Sleep-Disordered Breathing as Risk Factor for Stroke

Habitual snoring—that is, snoring occurring always or almost always—affects 4% to 24% of the adult population, with a maximum prevalence around the age of 50 to 60 years, and is strongly associated with OSA.¹⁶ Five case-control and three cohort studies have shown that habitual snoring represents an independent risk factor for stroke, with a pooled risk of about 1.5.¹⁷

Longitudinal and population studies have suggested a direct link between polysomnographically proven SDB and stroke. In a study of 1022 patients admitted to a single sleep center and followed up over up to 6 years, Yaggi and colleagues found that OSA is associated with an increased risk for stroke and death, even after adjusting for cardiovascular risk factors, such as arterial hypertension, diabetes mellitus, atrial fibrillation, hyperlipidemia. and smoking status, with an hazard ratio (HR) of 2.0 (95% confidence interval [CI], 1.12 to 3.48). The risk was even higher (HR, 3.3; 95% CI, 1.74 to 6.26) in patients with severe OSA (apnea–hypopnea index [AHI] > 36/hr).¹⁸

In a single-center study of 1387 male patients with OSA, 377 simple snorers, and 264 controls, matched for age and body mass index (BMI), Marin and coworkers found, after a mean follow-up period of 10 years, a significantly higher incidence of fatal and nonfatal cardiovascular events (e.g., stroke, myocardial infarction, coronary artery bypass graft surgery, percutaneous transluminal coronary angiography) in patients with severe OSA (AHI > 30/hr) as compared to patients with mild to moderate OSA, patients with OSA of any severity treated with continuous positive airway pressure (CPAP), simple snorers, and controls.¹⁹ Multivariate statistics, adjusting for potential confounders (hypertension, diabetes, cardiovascular disease, lipid disorders, smoking status) showed that untreated severe OSA increased the risk of fatal and nonfatal cardiovascular events compared with healthy participants.

In a cross-sectional analysis of 1475 subjects, Arzt and colleagues showed that those with an AHI of at least 20/hr have an odds ratio (OR) of 4.3 (95% CI, 1.32 to 15.24) compared to those with an AHI less than 5/hr, even after adjustment for known risk factors.²⁰

In an elderly population of 394 noninstitutionalized stroke-free subjects (>70 years), Munoz’s group showed after a follow-up of 6 years that persons with severe SDB (AHI > 30/hr) have an increased risk for stroke with a hazard ratio of 2.5 (95% CI, 1.04 to 6.01) even when data are adjusted for sex.²¹

A link between SDB and silent stroke or white matter disease on magnetic resonance imaging (MRI) has also been examined.^22–25 In a study of 50 OSA patients, Minoguchi and colleagues found an increased risk for silent brain infarction (as detected by MRI) for those with an AHI of at least 30 when compared with weight-matched patients with mild or no OSA.²⁵

Sleep-Disordered Breathing in Stroke Patients

More than 20 studies (Table 88-1) on almost 2000 patients have shown, despite different diagnostic approaches (polysomnography or respirography), that at least 50% of stroke patients exhibit SDB, as defined by an AHI of 10/hr or more.^26–49 Few studies suggested a higher incidence of SDB depending upon type (hemorrhagic versus ischemic), presumed etiology (macroangiopathy versus other origins), and severity or clinical characteristics (facial or bulbar involvement, recurrent versus first-ever event) of stroke. In seven studies, the prevalence of SDB was found to be similar also in patients with TIAs and ischemic stroke (see Table 88-1).^{26,27,31,38,44,51a,51b}

In a 2010 metaanalysis of 29 articles (2343 ischemic or hemorrhagic stroke and TIA patients), the frequency of SDB with AHI > 5 was 72% and with AHI > 20 was 38%. Only 7% of the SDB was primarily central apnea. There was no significant difference in SDB prevalence by event type, timing after stroke, or type of monitoring. Males had a higher percentage of SDB (AHI > 10) than females (65% compared to 48%, P = 0.001). Patients with recurrent strokes had a higher percentage of SDB (AHI > 10) than initial strokes (74% compared to 57%, P = 0.013). Patients with unknown etiology of stroke had a higher and cardioembolic etiology a lower percentage of SDB than patients with other etiologies.^51c

From the acute to the subacute phase of stroke, SDB improves.^{31,35–37,43} Nevertheless, a significant percentage of patients (up to 50%) exhibit an AHI of at least 10/hr months to years after the acute event (see Table 88-1). Central events improve more than obstructive events. One study suggested a better improvement of SDB in hemorrhagic strokes as compared to ischemic strokes.³⁷

