Screening for pediatric obstructive sleep apnea : performance of the Sleep Clinical Record in a Canadian cohort

Objectives: Pediatric obstructive sleep apnea (OSA) is underdiagnosed due to limited access to polysomnography (PSG), the diagnostic gold standard. The Sleep Clinical Record (SCR) is a clinical screening tool developed in Italy that incorporates physical examination findings, sleep-related symptoms, and neurobehavioural measures. This study aimed to evaluate the performance of the SCR in a Canadian pediatric population and to examine whether craniofacial and orthodontic features differ between children with and without PSG-confirmed OSA. Methods: This prospective cohort study enrolled 60 children aged 4–16 years undergoing level I overnight PSG at BC Children's Hospital. Each participant completed the SCR and a comprehensive orthodontic examination before PSG. Diagnostic performance was assessed for three Apnea-Hypopnea Index (AHI)-based OSA definitions (≥1, ≥2, ≥5 events/hour) using sensitivity, specificity, predictive values, likelihood ratios, and area under the receiver operating characteristic curve (AUC). An optimal SCR cutoff was determined via the Youden index, with the Brouillette score as comparator. Craniofacial and orthodontic parameters were compared between OSA and non-OSA groups using Mann-Whitney U and Fisher's exact tests, with Bonferroni correction for multiple comparisons. Results: Mean age was 9.0 ± 3.0 years; 61.7% were male. OSA prevalence was high across all thresholds (90.0% at AHI ≥1), leaving only six non-OSA participants and limiting the precision of specificity and negative predictive value estimates. The SCR showed moderate sensitivity (74.3–91.7%) but limited specificity (31.8–50.0%), low negative predictive value (18.8% at AHI ≥1), and modest discrimination (AUC 0.54–0.68), below the original Italian validation cohort. The Brouillette score performed comparably (AUC 0.689). A Youden-optimal cutoff of ≥8.4 improved specificity to 83.3% but reduced sensitivity to 64.7%. No anthropometric, craniofacial, or orthodontic parameters were significantly associated with OSA. Conclusions: The diagnostic performance of the SCR in this Canadian cohort did not replicate that of the original Italian validation study, likely reflecting differences in population characteristics, OSA prevalence, and referral patterns. The Brouillette score, a simplified three-item tool, performed comparably to the SCR, raising questions about the added value of its additional complexity. Further validation in community-based settings with lower OSA prevalence is warranted.