Association between Drug-Induced Sleep Endoscopy and Measures of Sleep Apnea Burden

To test the following associations: (1) complete obstruction on drug-induced sleep endoscopy (DISE) and polysomnographic and subjective measures of obstructive sleep apnea; (2) tongue base/epiglottic obstruction and apnea index.

Retrospective cohort.

Academic medical center.

Subjects included surgically naïve adult patients with DISE. Chart extraction included demographics, polysomnography, and Epworth Sleepiness Scale and SNORE25 (Symptoms of Nocturnal Obstruction and Related Events 25) scores. Each DISE video was examined for complete obstruction at velum, oropharynx, tongue, epiglottis (VOTE system). Student's t test, correlation, and multivariate linear regression were performed.

Among 65 subjects, complete obstruction was observed at 0 (3%), 1 (46%), 2 (48%), and 3 (3%) subsites, respectively. Subjects with 0-1 subsites vs 2-4 subsites of complete obstruction had similar apnea indexes (13 ± 24 vs 12 ± 17, P = .78, 83% power to detect difference of 15), apnea-hypopnea indexes (30 ± 25 vs 31 ± 28, P = .96, 54% power to detect difference of 15), Epworth Sleepiness Scale scores (11 ± 7 vs 12 ± 5, P = .34, 91% power to detect difference of 5), and SNORE25 scores (2.0 ± 1.1 vs 1.9 ± 1.0, P = .70, 96% power to detect difference of 1.0), with similar results after adjusting for age, sex, body mass index, and tonsil status. Neither tongue base nor epiglottic obstruction was associated with apnea index.

The number of subsites with complete obstruction on DISE was not associated with polysomnographic, subjective sleepiness, and quality-of-life measures. Tongue base and epiglottic obstruction were not associated with apnea index. Larger detailed analyses are needed to determine the importance of each site and degree of obstruction seen on DISE.