Symptoms

Sleep-disordered breathing can manifest with a variety of symptoms and signs that are sometimes attributed to the underlying brain damage. Nighttime symptoms of SDB include difficulties falling asleep (sleep-onset insomnia, see later); respiratory noises (snoring, stridor); irregular or periodic respiration, apneas; disrupted sleep with increased motor activity and frequent awakenings; sudden awakenings with or without choking sensations, shortness of breath, palpitations, and panic attacks; orthopnea; and increased sweating. In patients with severe hypoventilation, arousal responses can be suppressed by increasing sleep debt and can lead to death during sleep. Daytime symptoms of SDB include headaches, fatigue and excessive daytime sleepiness, concentration and memory difficulties, irritability, and depression. Some patients also exhibit breathing irregularities during wakefulness including dyspnea, apneas, inspiratory breath holding (apneustic breathing), irregular breathing, rapid shallow breathing (hyperpnea and central hyperventilation), hiccup, and other breathing abnormalities.

Breathing Disturbances during Sleep

The most common form of SDB in stroke patients is OSA (Fig. 88-1). Severe OSA (with AHI 20 to 30) is found in 20% to 30% of patients. Occasionally, patients present with both OSA and Cheyne-Stokes breathing, with predominance of the first in rapid eye movement (REM) sleep and of the latter in light non-REM (NREM) sleep.²⁶

Figure 88-1 Obstructive sleep apnea in acute ischemic stroke. This 70-year-old man (HRS) has left middle cerebral artery stroke, carotid artery occlusion, and atrial fibrillation. He has habitual snoring but no excessive daytime sleepiness on history. Aphasia and severe hemiparesis are clinically apparent. National Institutes of Health (NIH) stroke score is16, and there are no signs of heart failure. Polysomnography 2 days after stroke onset shows an apnea–hypopnea index (AHI) of 79 (33 obstructive, 42 mixed) and minimal oxygen desaturation of 85%. The AHI is normalized (<5/hr) with continuous positive airway pressure (CPAP).

(MRI pictures courtesy of Prof. A. Valavanis, Institute of Neuroradiology, University Hospital, Zürich, Switzerland.)

Cheyne-Stokes breathing is a type of periodic breathing in which central apneas and hypopneas are separated by a crescendo–decrescendo respiratory pattern (Fig. 88-2). Central periodic breathing can be defined as three or more cycles of regular crescendo–decrescendo breathing associated with a reduction of at least 50% in nasal airflow and effort lasting at least 10 seconds. In the first few days after stroke, Cheyne-Stokes breathing and central periodic breathing during at least 10% of recording time is present in 10% to 40% of patients^{31,33,48,52,53}

Figure 88-2 Central sleep apnea in acute ischemic stroke. This 63-year-old man (PD) had a left subcortical stroke of unknown origin, with arterial hypertension and habitual snoring but no excessive daytime sleepiness on history. He had clinically mild hemiparesis, with a National Institutes of Health (NIH) Stroke Score of 8 and no signs of heart failure (cardiac ejection fraction, 55%). Polysomnography the first night after stroke onset showed an apnea–hypopnea index (AHI) of 53 (mainly central). The patient spontaneously improved after 1 week (AHI, 16).

(MRI pictures courtesy of Prof. G. Schroth, Institute of Neuroradiology, University Hospital, Bern, Switzerland.)

Central hypoventilation, failure of automatic breathing (Ondine’s curse) and other neurogenic breathing disturbances are less common and usually associated with brainstem strokes.

Breathing Disturbances while Awake

Breathing abnormalities occurring during wakefulness following hemispheric strokes have not been well studied. Such abnormalities can include selective impairment of behavioral or volitional respiratory control with preservation of metabolic control mechanisms. Strokes in the frontal cortex, basal ganglia, or capsula interna can cause respiratory apraxia, with impairment of voluntary modulation of breathing amplitude and frequency, leaving patients unable to take a deep breath or hold the breath.^54,55

Several abnormal patterns of breathing, though uncommon, can occur after brainstem stroke, especially during sleep or with decreased levels of consciousness. Sustained respiratory rates above 25 to 30/min in the absence of hypoxemia (neurogenic hyperventilation) were originally described in six comatose patients with ventrotegmental pontine strokes,⁵⁶ but these rates were subsequently attributed to stimulation of lung and chest wall afferent reflexes due to pulmonary congestion.⁵⁷ Neurogenic hyperventilation after stroke usually indicates a poor prognosis.^58,59

