Reliability of drug-induced sedation endoscopy: interobserver agreement

Several studies have demonstrated the validity, reliability, and safety of drug-induced sedation endoscopy (DISE) in assessing the collapse of the upper airway (UA) in patients with obstructive sleep apnea hypoapnea syndrome (OSAHS). The aim of this study was to assess the interobserver agreement on DISE and on therapeutic decision between an expert observer and an observer in training.

This was a cross-sectional study. Thirty-one DISE videos performed in our service were randomly selected. Videos belonged to patients with OSAHS who wanted alternative treatments to CPAP. The videos were reviewed by an ENT experienced on DISE and a second observer in formation. Each observer independently assessed the presence of collapse of the UA according to a modified VOTE classification and proposed an alternative treatment to CPAP. Interobserver agreement kappa values were calculated.

In assessing the presence of collapse at different levels of the upper airway, the percentage of agreement was 80 % at the level of the soft palate (kappa = 0.1667), 89.29 % in the oropharynx (k = 0.7742), 80.65 % at the tongue base (k = 0.5571), and 74.17 % at the epiglottis (k = 0.4768). When degree and configuration of the collapse was evaluated, the interrater agreement was moderate to good, except at the level of the tongue base where the agreement was weak for both degree and configuration of collapse (k = 0.34 and 0.38, respectively). Interobserver agreement was moderate when the indication of alternative treatments to CPAP is valued based on the findings of DISE.

Overall, DISE is a reliable technique even when assessing interobserver agreement between an experienced observer and one in training; however, tongue base is the level of the upper airway that presents the greatest difficulties when assessing the collapse with DISE. Therefore, it is important to develop learning curves for this technique in order to obtain more reliable results.