Hypoxic Burden Predicts Symptomatic but Not Sleep Architecture Response to CPAP Therapy in Obstructive Sleep Apnea

The apnea-hypopnea index (AHI) remains the standard metric for assessing obstructive sleep apnea (OSA) severity, although it does not fully capture the physiological consequences of respiratory events. Hypoxic burden (HB), which integrates the frequency, depth, and duration of oxygen desaturations, has been proposed as a more comprehensive measure of OSA-related hypoxic stress. This study compared the prognostic value of HB and AHI for subjective and objective treatment outcomes in patients undergoing continuous positive airway pressure (CPAP) therapy. Data from 412 patients with OSA who underwent diagnostic and CPAP-treated polysomnography were analyzed, including 52 patients with available follow-up questionnaires. HB was calculated using a validated event-based algorithm. Multivariable regression models were used to predict patient-reported outcomes and changes in sleep architecture, adjusted for age, sex and body mass index. HB consistently demonstrated stronger associations with symptomatic outcomes than AHI. Models including HB showed improved model fit and explained variance. HB remained independently associated with lower follow-up fatigue and better sleep quality after adjustment for AHI. In contrast, AHI was the dominant predictor of objective sleep architecture improvements, including increases in N3 sleep, reductions in N1 sleep, and reductions in arousal index. HB provided additional independent information only for changes in arousal burden. These findings suggest that HB and AHI capture complementary dimensions of OSA severity. While AHI primarily reflects sleep fragmentation and physiological response to CPAP, HB appears more closely related to subjective treatment benefit. Incorporating HB alongside AHI may improve clinical phenotyping and prediction of treatment response in OSA.