Inspiratory breath-holding (apneustic breathing), originally described in two patients with bilateral ventrotegmental mediocaudal (infratrigeminal) pontine stroke,⁶⁰ is rare and usually secondary to basilar artery occlusion. Erratic variations in breathing frequency and amplitude (ataxic or Biot’s breathing), and failure of automatic breathing (central sleep apnea or Ondine’s curse), usually imply a lateral medullary stroke, often bilateral.^61–64 Damage to the medullary reticular formation and nucleus ambiguus may be sufficient to cause a loss of automatic breathing, and a lesion that also includes the nucleus tractus solitarius is necessary to cause failure of both automatic and voluntary respiration.⁶³ Volitional breathing can be impaired by brainstem strokes involving corticobulbar and corticospinal pathways at pontine and medullary levels.⁵⁵ Repetitive yawning can accompany hypersomnia (e.g., in patients with thalamic stroke) and can also occur as a release phenomenon after extensive corticospinal lesions (e.g., in patients with a locked-in syndrome.⁶⁵

Spinal cord stroke can impair both automatic and voluntary breathing control.^54,66 The topography and extent of the lesion can determine respiratory effects. Anterior spinal artery strokes can affect reticulospinal pathways, located anteriorly in the lateral columns of the first three cervical segments, which are crucial for automatic breathing.⁶⁶ In contrast, posterior spinal artery strokes can damage corticospinal pathways in the dorsolateral spinal cord and impair voluntary control of breathing.⁶⁷ Strokes that extend up to the C1 level usually cause severe respiratory insufficiency and necessitate ventilatory support.

Sleep-Disordered Breathing as a Risk Factor for Stroke

Several studies have suggested a variety of physiologic mechanisms by which obstructive sleep-disordered breathing could predispose to stroke. Chronically, SDB and even habitual snoring are associated with hypertension, which in turn is an important risk factor for stroke. Findings from the Wisconsin Sleep Cohort have shown that an AHI greater than 15 is independently associated with a threefold increased risk of developing new hypertension within a 4-year period.⁶⁸ Furthermore, SDB is associated with coronary heart disease, myocardial infarction, heart failure, and atrial fibrillation, all of which can cause stroke.^69–71 The consequences of nocturnal respiratory events include intermittent hypoxemia, hypercapnia, sleep fragmentation (occasionally with loss of slow-wave and REM sleep), intrathoracic or intracranial pressure changes, and sympathetic activation. A series of hemodynamic (e.g. blood pressure swings), neural (e.g. sympathetic hyperactivity), metabolic (e.g. insulin resistance), and inflammatory or oxidative changes (e.g., increased levels of C-reactive protein, tumor necrosis factor [TNF]-α, interleukin [IL]-6, adhesion molecules) result, contributing to diabetes, increased platelet aggregation, decreased fibrinolysis, endothelial damage, and atherogenesis.^69,72,73 Patients with acute stroke also show an increase in some of the molecular markers and consequences of SDB such as elevated levels of fibrinogen and C-reactive protein.^74,75 The observation of an increase in the intima-media thickness of the common carotid artery documented by ultrasound techniques in SDB patients as compared with controls matched for age and vascular risk factors gives further support for an increased (chronic) atherogenesis in SDB.^76,77

Acutely, apneas and hypopneas during sleep are accompanied by decreased cardiac output, cardiac arrhythmias, systemic hypotension or hypertension, vasodilation due to hypoxia and hypercapnia, and increased intracranial pressure.⁷⁸ These factors lead to an overall 15% to 20% (in single studies, up to 50%) reduction in cerebral blood flow velocities during respiratory events^79–82 (Fig. 88-3). Type, duration, and timing of respiratory events affect hemodynamic consequences. Obstructive apneas of long duration and those that occur during REM sleep may be particularly detrimental.^81,83,84 Large fluctuations in cerebral blood velocities (and flow) may be particularly detrimental because patients with SDB have been shown to have a diminished vasodilator reserve.⁸² Cheyne-Stokes respiration and central apneas also can alter cerebral blood flow.^81,85,86 Paradoxical embolization resulting from right-to-left shunting in patients with patent foramen ovale during long apneas is another potential mechanism of stroke.^87,88 In view of these observations, it is not surprising that SDB has been associated with the onset of cerebrovascular events at night.^32,43 A direct link between SDB and nocturnal cerebrovascular ischemic events has been suggested for retinal infarcts of embolic origin and in single patients with TIA or minor stroke.^89–91

Figure 88-3 Cerebral blood flow velocity and obstructive sleep apnea. Flow velocity (FV) and mean flow velocity (MFV) in the right middle cerebral artery, peripheral blood pressure (BP), and end-expiratory PCO₂ in a patient with obstructive sleep apnea. Intervals with low end-expiratory PCO₂ show apneic episodes.

