Obstruction patterns during drug-induced sleep endoscopy vs natural sleep endoscopy in patients with obstructive sleep apnea

Abstract

Importance: Drug-induced sleep endoscopy (DISE) has been suggested to be a valuable technique for identifying the obstruction site associated with sleep-disordered breathing. However, the reliability of DISE findings is controversial because the procedure uses sedative drugs, which may have implications for the obstruction patterns observed on DISE. Objective: To compare the obstruction patterns during DISE with the obstruction patterns during natural sleep endoscopy (NSE). Design, Setting, and Participants: This prospective cohort study was conducted between June 2013 and May 2018 in Kyungpook National University Chilgok Hospital in Daegu, South Korea. All analysis took place from June 1 to July 31, 2018. Participants had an obstructive sleep apnea diagnosis, were older than 18 years, and had an apnea hypopnea index higher than 5 on type I polysomnography. The patients initially enrolled were excluded from the study for not reaching adequate sleep depth and waking up during insertion of the nasopharyngoscope. Intervention: Patients underwent DISE using midazolam and NSE without sedatives on 2 different days. Main Outcomes and Measures: Obstruction findings were observed only in the same range (65-75) of the bispectral index on both days. Obstruction findings were classified according to the VOTE (velum, oropharynx lateral wall, tongue base, and epiglottis) classification. The extent of agreement between DISE and NSE findings was evaluated using Cohen weighted κ value. Results: The study included 26 patients with snoring or obstructive sleep apnea (mean [SD] age, 44.7 [10.3] years; predominantly male [22 (85%)]). The mean (SD) apnea hypopnea index was 41.9 (17.2) and the lowest mean (SD) oxygen saturation was 79.8% (12.2%). The degree of agreement in upper-airway obstruction between DISE and NSE was 76.9% (Cohen weighted κ = 0.42; 95% CI, 0.02-0.83) in the velum, 88.5% (Cohen weighted κ = 0.84; 95% CI, 0.67-1.01) in the oropharynx lateral wall, 69.2% (Cohen weighted κ = 0.66; 95% CI, 0.46-0.86) in the tongue base, and 92.3% (Cohen weighted κ = 0.67; 95% CI, 0.24-1.11) in the epiglottis. Agreement of configuration of the velum was 88.5% (Cohen weighted κ = 0.50; 95% CI,-0.03 to 1.03) and the epiglottis was 92.3% (Cohen weighted κ = 0.67; 95% CI, 0.24-1.11). Conclusions and Relevance: Obstruction patterns of the upper airway appeared to be in agreement between DISE and NSE, suggesting that DISE may be a reliable test; future studies of multiple positions and sleep stages with larger sample sizes may confirm these results.