Abstract
Objective: To determine the prevalence of sleep apnea (SA) during the first night after hemispheric ischemic stroke and its influence on clinical presentation, course, and functional outcome at 6 months.
Methods: The first night after cerebral infarction onset, 50 patients underwent polysomnography (PSG) followed by oximetry during the next 24 hours. Neurologic severity and early worsening were assessed by the Scandinavian Stroke Scale and outcome by the Barthel Index. Patients were evaluated on admission, on the third day, at discharge, and at 1, 3, and 6 months.
Results: There were 30 males and 20 females with a mean age of 66.8 ± 9.5 years. Latency between stroke onset and PSG was 11.6 ± 5.3 hours. Thirty-one (62%) subjects had SA (apnea–hypopnea index [AHI] ≥ 10). Of these, 23 (46%) had an AHI ≥20 and 21 (42%) an AHI ≥25. Sleep-related stroke onset occurred in 24 (48%) patients and was predicted only by an AHI ≥25 on logistic regression analysis. SA was related to early neurologic worsening and oxyhemoglobin desaturations but not to sleep history before stroke onset, infarct topography and size, neurologic severity, or functional outcome. Early neurologic worsening was found in 15 (30%) patients, and logistic regression analysis identified SA and serum glucose as its independent predictors.
Conclusions: SA is frequent during the first night after cerebral infarction (62%) and is associated with early neurologic worsening but not with functional outcome at 6 months. Cerebral infarction onset during sleep is associated with the presence of moderate to severe SA (AHI ≥ 25).
Sleep apnea (SA) is characterized by recurrent cessation of airflow, oxyhemoglobin desaturation, and sleep fragmentation and affects 2 to 5% of the population.1 Several risk factors for ischemic stroke, such as arterial hypertension, coronary heart disease, cardiac arrhythmias, obesity, and habitual snoring, are common among patients with SA.1,2⇓ Conversely, SA is frequent in patients with recent cerebral infarction,3-12⇓⇓⇓⇓⇓⇓⇓⇓⇓ and it has been suggested that hemodynamic disturbances such as decline in cerebral blood flow13 and episodes of hypotension and bradyarrhythmia14 that occur during the apneas may precipitate cerebral infarction onset during sleep and aggravate an existent ischemic damage. Thus, it can be speculated that treatment of SA would reduce the risk for cerebral infarction during sleep and also prevent early clinical worsening when the phenomenon of the ischemic penumbra is taking place.15 However, the influence of SA on early neurologic worsening and functional outcome is unknown since previous studies did not perform polysomnographic (PSG) or oximetric evaluation during the first hours after stroke onset,3-12⇓⇓⇓⇓⇓⇓⇓⇓⇓ the most critical period for irreversible ischemic tissue necrosis. The aims of this study were to determine the prevalence of SA during the first night after stroke onset in a group of patients with hemispheric ischemic stroke and to evaluate its possible association with sleep-related stroke onset, early neurologic worsening, and functional outcome at 6 months.
Patients and methods.
We prospectively studied 50 consecutive adult patients with a first-ever hemispheric ischemic stroke that included at least some motor impairment. The first night after stroke onset all patients underwent all-night PSG followed by continuous-pulse oximetry during the next 24 hours. Exclusion criteria were age <40 and >80 years, stupor or coma, fixed gaze deviation plus hemiplegia, seizures on stroke onset, baseline oxyhemoglobin saturation <95%, chronic obstructive pulmonary disease, acute heart failure, and presence of an illness with a life expectancy of <1 year. Patients entered into therapeutic clinical trials were also excluded. This study was approved by the ethics committee at our institution, and written informed consent was obtained from each patient or relatives.
Clinical and neuroradiologic evaluation.
On admission, all patients underwent a comprehensive history that included the prevalence of risk factors for stroke, physical and neurologic examination, axillary temperature, blood pressure measurement, electrocariogram, chest radiography, and standard blood tests. Body mass index and cervical perimeter were measured during admission. Transthoracic echocardiography, carotid Doppler ultrasonography, brain MR angiography, and cerebral arteriography were performed when appropriate. Sleep-related stroke onset was diagnosed in patients who first noticed their neurologic symptoms immediately upon awakening. Stroke subtype was classified according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria.16 Neurologic impairment was evaluated by the Scandinavian Stroke Scale17 (SSS) at admission, on the third day of hospitalization, at discharge, and at 1, 3, and 6 months after admission. Early neurologic worsening was defined as a decrease of at least 2 points in the SSS score from admission to the third day examination.18 Functional disability was assessed by the Barthel Index19 (BI) at discharge and at 1, 3, and 6 months after stroke onset and was classified as “good” when the BI score was between 60 and 100 and “poor” when the BI score was <60.20
On admission, all 50 patients underwent a brain CT scan within 6 hours from onset of symptoms, and the presence of early signs of infarction (focal hypodensity, cerebral edema, and hyperdensity of the middle cerebral artery) was evaluated. Brain MRI was performed in 47 patients during the first week after stroke onset (4 ± 2 days), and the infarct area was measured at the slice where the largest lesion was visible. The volume of the infarct was calculated manually by multiplying the areas where the lesion was visible by the thickness of the section (5 mm) and summing the value of each section. Location of the infarct was classified as cortical, subcortical, cortical–subcortical, or absent. Neuroimaging studies were assessed by a neuroradiologist blinded to the clinical and PSG data.