(From Hajak G, Klingelhöfer J, Schulz-Varszegi M, et al. Sleep apnea syndrome and cerebral hemodynamics. Chest 1996;110:670-679.)

A pathogenic role of SDB in vascular disease is further suggested from studies of continuous positive airway pressure (CPAP) treatment. A few studies have shown favorable, although modest, effects of CPAP on mean arterial pressure (MAP).⁹² Pepperell and coworkers reported that therapeutic CPAP reduced MAP by 2.5 mm Hg, whereas subtherapeutic CPAP levels increased blood pressure by 0.8 mm Hg. Such an effect can lead to a reduction in stroke risk of about 20%.⁹³ Treatment with CPAP can also reverse the detrimental hemodynamic, neural, and molecular effects of SDB-influenced surrogate markers of vascular disease such as factor VII clotting activity, fibrinogen levels, and platelet activation or aggregation.^25,72,73,94

Sleep-Disordered Breathing as a Consequence of Stroke

Although patients at risk for cerebrovascular disease often have SDB before they experience a stroke, some develop the sleep disorder as a consequence of stroke. The observation that recovery from stroke may be accompanied by improvement of SDB gives support to the assumption that SDB is aggravated or even appears de novo after stroke (see earlier).

A disturbed coordination of upper airway, intercostal, and diaphragmatic muscles due to brainstem or hemispheric lesions (bulbar and pseudobulbar palsies) might favor the appearance of OSA.^95,96 In patients with rostrolateral medullary lesions and SDB, a reduced ventilatory sensitivity to inhaled CO₂ can play a role.⁹⁷ Other factors, including hypoxemia due to aspiration or respiratory infections, reduction of voluntary chest movements on the paralyzed side, supine position, and sleep fragmentation secondary to stroke or stroke complications, can also be involved.^45,98

The presence of Cheyne-Stokes breathing has been attributed to CO₂-hypersensitivity induced by bilateral supratentorial lesions.⁹⁹ The presence of bilateral or severe strokes, heart failure, and a decreased level of consciousness, traditionally described in stroke patients with Cheyne-Stokes breathing, is not always necessary.^41,48,52,100 Conversely, the damage of specific brain areas (e.g., those belonging to the autonomic network) may play a so-far neglected role, particularly in the acute phase of stroke.^48,53 Patients with brainstem stroke can also develop Cheyne-Stokes breathing.^52,101–103

Hypersomnia and Excessive Daytime Sleepiness

Hypersomnia is defined clinically as a reduced latency to sleep, increased sleep, or excessive daytime sleepiness. In patients with strokes that affect activating arousal pathways or the paramedian thalamus, hypersomnia can alternate with insomnia (see earlier). Façon and coworkers described, for example, a 78-year-old patient with a tegmental mesencephalic infarct in whom severe, persistent hypersomnia was accompanied by an inversion of the sleep–wake cycle with nocturnal agitation.¹²⁴

Poststroke hypersomnia, with or without excessive daytime sleepiness, is usually observed in association with thalamic (Fig. 88-4), mesencephalic, or upper pontine strokes. Less commonly, strokes in the caudate, striatum, lower pons (Fig. 88-5), medial medulla, and cerebral hemispheres (with or without mass effect) (Fig. 88-6) may be complicated by hypersomnia.¹²⁵

Figure 88-4 Hypersomnia after bilateral paramedian thalamic stroke. This 65-year-old male patient (KM) had initial coma, followed by severe hypersomnia (A), vertical gaze palsy (B), amnesia, and disturbed time perception (Zeitgefühl). Brain MRI shows a bilateral paramedian thalamic stroke (C, D). Polysomnography performed 12 days after stroke onset demonstrates (E) a drastic reduction of sleep spindles and (F) loss of spindle peak (12-14 Hz activity) on spectral analysis (as compared to normal control [G]). A severe central apnea (apnea–hypopnea index, 54/hr) was observed in the acute phase (in the absence of any signs of cardiac dysfunction) but not on follow-up few months later. Actigraphy performed within the first month after stroke shows time “asleep” (rest or sleep) during 61% of the recording time (2 weeks) (H). One year after stroke, the patient still reports increased sleep needs (15 hours per day), apathy (athymormia) and attentional and memory deficits. Modafinil at a dose of 200 mg per day improved his hypersomnia.

