Background and purpose: We objectively evaluated patients with recent stroke to determine the prevalence of sleep-disordered breathing (SDB) and whether SDB was associated with unfavorable clinical outcomes.
Methods: Forty-seven patients with recent ischemic stroke (median, 13 days) were studied with computerized overnight oximetry for evidence of arterial oxyhemoglobin desaturation (SaO2). Polysomnography was also performed on 19 patients. Medical history, sleep history, location of stroke, and severity of neurological deficit were recorded, and patients were observed by staff for evidence of snoring and excessive daytime sleepiness. Functional abilities were measured with the use of the Barthel Index (BI). Outcome variables included ability to return home at discharge, continued residence at home at 3 and 12 months, BI at discharge, BI at 3 and 12 months, and death from any cause at 12 months.
Results: Mean SaO2 during oximetry was 94.0 +/- 1.7%, and percentage of recording time spent at < 90% SaO2 was 4.3 +/- 5.7%. The number of desaturation events per hour of recording time (desaturation index [DI]) was 9.5 +/- 9.67, with 15 of 47 (32%) having DI > 10 and 6 of 47 (13%) having DI > 20. Oximetry measures of SDB correlated with lower BI scores at discharge and lower BI at 3- and 12-month follow-ups (P < or = .05, Pearson coefficients). Oximetry measures correlated with return home after discharge, but the association between oximetry measures and living at home was lost at 12 months. Two oximetry variables correlated with death at 1 year. Brain stem location correlated with higher DI and time at < 90% SaO2, but patients with hemispheric stroke and oximetry abnormalities also had worse functional outcome. No correlation was found between oximetry values and sex, age, preexisting medical conditions (except previous stroke), or severity of neurological deficit. Oximetry abnormalities were associated with a history of snoring. Polysomnography on 19 patients confirmed oximetry evidence of severe SDB. Eighteen of 19 patients (95%) had an apnea-hypopnea index (AHI) of > 10 events per hour of recording, 13 of 19 (68%) had an AHI > 20, and 10 of 19 (53%) had an AHI > 30. Desaturation events were largely due to obstructive apneas.
Conclusions: SDB accompanied by arterial oxyhemoglobin desaturation is common in patients undergoing rehabilitation after stroke and is associated with higher mortality at 1 year and lower BI scores at discharge and at 3 and 12 months after stroke. SDB may be an independent predictor of worse functional outcome. Obstructive sleep apnea appeared to be the most common form of SDB, and the frequent history of snoring suggests that SDB preceded the stroke in most patients.