Manifestations and Prevalence
Sleep-disordered breathing refers to a spectrum of disorders that feature breathing pauses during sleep. The most common disorder is characterized by obstructive apneas and hypopneas (White, 2005), where repeated episodes of collapse (apneas) or partial collapse of the pharyngeal airway occur, usually a result of obstruction by soft tissue in the rear of the throat. Snoring, which is produced by vibrations of the soft tissues, is a good marker for OSA (Netzer, et al., 2003). Apneas or hypopneas (a reduction without cessation in airflow or effort) typically result in abrupt and intermittent reduction in blood oxygen saturation, which leads to sleep arousal, often accompanied by loud snorts or gasps as breathing resumes. Episodic interruptions of breathing also frequently cause cortical and brainstem arousals, interrupting sleep continuity, reducing sleep time, and causing increased sympathetic nervous system activation. These broad systemic effects on gas exchange and nervous system activation may lead to a range of systemic effects that affect vascular tone, levels of inflammatory mediators, and hormonal changes. As discussed in the following sections, these in turn may contribute to the development of hypertension, coronary artery disease, congestive heart failure, arrhythmias, stroke, glucose intolerance, and diabetes.
The defining symptom of sleep-disordered breathing is excessive daytime sleepiness. The symptom is likely influenced by sleep fragmentation tied to recurrent arousals that occur in response to breathing pauses. Other symptoms of fragmented sleep include decreased concentration and mood changes. The diagnosis of OSA requires detection, by polysomnography, of at least five or more apneas or hypopneas per hour of sleep (Thorpy, 2005). This rate is expressed as an index, the apnea-hypopnea index (or respiratory disturbance index), which is the average hourly number of apneas plus hypopneas.
OSA is found in at least 4 percent of men and 2 percent of women in the middle-aged workforce, according to the first major United States population-based study of the condition conducted about 15 years ago (Young et al., 1993). Those prevalence figures are based on a cutoff apnea-hypopnea index of 5 or higher, plus a requirement for daytime sleepiness. The prevalence is higher, 9 percent of women and 24 percent of men, with the same apnea-hypopnea index cutoff (Box 3-1), but without the daytime sleepiness requirement. In view of the epidemic increase of obesity (an important determinant of OSA) in recent years, these numbers might underestimate the current prevalence. However, other more recent population-based studies support these prevalence figures (Bixler et al., 1998, 2001).
The metric used most commonly to define obstructive sleep apnea and to quantify its severity is the apnea-hypopnea index, derived by identifying and manually counting each respiratory disturbance (apnea (more...)OSA prevalence appears to increase with age. Adults 65 to 90 years of age had a threefold higher prevalence rate than middle-aged adults (Ancoli-Israel et al., 1991), while the prevalence in children has been reported to be around 2 percent (Ali et al., 1993; Rosen et al., 2003), with higher estimates occurring in ethnic minorities (Gislason and Benediktsdottir, 1995; Redline et al., 1999; Rosen et al., 2003). Underdiagnosis of OSA is common, with between 10 and 20 percent of OSA being diagnosed in adults (Young et al., 1997b). Less than 1 percent of older adults in primary care are referred for polysomnography (Haponik, 1992), although these numbers might have increased in recent years due to increased awareness of the disease. Similarly, children’s OSA often goes undiagnosed too, partly because the implications of snoring are not often recognized by pediatricians. Although OSA can occur in children of any age, it is most common at preschool ages, a time coincident with tonsils and adenoids being largest relative to the underlying airway (Jeans et al., 1981).
Obstructive Sleep Apnea Causes Hypertension
OSA causes chronic elevation in daytime blood pressure (Young et al., 2002a; Young and Javaheri, 2005). The strongest evidence for a rise in systemic hypertension comes from several large, well-designed epidemiological studies, both cross-sectional (Young et al., 1997a; Nieto et al., 2000; Bixler et al., 2000; Duran et al., 2001) and prospective (Peppard et al., 2000). The Wisconsin Sleep Cohort study, a prospective study, tracked adults with sleep-disordered breathing for at least 4 years to determine new onset hypertension and other outcomes. The hypertensive effect was independent of obesity, age, gender, and other confounding factors. Controlling for obesity is especially important because it is a risk factor for hypertension as well as for OSA.
