Abstract:Abstract Introduction Prediction of oral appliance therapy (OAT) response to treat obstructive sleep apnea (OSA) remains a major clinical challenge (currently ~50% or less depending on treatment definition). Recent physiology-based studies indicate that different combinations of pathophysiological traits or “OSA endotypes” are associated with OAT outcomes. However, methodologies to identify OSA endotypes to inform patient selection for OAT have largely been limited to in-laboratory and detailed physiology testing which limits clinical translation potential. Accordingly, we recently developed a physiology-based model that uses standard polysomnography and clinical data to predict OAT outcomes with correct OAT treatment classification in ~60 to >90% of cases depending on the treatment outcome definition (Dutta et al, J Clin Sleep Med, 2022). Here we applied this model to a large clinical cohort who had a home-based diagnostic study to determine the proportion of people estimated to respond to OAT. Methods Model based OAT predictions were performed using clinical data (i.e., age and BMI) plus standard sleep study outputs from >6,500 consecutive home-based diagnostic studies (Sleep Profiler PSG- Advanced Brain Monitoring). All participants had an apnea/hypopnea index (AHI) >5 events/h. The proportion of participants who were estimated to respond to OAT according to the model was calculated across three commonly used treatment outcome definitions: 1) AHI< 5, 2) AHI< 10 events/h and 3) 50% reduction in AHI. Results On average, participants (39% female) were middle-aged (49±15 years ([Mean±SD]), had moderately severe OSA (AHI=28±23 events/h) and were obese (BMI=32±7 kg/m2). ~10% were predicted to fully respond (AHI< 5 events/h) to OAT, ~43% to achieve an AHI< 10 events/h and 42% to have a 50% reduction in AHI. Conversely, ~41% of participants were predicated to not respond to OAT across all 3 treatment definitions. Conclusion While further prospective validation is required, these real-world data suggest that application of a novel physiology based OSA endotype algorithm using standard clinical and home sleep study information may be helpful to identify successful candidates for OAT (and vice versa) in a substantial proportion of people diagnosed with OSA with greater accuracy that current clinical practice. Support (if any) DJE is funded by a National Health and Medical Research Investigator Grant (1196261).

0602 Application of a Physiology Based OSA Endotype Model to Predict Oral Appliance Outcomes Using Home Sleep Study Data