A79-01 Mouth Opening Driven by Residual Increases in Respiratory Effort Causes Unintentional Air Leaks During CPAP Therapy in OSA Patients

Rationale Unintentional air leaks are common during continuous positive airway pressure (CPAP) therapy and represent a leading cause of treatment failure. This study aims to investigate the temporal association between mouth opening (MO) driven by increases in respiratory effort (RE) and the occurrence of mouth leaks in patients with obstructive sleep apnea (OSA) during CPAP therapy. Methods In 200 OSA patients under CPAP therapy, polysomnography combined with simultaneous mandibular jaw movement (MJM) recording - using a chin-mounted sensor including a gyroscope (Gyr) and an accelerometer (Acc) (Sunrise, Namur, Belgium), and leak-flow monitoring were performed. The primary analysis covered 1,494 data fragments (duration: 9.34 ± 6.33 minutes) - stratified into 3 groups: leak-free (control), sequential intermittent (Group A), and isolated sustained leaks (Group B) (see Figure). The impact of leaks on sleep fragmentation and obstructive events rate was evaluated by Negative-binomial and Dirichlet regression analyses. Bidirectional Granger causality testing was used to test the putative causal relationship between mouth-opening and leak generation. Results Sleep studies under CPAP indicated residual moderate sleep fragmentation (median: 14.56/h) and a persistent apnea-hypopnea index of 10.12/h. Compared with control fragments, both leak patterns were significantly associated with higher RE levels and greater burden of mouth opening, as quantified by MJM amplitudes. The use of nasal (75%) or oro-nasal (25%) masks had no significant effect on these associations. Regression analyses showed that the presence of mouth leaks was associated with significant reduction in N3 sleep stage proportion (-5.34%), greater sleep fragmentation, and more residual obstructive events under CPAP. Each 1-SD increase in the number of leak episodes corresponded to a 38% (CI: 34-42%) rise in arousal index and a 34% (CI: 37-41) rise in residual obstructive-event rate. Bidirectional Granger causality tests indicated that, in most cases (93.6% in Group A; 91.5% in Group B), MO was preceded by RE and induced subsequent leaks. Conclusion Our findings indicate that unintentional mouth leaks during CPAP therapy are temporally linked with previous residual RE increase which in turn triggers mouth opening. Integrating MJM analysis to the current leak monitoring framework might enable a more comprehensive assessment of leak origins, thereby informing targeted intervention strategies to reduce unintentional leaks and, ultimately, significantly improve patient adherence and the overall effectiveness of PAP therapy.