A causal association between OSA and hypertension is supported by evidence of a dose-response relationship; the higher the apnea-hypopnea index, the greater the increase in blood pressure (Peppard et al., 2000; Nieto et al., 2000). Both the Wisconsin Sleep Cohort study and the Sleep Heart Health Study showed dose-response relationships. The Sleep Heart Health Study is a community-based multicenter study of more than 6,000 middle-aged and older adults whose apnea-hypopnea index was measured by polysomnography. The likelihood of hypertension was greater at higher apnea-hypopnea index levels. Case-control studies reveal that approximately 30 percent of patients diagnosed with essential hypertension (hypertension in which the underlying cause cannot be determined) turn out to have sleep apnea (Partinen and Hublin, 2005). Further, evidence from pediatric studies indicate elevations in systemic blood pressure during both wakefulness and sleep in children with sleep apnea (Amin et al., 2004), with additional evidence of left ventricular wall changes by echocardiography.
The causal nature of the relationship between OSA and hypertension is reinforced by randomized controlled clinical trials showing that the most effective treatment for OSA, continuous positive airway pressure (CPAP) therapy, can reduce blood pressure levels. Although findings have been mixed in other studies, a critical review article that evaluated each study’s methodology and results concluded that the trials show convincing decreases in blood pressure in those patients with severe OSA. The benefit is greatest in patients with severe OSA, determined by objective (polysomnography) and subjective (daytime sleepiness) criteria. The review also concluded that there was a lack of benefit in patients who had no daytime sleepiness (Robinson et al., 2004b). However, each of these studies was relatively small (less than 150 individuals), and findings can be considered only tentative.
How does OSA cause sustained hypertension? During the night, the apneas and hypopneas of OSA cause a transient rise in blood pressure (30 mm Hg or more) and increased activity of the sympathetic nervous system (Figure 3-5). Over time, the transient changes become more sustained and are detectable during the daytime, including evidence of sympathetic overactivity (Narkiewicz and Somers, 2003). Studies have found that people with OSA (versus those with similar blood pressure, but no OSA) have faster heart rates, blunted heart rate variability, and increased blood pressure variability—all of which are markers of heightened cardiovascular risk (Caples et al., 2005). The precise pathophysiological steps from transient vascular changes to systemic hypertension are far from clear but may involve oxidative stress, upregulation of vasoactive substances (Caples et al., 2005), and endothelial dysfunction (Faulx et al., 2004; Nieto et al., 2004; Young and Javaheri, 2005).
During even the lowest phase, blood pressure during REM was higher than in the awake state. Electrooculogram (EOG), electroencephalogram (EEG), electromyogram (EMG), electrocardiogram (EKG), sympathetic
Obstructive Sleep Apnea Is Associated with Cardiovascular Disease and Stroke
Epidemiological studies reveal an association between OSA and cardiovascular disease, including arrhythmias (Guilleminault et al., 1983); coronary artery disease (Andreas et al., 1996) and specifically, myocardial infarction (Hung et al., 1990; D’Alessandro et al., 1990; Mooe et al., 1996a,b; Marin et al., 2005); and congestive heart failure (Javaheri et al., 1998). Most case-control studies detecting a relationship with myocardial infarction found adjusted odds ratios of around 4 (Young et al., 2002a,b). The large, cross-sectional Sleep Heart Health Study of nearly 6,500 (Shahar et al., 2001) found that participants in the highest apnea-hypopnea index quartile (index greater than 11) were 42 percent more likely to self-report cardiovascular disease (coronary heart disease, heart failure, or stroke) than those in the lowest quartile (adjusted OR = 1.42, 95% CI, 1.13–1.78). The adjusted OR for stroke was 1.58 (95% CI, 1.02–2.46). A higher probability of stroke associated with OSA is also supported by other studies (Bassetti and Aldrich, 1999; Parra et al., 2000; Yaggi et al., 2005; Bradley et al., 2005). In the Sleep Heart Health Study, apnea-hypopnea index was deter mined by polysomnography, and adjustments were made for a variety of confounding factors, including hypertension. That the hypertension adjustment did not eliminate the effect suggests that hypertension is not the exclusive means by which OSA may lead to cardiovascular disease. A limitation of cross-sectional and case-control analyses is that cause and effect cannot be determined: heart disease may have resulted in OSA or vice versa. However, an observational cohort study of 1,022 individuals, where 68 percent of individuals had OSA (apnea-hypopnea index of 5 or higher), showed that OSA syndrome significantly increased the risk of stroke or death from any cause, and the increase is independent of other risk factors, including hypertension (Yaggi et al., 2005). Other studies have confirmed the risk of OSA syndrome with stroke or death from any cause (Ayas et al., 2003; Gami et al., 2005). Furthermore, other large prospective studies also have shown an association between snoring—a marker for OSA—and incidence of cardiovascular diseases (Jennum et al., 1995; Hu et al., 2000), providing temporal associations in support of OSA playing a causal role in the development of heart diseases. As will be discussed in the next section, OSA is associated with glucose intolerance and diabetes, both of which are independent risk factors for cardiovascular disease.
