0577 Cluster Analysis for the Association Between Obstructive Sleep Apnea Phenotypes: A Population-Based Longitudinal Study. (27th May 2020)

• Record Type: Journal Article
• Authors: Tempaku, P F; Silva, L O; Guimaraes, T M; Vidigal, T A; D'Almeida, V; Andersen, M L; Bittencourt, L; Tufik, S
• Is Part Of: Sleep. Volume 43(2020)Supplement 1
• Page Start: A221
• Page End: A221
• Publication Date: 2020-05-27
• Subjects: Sleep -- Physiological aspects -- Periodicals; Sleep disorders -- Periodicals; Sommeil -- Aspect physiologique -- Périodiques; Sommeil, Troubles du -- Périodiques; Sleep disorders; Sleep -- Physiological aspects; Sleep -- physiological aspects; Sleep Wake Disorders; Psychophysiology; Electronic journals; Periodicals; 616.8498
• DOI: 10.1093/sleep/zsaa056.574
• Languages: English
• ISSNs: 0161-8105

Abstract

Abstract: Introduction: The identification of subgroups of obstructive sleep apnea (OSA) is critical to understand disease causality and ultimately develop optimal care strategies customized for each subgroup. In this sense, we aimed to perform a cluster analysis to identify subgroups of individuals with OSA based on clinical parameters. Furthermore, we aimed to analyze whether subgroups remain after 8 years. Methods: We used data derived from the Sao Paulo Epidemiologic Sleep Study (EPISONO) cohort, which was followed over 8 years. All individuals underwent polysomnography, answered questionnaires and had their blood collected for biochemical exams. OSA was defined according to an AHI equal or greater than 15 events per hour. Cluster analysis was performed using latent class analysis (LCA). Results: Of the 1, 042 individuals in the EPISONO baseline cohort, 68.3% accepted to participate in the follow-up study (n=712). We were able to replicate the OSA 3-cluster solution observed in previous studies: disturbed sleep, minimally symptomatic and excessively sleepy in both baseline (35.5%, 45.4% and 19.1%, respectively) and follow-up studies (41.9%, 43.4% and 14.8%, respectively). 44.8% of the participants migrated clusters between the two evaluations and the factor associated with this was a greater delta-AHI (B=-0.033, df=1, p=0.003). The optimal cluster solution for our sample based on Bayesian information criterion (BIC) was 2 clusters for baseline (disturbed sleep and excessively sleepy) and 3 clusters for follow-up (disturbed sleep, minimally symptomatic and excessively sleepy). Conclusion: The results found replicate and confirm previously identified clinical clusters in OSA even in a longitudinal analysis. Support: This work was supported by grants from AFIP, FAPESP and CAPES.