Predictive Value of Sleep Architecture and Arousal Index for Surgical Outcomes in Obstructive Sleep Apnea

Abstract

Objective: While palatal surgery is an established treatment for obstructive sleep apnea (OSA), variability in outcomes has prompted investigation into factors predicting surgical success. This study aimed to evaluate whether preoperative polysomnographic (PSG) parameters, along with drug-induced sleep endoscopy (DISE) findings, can predict surgical outcomes in patients with OSA. Methods: This retrospective study included 56 adult patients with OSA who completed both pre- and postoperative PSG and DISE assessments. Patients were classified into a successful outcome group (SG; >= 20 events/h reduction in apnea-hypopnea index [AHI]) or a worse outcome group (WG; >= 20 events/h increase in AHI). Baseline PSG and DISE findings were compared between groups, and multivariate regression analysis was used to identify independent predictors of surgical success. Results: Patients in the SG exhibited significantly higher preoperative AHI and larger tonsil size compared to those in the WG. While DISE collapse patterns did not differ significantly between groups, key differences emerged in PSG metrics. SG patients had a significantly greater proportion of non-REM stage N3 sleep within non-REM sleep and a lower arousal index, both of which were independently associated with improved surgical outcomes. Additionally, respiratory events were more prominent during non-rapid eye movement sleep in SG patients. Multivariate analysis confirmed that a greater proportion of non-REM stage N3 sleep and lower arousal index were strong predictors of postoperative AHI reduction. Conclusions These findings may offer more reliable predictors of which OSA patients will benefit from sleep surgery and underscore the importance of incorporating detailed PSG analysis into preoperative evaluations. Level of Evidence: 4.