Abstract
Background: We describe a new digital medicine solution (DMS) to identify relevant clinical phenotypes of obstructive sleep apnea syndrome (OSA), by combining self-reported clinical data and indices provided by mandibular movements (MM) analysis supported by artificial intelligence (AI).
Method: 1117 consecutive adults (18 to 89 yrs old) with OSA suspicion were included. Input data for phenotyping were self-reported symptoms, anthropometric/demographic data from on-line questionnaire and 5 metrics from a connected home sleep device (Sunrise, Namur, Belgium) using an AI-based analysis of MM signal (Fig A).
Results: Using k-prototype clustering method, we identified from DMS 4 distinct clinical phenotypes: C1 to C4. These clusters were then compared to OSA severity evaluated by polysomnography (PSG)(Fig B). Severity revealed by DMS was consistent to PSG OSA severity: from C1 to C4, mean apnea-hypopnea index = 6.3, 15.6, 37.5 and 72.1; mean O2 desaturation index = 3.9, 13.8, 37.3 and 69.8, respectively. The clusters C2, C3 and C4 were associated with higher risks of cardiovascular comorbidities (OR = 4.5 to 7.8 for hypertension, 4.5 to 6. 9 for diabetes, 6.8 to 18.4 for heart failure and 3.1 to 5.4 for cerebrovascular stroke; all p values < 10-6) (Fig C).
Conclusion: AI-based home sleep testing of MM combined to on-line questionnaire provides easy patient assignment to relevant clinical phenotype and dedicated care.
Footnotes
Cite this article as: European Respiratory Journal 2021; 58: Suppl. 65, OA3017.
This abstract was presented at the 2021 ERS International Congress, in session “Prediction of exacerbations in patients with COPD”.
This is an ERS International Congress abstract. No full-text version is available. Further material to accompany this abstract may be available at www.ers-education.org (ERS member access only).
- Copyright ©the authors 2021
