SummaryBackground Obstructive sleep apnoea (OSA) is often underdiagnosed, highlighting the need for scalable diagnostic alternatives. The SUNSAS study compared a new device for at-home diagnosis of OSA (artificial intelligence AI-supported analysis of mandibular jaw movements MJM) with polysomnography (PSG) for time to diagnosis and treatment, and patient-reported outcomes. Methods This prospective, multicentre, randomised, controlled, open-label study was conducted in France (October 2021–October 2024). Adults aged 18–80 years with suspected OSA were randomised (1:1) to undergo diagnostic testing using MJM monitoring (Sunrise) or PSG. Primary endpoints were assessed using hierarchical testing: 1. daytime sleepiness (Epworth Sleepiness Scale ESS score) at 3 months post-diagnosis and time to diagnosis; 2. time to treatment; and 3. daytime sleepiness at 3 months post-randomisation. Secondary endpoints included quality of life (Short Form-36, Quebec Sleep Questionnaire), work productivity (Work Productivity and Activity Impairment questionnaire), and positive airway pressure therapy adherence at 3 months after treatment initiation. Findings Of 849 participants randomised (58·7% male, median age 50 years, body mass index 28·0 kg/m2, apnoea-hypopnoea index 15·2/h), 774 received a diagnosis: 133 no OSA, 239 mild OSA, 220 moderate OSA, and 182 severe OSA. Median time to diagnosis (15 vs. 106 days) and to treatment initiation (50 vs. 124 days) were significantly shorter with MJM analysis versus PSG (both p Interpretation OSA diagnosis based on MJM monitoring with AI-supported analysis is noninferior to PSG in reducing daytime sleepiness at 3 months after diagnosis, while significantly accelerating time to diagnosis and treatment initiation, resulting in earlier improvement in daytime sleepiness. Funding Sunrise, with support from the French Ministry of Health through the Forfait Innovation programme.
Pépin et al. (2026) studied this question.