Estimating the prevalence of sleep-disordered breathing in community-based, long-term stroke survivors using a validated predictive model.

Abstract

Assessing sleep-disordered breathing (SDB) in aged or disabled populations is difficult.To validate a multivariable apnea risk prediction index (MAPI) in stroke survivors and estimate SDB prevalence (apnea-hypopnea index >or=10) in a community-based stroke cohort.Self-reported sleep apnea symptoms, demographic and anthropometric data were obtained from a hospital-based (SCOPES II, n = 152) and a community-based (NEMESIS, n = 431) cohort at about 3 years after stroke. Logistic regression models for prediction of SDB using the MAPI were validated with home-based nocturnal polygraphic data recordings obtained in 74 SCOPES II participants.NEMESIS subjects (median age 76, females 45%) were older than SCOPES II subjects (median age 70, females 42%). SDB prevalence was 44.9% in the nocturnal polygraphy subset, and estimated as 44.1% (SCOPES II) and 50.6% (NEMESIS) using a modified MAPI.The MAPI can be reliably used in stroke cohorts and SDB affects about half of chronic stroke survivors.