0861 HSAT Diagnostic Performance in Identifying Residual Obstructive Sleep Apnea Compared to Polysomnography in Children Post Adenotonsillectomy

Abstract Introduction Obstructive sleep apnea (OSA) persists in up to 73% of children following adenotonsillectomy, yet post-operative OSA evaluation remains uncommon. Polysomnography is the gold standard for pediatric OSA assessment, but availability is limited. Home sleep apnea testing (HSAT) represents a potential alternative and is extensively validated in adult populations, though its accuracy in pediatrics remains underexplored. This investigation sought to determine the diagnostic concordance of HSAT with PSG for the diagnosis of OSA in children post-adenotonsillectomy. Methods In this prospective comparative effectiveness trial, children aged 5–12 years scheduled for clinical PSG to evaluate residual OSA following AT were recruited for comprehensive HSAT with electroencephalography in their home within one week of their clinical PSG. Obstructive apnea hypopnea index (OAHI) derived from HSAT and PSG were compared using Spearman correlation and Bland-Altman analysis. Concordance in OSA classification was assessed using an OAHI threshold of ≥2 and ≥5 events/hour. Diagnostic accuracy was evaluated using nonparametric AUC ROC analysis at cutoffs of 1, 2, and 5 events/hour. Results Forty-nine children completed both HSAT and PSG, median (range) age 9.9 (5.6 – 13.2) years including 24 (49%) females, with comorbidities including asthma (50%), environmental allergies (42%), and obesity (17%). OAHI based on PSG was 1.9 (0.1 – 33) events/hour. HSAT and PSG OAHI correlated strongly (Spearman r=0.74, p< 0.0001). Bland Altman analysis showed a mean difference of 0.9 events/hour (SD=4.7), with limits of agreement −8.2 to 10. Using OAHI cutoff of 2 events/hr, PSG identified OSA in 23 (47%) children, while HSAT identified 24 (49%). Relative to PSG, HSAT over-diagnosed 9 (18%) and under-diagnosed 7 (14%). At 5 events/hour, both modalities classified 16 (33%) identically, with minimal misclassification. AUC ROC analysis demonstrated excellent diagnostic accuracy: 0.8, 0.9, and 0.9 for cutoffs of 1, 2, and 5 events/hour, respectively. Conclusion Despite some night-to-night variability in OSA diagnosis at low OAHI thresholds, comprehensive HSAT with electroencephalography shows strong concordance and excellent diagnostic accuracy compared with PSG for the diagnosis of OSA in children. These findings support consideration for use of HSAT for pediatric OSA evaluation following adenotonsillectomy, with the addition of electroencephalography potentially enhancing accuracy. Support (if any) AASM Foundation