Diagnosing Tongue Base Obstruction in Pierre Robin Sequence Infants: Sleep vs Awake Endoscopy

Objective: To investigate whether awake endoscopy can diagnose base-of-tongue obstruction as reliably as sleep endoscopy in infants with Pierre Robin sequence (PRS). Design: The study was retrospective with the clinicians blinded to patient identity. Endoscopy findings were assessed and measured by the performing pediatric otolaryngologist. Setting: Tertiary care children's hospital. Patients: All infants with PRS managed between January 2005 and July 2015 were included. There were 141 patients, of which 35 underwent both awake endoscopy (AE) and drug-induced sleep endoscopy (DISE). Interventions: Bedside AE and DISE in the operating room. Main Outcome Measures: Presence of moderate or severe base-of-tongue collapse was assessed. Sensitivity, specificity, and positive likelihood ratio of AE findings as well as intertest differences between AE and DISE were calculated. Results: AE had 50.0% sensitivity (95% confidence interval [CI] 27.2%-72.8%) and 86.7% specificity (95% CI 59.5%-98.3%) for detecting base-of-tongue obstruction compared to DISE; false negative rate was 50.0% (n = 10). Positive likelihood ratio was 3.75 (CI 0.96-14.65). Compared to AE, DISE demonstrated significantly more cases of base-of-tongue obstruction (P = .039). Conclusions: Bedside AE has low sensitivity for detecting base-of-tongue collapse in infants with PRS. Because of the substantial false negative rate, AE may not be a reliable diagnostic modality for ruling out base-of-tongue obstruction in this susceptible population. DISE may be indicated in high-risk patients to avoid underdiagnosing upper airway obstruction.