Clusters of sleep apnoea phenotypes: A large pan‐European study from the European Sleep Apnoea Database (ESADA)

• Published in: Respirology (Carlton, Vic.) Vol. 26; no. 4; pp. 378 - 387
• Main Authors: Bailly, Sébastien, Grote, Ludger, Hedner, Jan, Schiza, Sofia, McNicholas, Walter T., Basoglu, Ozen K., Lombardi, Carolina, Dogas, Zoran, Roisman, Gabriel, Pataka, Athanasia, Bonsignore, Maria R, Pepin, Jean‐Louis, Steiropoulos, P, Verbraecken, J, Georgia, Trakada, Penzel, T, Ondrej, Ludka, Bouloukaki, I, Ryan, S, Riha, RL, Kvamme, JA, Tamisier, R, Hein, H, Tasbakan, MS, Joppa, P, Staats, R, Parati, G, Marta, Drummond, Marrone, O, Petitjean, M, Galic, T, Saaresranta, T, Sliwinski, P, Haralampos, Gouveris, Stefan, Mihaicuta, Francesco, Fanfulla, Winfried, Randerath, Dries, Testelmans, Buskova, J
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 01.04.2021; Wiley Subscription Services, Inc; Wiley
• Subjects: Apnea; Applications; Cluster analysis; Gender; Geographical variations; Human health and pathology; international database; Life Sciences; Neurologi; Neurology; Patients; personalized medicine; Phenotypes; Sleep; Sleep apnea; sleep apnoea; Sleep disorders; Statistics; international database; phenotypes; sleep apnoea; cluster analysis; personalized medicine
• ISSN: 1323-7799; 1440-1843
• DOI: 10.1111/resp.13969

ABSTRACT Background and objective To personalize OSA management, several studies have attempted to better capture disease heterogeneity by clustering methods. The aim of this study was to conduct a cluster analysis of 23 000 OSA patients at diagnosis using the multinational ESADA. Methods Data from 34 centres contributing to ESADA were used. An LCA was applied to identify OSA phenotypes in this European population representing broad geographical variations. Many variables, including symptoms, comorbidities and polysomnographic data, were included. Prescribed medications were classified according to the ATC classification and this information was used for comorbidity confirmation. Results Eight clusters were identified. Four clusters were gender‐based corresponding to 54% of patients, with two clusters consisting only of men and two clusters only of women. The remaining four clusters were mainly men with various combinations of age range, BMI, AHI and comorbidities. The preferred type of OSA treatment (PAP or mandibular advancement) varied between clusters. Conclusion Eight distinct clinical OSA phenotypes were identified in a large pan‐European database highlighting the importance of gender‐based phenotypes and the impact of these subtypes on treatment prescription. The impact of cluster on long‐term treatment adherence and prognosis remains to be studied using the ESADA follow‐up data set. A cluster analysis of a pan‐European database of over 23 000 OSA patients found eight distinct clinical phenotypes, with clear gender‐based differences and implications for more personalized treatment prescriptions. See related Editorial