Effectiveness of pediatric drug-induced sleep endoscopy for REM-predominant obstructive sleep apnea

Study objectives Because dexmedetomidine (DEX)-induced sedation mimics non-rapid eye movement (NREM) sleep, its utility in sedating children with REM-predominant disease is unclear. We sought to determine the effectiveness of pediatric drug-induced sleep endoscopy (DISE) using DEX and ketamine for children with REM-predominant OSA, specifically whether or not at least one site of obstruction could be identified. Methods A retrospective case series of children without tonsillar hypertrophy undergoing DISE at a tertiary pediatric hospital from 10/2013 through 9/2015 who underwent subsequent surgery to address OSA with polysomnography (PSG) before and after. Results We included 56 children, mean age 5.6 +/- 5.4 years, age range 0.1-17.4 years, mean BMI 20.3 +/- 7.4 kg/m2 (76 +/- 29 percentile). At least one site of obstruction was identified in all patients, regardless of REM- or NREM-predominance. The mean obstructive apnea-hypopnea index (oAHI) improved (12.6 +/- 10.7 to 9.0 +/- 14.0 events/h) in children with REM-predominant (P= 0.013) and NREM-predominant disease (21.3 +/- 18.9 to 10.3 +/- 16.2 events/h) (P= 0.008). The proportion of children with a postoperative oAHI < 5 was 53% and 55% for REM- and NREMpredominant OSA, respectively. Unlike children with NREM-predominant disease, children with REM-predominant disease had significant improvement in the mean saturation nadir (P< 0.001), total sleep time (P= 0.006), and sleep efficiency (P= 0.015). Conclusions For children with OSA without tonsillar hypertrophy, DISE using DEX/ketamine was useful to predict at least one site of obstruction, even for those with REM-predominant OSA. DISE-directed outcomes resulted in significant improvements in mean oAHI, total sleep time, sleep efficiency, saturation nadir, and the proportion with oAHI < 5, after surgery for some children with REM-predominant disease.