(MRI pictures courtesy of Prof. A. Valavanis, Institute of Neuroradiology, University Hospital, Zürich, Switzerland.) (From Bassetti CL, Hermann DM. Sleep and stroke. In: Vinken PJ, Bruyn GW. Handbook of clinical neurology. Sleep disorders. New York: Elsevier; 2010. In press.)

Figure 88-5 Hypersomnia or excessive daytime sleepiness after pontomedullary stroke. This 39-year old male patient (UM) had pontomedullary ischemia following subarachnoid hemorrhage and embolization of a giant aneurysm of the basilar artery. This resulted in clinical brainstem syndrome with singultus, left IX, X, XII palsy, dysarthria, gait ataxia, and mild left hemiparesis. Sleep symptoms were postintervention severe excessive daytime sleepiness (EDS; Epworth Sleepiness Score: 23/24) and increased sleep needs (12-14 hours/day). Polysomnography showed sleep efficiency 97%, slow-wave sleep 8% of total sleep time, no sleep apnea, and no periodic limb movements in sleep. The multiple sleep latency test showed mean sleep latency 1 minute and no sleep-onset REM periods. Actigraphy showed time “asleep” (rest or sleep) during 43% of the recording time (2 weeks). Cerebrospinal fluid levels of hypocretin-1 were normal. The patient does not wish any treatment for his EDS and hypersomnia.

(MRI pictures courtesy of Prof. A. Valavanis, Institute of Neuroradiology, University Hospital, Zürich, Switzerland.)

Figure 88-6 Hypersomnia and sleep architecture changes after middle cerebral artery stroke. This 39-year old female patient (FF) had aphasia, right hemiparesis, depressed mood, and crying spells. National Institutes of Health (NIH) Stroke Score was 16. The patient’s sleep needs increased in the first 1 to 2 weeks after the stroke (12 hours/day as compared to 7 hours/day before stroke), followed by mild excessive daytime sleepiness (Epworth Sleepiness Score, 12). At 12 months, the patient reported a decrease in sleep needs to 10 hours per 24 hours. Repeated polysomnographic recordings (day 2, day 8, and 70 days after stroke) demonstrate a progressive recovery of spindling activity (coherent activity around 12 Hz) over both the affected (left) and the nonaffected (right) hemisphere.

(MRI pictures courtesy of Prof. G. Schroth, Institute of Neuroradiology, University Hospital, Bern, Switzerland.)

In deep (subcortical) hemispheric and thalamic stroke, hypersomnia may correspond to presleep behavior, during which patients yawn, stretch, close their eyes, curl up, and assume a normal sleeping posture, while complaining of a constant sleep urge.¹²⁶ Some of these patients are able to control this behavior when stimulated or given explicit, active tasks to perform. This presleep behavior may be compulsive in that removing the patient from bed can result in repeated attempts to lie down and adopt a sleeping posture. However, during what appear to be daytime sleep periods, relatively quick responses to questions or requests suggest wakefulness. For this peculiar dissociation between lack of autoactivation in the presence of preserved heteroactivation, Laplane suggested the term athymormia, or “pure psychic akinesia.”¹²⁷

In some patients, hypersomnia evolves to extreme apathy, with lack of spontaneity and initiative, slowness, poverty of movement, and catalepsy, a condition for which the term akinetic mutism was coined.¹²⁸ Akinetic mutism, and a less severe form, usually referred to as abulia, can persist despite normalization of vigilance or even after appearance of insomnia. In some of these patients, the eventual diagnosis is poststroke fatigue (see later) or post-stroke depression.

Narcolepsy-like phenotypes have rarely been seen following stroke, even in the absence of human leukocyte antigen (HLA) positivity and cerebrospinal fluid (CSF) hypocretin-1 deficiency.^129–131 Hypersomnia with hyperphagia (Kleine-Levin–like syndrome) was reported after multiple cerebral strokes.¹³²