Sleep evaluation.
Patients, with the assistance of their bed partners, were interviewed by means of a comprehensive standardized sleep questionnaire covering their sleep habits and disturbances that were noticed before the stroke. Daytime sleepiness preceding the stroke was estimated by the Epworth Sleepiness Scale21 and was considered abnormal when the score was >10. PSG was performed at the hospital ward from 11:00 pm to 7:00 am without interfering with conventional care of the patient and included eight EEG channels (F3-A1, F4-A2, C3-A1, C4-A2, P3-A1, P4-A2, O1-A1, O2-A2), electro-oculogram, chin and left and right anterior tibial surface electromyogram, electrocardiogram, nasal and oral airflow, thoracic and abdominal movements, and oxyhemoglobin saturation, which was also recorded continuously by finger pulse oximetry during the next 24 hours. Sleep stages were scored according to standard criteria.22 Apnea was defined as the absence of airflow for at least 10 seconds. In obstructive apnea, respiratory effort was maintained, whereas in central apnea, breathing movements were absent. Mixed apnea was defined as a combination of central and obstructive apnea. Hypopnea was defined as a thoracoabdominal amplitude decrease >50% for at least 10 seconds with either an arousal or an oxygen desaturation >3%. Cheyne–Stokes respiration was defined as a periodic crescendo and decrescendo breathing pattern with central apnea or hypopnea. The apnea–hypopnea index (AHI) was defined as the average number of apneas and hypopneas per hour of sleep, and SA was defined by an AHI ≥107. The percentage of time with oxyhemoglobin saturation below 90% (CT90) was also evaluated.
Statistical analysis.
In univariate analyses, the χ2 test was used for categorical data and analysis of variance for continuous data in variables with normal distribution. Otherwise, nonparametric tests were used.
Forward stepwise logistic regression analyses were used to identify the independent predictors of SA, sleep-related stroke onset, early neurologic worsening, and functional outcome. The independent contribution of variables with a p value <0.1 on univariate analyses was assessed. Age, gender, body temperature on admission, and early signs of infarction on CT scan were forced into the model of early neurologic worsening. Multiple linear regression analysis was used to determine which variables were independently correlated with the AHI. Results are expressed as means ± SD, adjusted odds ratios (OR), and corresponding 95% CI. In all tests, a p value <0.05 was considered significant. All statistical analyses were done with the SPSS package (version 9.0 for Windows; Chicago, IL).
Results.
There were 30 men and 20 women with a mean age of 66.8 ± 9.5 (range 41 to 80) years. Subtypes of stroke were large-artery atherosclerosis in 17 (34%) patients, lacune in 15 (30%), cardioembolism in 13 (26%), and stroke of undetermined etiology in 5 (10%). Before stroke, 37 (74%) patients reported habitual snoring, 22 (44%) witnessed apneas, and 5 (10%) hypersomnia. None of the patients studied showed signs of dehydration, and blood tests on admission showed no evidence of renal dysfunction or electrolyte disturbances. On admission, no patient received medications such as hypnotics that could have induced SA or worsened the AHI. SA (AHI ≥ 10) was disclosed in 31 (62%) patients, sleep-related stroke onset in 24 (48%), and early worsening in 15 (30%). Poor functional outcome at discharge occurred in 54% of the patients, in 42% at 1 month after stroke onset, in 34% at 3 months, and in 28% at 6 months.
The mean latency between stroke onset and PSG was 11.6 ± 5.3 hours. The main PSG and oximetric findings are shown in table 1. PSG showed a mean AHI of 27.7 ± 26.6 (range 0 to 98). Thirty-one patients had an AHI ≥10. Of these, 8 had an AHI between 11 and 20, 2 an AHI between 21 and 25, 1 an AHI between 26 and 30, and 20 an AHI ≥30. Among the 31 patients with SA, apneas were predominantly obstructive in 15 (48.3%) subjects, central in 2 (6.5%), and a combination of central and obstructive in 5 (16.2%). Cheyne–Stokes respiration was detected in 9 (29%) patients in whom clinical examination and history did not disclose signs or symptoms suggestive of heart failure, and echocardiography showed a ventricular ejection fraction within normal values (>60%) in 6.
