds, but were not predictive at oAHI > 25/h. The addition of demographic data (age and gender) improved the classification of the SRBD scale.

CONCLUSIONS

When utilized in children referred for attended PSG due to concerns for an underlying sleep disorder, the PSQ snoring subscale was more predictive of OSA at varying thresholds than the SRBD scale. While the original intent of the PSQ was not for the purpose of predicting severity in children referred for PSG, future directions include augmenting the questionnaire with additional clinical variables.