0600 Identifying Responders to Mandibular Advancement Devices in ADVANCE: A Subgroup Analysis in Prosomnus Devices

Abstract Introduction Obstructive sleep apnea (OSA) is common and harmful, yet difficult to treat. Continuous positive airway pressure is often limited by adherence, while mandibular advancement devices (MAD) are more tolerable but work well only in some patients. Identifying mechanistic subgroups with the greatest likelihood of benefit could enable MAD as first-line therapy in selected individuals. Using pooled data from six studies, we recently developed an endotype-based prediction model. Because response may vary by device, this presentation examines outcomes specifically with Prosomnus devices (a sponsor of FLOSAT and Praxis). Methods The ADVANCE consortium (US, Belgium, Australia, France) pooled individual patient data from six MAD studies (PROMAD, OASSC, UGAMAD, MADOx, FLOSAT, Praxis) comprising 281 patients with baseline apnea-hypopnea index (AHI) > 10 events/hr. In the parent study, patients in the strong-response endotype—characterized by lower loop gain, higher AHI, younger age, lower BMI, and greater sleep depth—exhibited a 62 [54, 69] %reduction in AHI (N=55), stronger than the 37 [26, 48] %reduction (P=0.001) seen in the remaining patients (moderate-response endotype, N=35). Here we reassessed outcomes by subgroup in patients using Prosomnus devices. Results were presented for all patients (N=119, development+validation), and the validation subset (N=58; 33% of FLOSAT; plus Praxis, collected prospectively for validation). Linear regression compared the %reduction in AHI between groups. The primary test was the subgroup difference in the validation sample. Results Participant characteristics: 94M:25F, median[IQR] age=54[44-62] years, BMI=28[26-31] kg/m2, AHI=25[18-34] events/hr. In the overall sample, patients in the strong-response endotype (N=64) exhibited a 70% [64, 75] reduction in AHI, which was stronger than the 53% [46, 59] reduction in the moderate-response endotype (N=55). In validation: the strong-response endotype (N=29) exhibited a 68% [58, 76] reduction in AHI, which was stronger than 51% [39, 62] in the moderate-response endotype (N=29). The therapeutic difference was 17% [92.3, 31] (P=0.026), exceeding the minimal clinically-important difference (15%). Conclusion The strong-response endotype subgroup—characterized by lower loop gain, higher AHI, younger age, lower BMI, and greater sleep depth—achieved profound OSA improvements with Prosomnus MAD devices, knowledge which may facilitate a precision model of care employing MAD as a first-line intervention in selected patients. Support (if any) 5R01HL102321