PSG and oximetric findings
Clinical differences between patients with and without AHI ≥10 on admission
Tables 2 and 3⇓ show the main differences between patients with and without SA. With use of logistic regression analysis, the independent predictors of SA were early neurologic worsening (OR, 9.7; 95% CI, 1.6 to 57.5) and oxyhemoglobin desaturations (CT90 > 5%) (OR, 1.5; 95% CI, 1.19 to 1.86). Although patients with SA were more often men and habitual snorers and had greater infarct size and cervical perimeter, differences did not reach the level of significance. Between patients with and without SA there were no differences in age, body mass index, previous history of witnessed apneas or hypersomnia, risk factors for stroke, systolic and diastolic blood pressure, hematocrit, temperature on admission, fasting glucose, early signs of infarction on CT, stroke location and subtype, neurologic impairment, or functional outcome during the 6 months of follow-up.
Radiologic, PSG, and clinical course differences between patients with and without AHI ≥10
Multiple linear regression analysis identified early neurologic worsening (p = 0.001), sleep-related stroke onset (p = 0.002), gender (men) (p = 0.001), and oxyhemoglobin desaturations (p < 0.001) as independent predictors of higher AHI, which accounted for 50.5% of the variance in the AHI.
Stroke was noticed on awakening in 24 patients (48%) and occurred during wakefulness in 26 (52%). Among the 24 patients with sleep-related stroke, symptoms were first noticed on awakening between 2:00 am and 7:00 am in 8 (33%) and between 7:01 am and 9:00 am in 16 (66%) subjects. None of the patients had stroke onset while napping during the daytime. With use of multiple logistic regression analysis, the only factor independently associated with sleep-related stroke onset was the AHI (OR, 1.02; 95% CI, 1.00 to 1.05). Sleep-related stroke onset was more frequent in patients with AHI ≥ 25 (p = 0.024). No other variables except for the AHI, including age, gender, sleep history, cardiovascular risk factors, body mass index, systolic and diastolic blood pressure, oxyhemoglobin desaturations, neuroradiologic findings, early neurologic worsening, neurologic impairment, and functional outcome, were associated with sleep-related stroke onset (table E4 is available on the Neurology Web site at www.neurology.org).
Univariate analyses showed that early neurologic worsening was associated with SA (p = 0.018), higher AHI (p = 0.011), higher level of fasting serum glucose (p = 0.010), and larger infarct volume (p = 0.042), although by logistic regression analysis, only SA (OR, 8.2; 95% CI, 1.3 to 51.2) and glucose (OR, 1.01; 95% CI, 1.00 to 1.02) remained independently correlated with early neurologic worsening. Early neurologic worsening was correlated with more severe impairment and worse functional outcome both at discharge (p < 0.001) and at 1 month after admission (p = 0.02), but this relation was not maintained after 3 and 6 months (table E5 is available on the Neurology Web site at www.neurology.org).
During the 6 months of follow-up, poor functional outcome was not associated with SA (see table 3). On logistic regression analysis, the independent predictors of poor functional outcome at discharge were neurologic impairment at admission (OR, 1.13; 95% CI, 1.03 to 1.24), early neurologic worsening (OR, 1.8; 95% CI, 1.21 to 3.23), and diabetes mellitus (OR, 1.13; 95% CI, 1.04 to 1.22). Evaluations at 1, 3, and 6 months after admission did not identify early neurologic worsening as an independent predictor of poor functional outcome on logistic regression analysis. At 6 months after stroke onset, patients with poor functional outcome had more severe neurologic impairment on admission (p < 0.001), larger infarct volumes measured on MRI within the first week of hospitalization (p < 0.001), early signs of infarction on admission brain CT (p = 0.002), and higher serum glucose levels (p = 0.003), although according to logistic regression analysis, the only independent predictors of poor functional outcome at 6 months were neurologic impairment on admission (OR, 1.08; 95% CI, 1.01 to 1.16) and volume of the infarct (OR, 1.00; 95% CI, 0.97 to 1.02).
Discussion.
Our study shows that in cerebral infarction, the presence of recurrent apneic events during the first night after stroke onset is frequent and is associated with stroke onset during sleep and early neurologic worsening but not with sleep history before stroke, initial severity, stroke size and location, and functional outcome at 6 months.
