Prevalence of sleep-disordered breathing after stroke and TIA: A meta-analysis.

Abstract

To perform a systematic review and meta-analysis on the prevalence of sleep-disordered breathing (SDB) after stroke. We searched PubMed, Embase (Ovid), the Cochrane Library, and CINAHL (from their commencements to April 7, 2017) for clinical studies reporting prevalence and/or severity of SDB after stroke or TIA. Only sleep apnea tests performed with full polysomnography and diagnostic devices of the American Academy of Sleep Medicine categories I-IV were included. We conducted random-effects meta-analysis. PROSPERO registration number: CRD42017072339. The initial search identified 5,211 publications. Eighty-nine studies (including 7,096 patients) met inclusion criteria. Fifty-four studies were performed in the acute phase after stroke (after less than 1 month), 23 studies in the subacute phase (after 1-3 months), and 12 studies in the chronic phase (after more than 3 months). Mean apnea-hypopnea index was 26.0/h (SD 21.7-31.2). Prevalence of SDB with apnea-hypopnea index greater than 5/h and greater than 30/h was found in 71% (95% confidence interval 66.6%-74.8%) and 30% (95% confidence interval 24.4%-35.5%) of patients, respectively. Severity and prevalence of SDB were similar in all examined phases after stroke, irrespective of the type of sleep apnea test performed. Heterogeneity between studies (I 2) was mostly high. The high prevalence of SDB after stroke and TIA, which persists over time, is important in light of recent studies reporting the (1) feasibility and (2) efficacy of SDB treatment in this clinical setting.