To evaluate a homogeneous group of patients, we studied only adult patients with hemispheric infarctions and excluded infratentorial infarctions, hemorrhagic strokes, and TIA. Previous investigations performed the sleep studies several days7-11⇓⇓⇓⇓ or weeks5-10,12⇓⇓⇓⇓⇓⇓ after stroke onset and grouped together stroke with TIA,7,9-11⇓⇓⇓ hemorrhagic with ischemic events,6,11,12⇓⇓ and hemispheric with infratentorial strokes.5-12⇓⇓⇓⇓⇓⇓⇓ However, the frequency of SA found in our study is similar to the frequency reported in TIA7,9-11⇓⇓⇓ and other stroke subtypes5-12⇓⇓⇓⇓⇓⇓⇓ and is much higher than in healthy elderly subjects and the general population.1 Thus, the available data suggest that SA existed prior to the cerebral infarction onset and is not always a consequence of it. However, since no patient had sleep studies performed before cerebral infarction onset, there is no way of knowing the exact frequency of SA before stroke, and this should be interpreted with caution.
In this study, the only variable that predicted sleep-related stroke onset was an AHI of ≥25. This finding suggests that repetitive apneic episodes may represent a risk factor for hemispheric ischemic stroke onset during sleep. The exact mechanisms by which recurrent apneas may precipitate cerebral infarction are not known but might be related to several hemodynamic disturbances that occur during or shortly after the apneic events, such as decline in cerebral blood flow,13 diminished cerebral vasodilator reserve,23 failure of cerebral autoregulation,24 decreased cardiac output,25 episodes of hypotension and bradyarrhythmia,14 increased platelet aggregability,26 decreased fibrinolytic activity,27 and increased sympathetic activity.28 In our study, the finding that oxyhemoglobin desaturation was not associated with sleep-related stroke or with early clinical deterioration might suggest that the hypoxemia induced by apneas does not seem to have a major effect on the development of hemispheric infarction.
It has been reported that night-time and daytime stroke onset may not differ in the AHI.9 This observation does not contradict our data when night-time stroke onset is defined as the period between midnight and 6:00 am and does not differentiate between stroke occurring while awake or asleep, as we did. The incidence of daytime stroke onset may be overestimated in patients who first noticed their neurologic deficits immediately on awakening between 6:00 am and 9:00 am, when presumably stroke occurred during sleep.
The most frequent symptoms of SA syndrome are habitual snoring, apneic episodes observed by the bed partner, and daytime somnolence.1 Our study, however, shows that although 90% of the patients with an AHI ≥10 were habitual snorers, sleepiness was reported by only 12.9% and witnessed apneas by 41.9%. Furthermore, there were no significant differences between patients with and without SA regarding habitual snoring, witnessed apneas, and sleepiness. These data confirm the previous finding7 that in patients with both stroke and SA, sleep history before stroke onset is not highly suggestive of SA. Since SA is related to cerebral infarction during sleep and early neurologic worsening, it might be important to exclude and treat SA in subjects with habitual snoring plus other known risk factors for stroke despite absence of somnolence and observed apneas. As nasal continuous positive airway pressure (CPAP) eliminates all types of apneic events (obstructive, central, mixed, Cheyne–Stokes pattern, and hypopneas)29 and may reduce morbidity and mortality due to cardiovascular disease,29-31⇓⇓ CPAP treatment in subjects with SA would reduce the risk for ischemic stroke onset during sleep.
We found that early neurologic worsening was predicted by SA and was associated with poor outcome at discharge. Since early neurologic worsening reflects the neuronal damage of the penumbra area that surrounds the central ischemic core,15 early identification and treatment of SA might help to prevent clinical deterioration and presumably to improve short-term outcome. Based on our data, however, SA did not predict a worse outcome within 6 months after stroke, although we observed a tendency in patients with SA to have a poorer outcome during the follow-up. We do not know the explanation for these paradoxical results. One possibility is that the study sample was too small to detect a negative influence of SA on long-term outcome. On the other hand, it can also be speculated that the severity of SA that our patients had was sufficient to produce a transient clinical deterioration during the acute phase of the stroke but not severe enough to produce a permanent impairment. Treatment trials of SA during the acute phase of cerebral infarction may be needed to determine whether correction of this sleep-disordered breathing improves outcome.
Acknowledgments
Supported by a grant from Carburos Metalicos ($35,000) and the Fondo de Investigaciones Sanitarias of Spain (FISS 97-1088).
Acknowledgment
The authors thank the sleep laboratory technicians (A. Arqueros, E. Iñiguez, and J. Port) for the PSG recordings and E. Morell for the statistical analyses.
Footnotes
-
Additional material related to this article can be found on the Neurology Web site. Go to www.neurology.org and scroll down the Table of Contents for the March 26 issue to find the title link for this article.
- Received June 25, 2001.
- Accepted December 1, 2001.
References
Disputes & Debates: Rapid online correspondence
- Reply to Letter to the Editor
- Prevalence and clinical importance of sleep apnea in the first night after cerebral infarction
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.