Studies of the benefits of CPAP further support an association between cardiovascular disease and OSA. Marin and colleagues (2005), in a large, observational study of 10 years’ duration, found that patients with untreated severe OSA (apnea-hypopnea index greater than 30), relative to those receiving CPAP treatment, with similar apnea-hypopnea index severity, had a higher incidence of fatal and nonfatal cardiovascular events. The events included myocardial infarction, stroke, and coronary artery bypass surgery. The untreated patients had refused CPAP but were followed regularly. A second study found an increased mortality rate from cardiovascular disease in individuals who did not maintain CPAP treatment over a 5-year follow-up period (Doherty et al., 2005). However, the number of new cases of cardiovascular disease was independent of CPAP treatment compliance. Although observational evidence of this type is not conclusive proof, because it may be subject to confounding by indication and other biases, it still lends weight to the strength of the association.
Most studies finding elevated cardiovascular disease risk have been conducted in adults. Whether or not children with sleep-disordered breathing are at risk for cardiovascular effects is not known. Children with OSA, as noted previously, do experience changes in blood pressure profiles, heart rate variability, and ventricular wall changes as measured by echocardiography (Marcus et al., 1998; Amin et al., 2005). The paucity of longitudinal data on OSA in children, in whom levels of OSA may vary during growth and development and in whom responses to therapies such as tonsillectomy may be variable (Morton et al., 2001), limits the ability to speculate on the long-term cardiovascular effects of untreated sleep-disordered breathing in children. Nonetheless, evidence that as many as 20 to 25 percent of children may have persistent OSA even after tonsillectomy underscores the potential importance of OSA as an early childhood risk factor for later cardiovascular diseases (Amin et al., 2005; Larkin et al., 2005).
Obstructive Sleep Apnea May Contribute to Obesity
Up to 40 percent of people who are morbidly obese have OSA (Vgontzas et al., 1994). This finding may reflect the role of obesity as a well-established risk factor for the development of OSA. It may also reflect obesity as a consequence of OSA, although the evidence is not yet conclusive (Grunstein, 2005b). Patients with newly diagnosed OSA, compared with controls matched for BMI and percent body fat, show recent weight gain (Phillips et al., 1999). Data from the Wisconsin Sleep Cohort also show that individuals with OSA have reduced levels of physical activity; OSA-related sleepiness may contribute to changes in activity and energy expenditure, and thus contribute to weight gain. OSA-related hormonal changes may also contribute to obesity. In general, patients with OSA have higher levels of leptin, the appetite-suppressing hormone (Phillips et al., 2000; Palmer et al., 2004; Patel et al., 2004) than controls. However, their morning levels are relatively lower than evening levels (Patel et al., 2004). Thus, either via leptin resistance (where high levels of leptin are present, but tissues are poorly responsive to leptin’s action) or because of disturbances in diurnal variability in leptin, individuals with OSA may be predisposed to lower effective levels of appetite suppressing hormones. Although CPAP reduces leptin levels, it is not known whether such effects relate to differences in the effectiveness of leptin’s actions (Chin et al., 2003). Furthermore, obesity also affects the severity of OSA. Significant weight loss in adolescents who underwent gastric bypass surgery (mean, 58 kg) was associated with a dramatic reduction of OSA severity (Kalra et al., 2005).
Etiology and Risk Factors
In simplest terms, OSA is caused by narrowing or collapse of the airway as a result of anatomical and physiological abnormalities in pharyngeal structures. Apnea episodes cause hypoxemia (insufficient oxygen in the blood) and hypercapnia (high concentration of blood carbon dioxide). The episodes also increase the output of the sympathetic nervous system (Narkiewicz and Somers, 2003), the effect of which is to restore pharyngeal muscle tone and reopen the airway. Although increased sympathetic activity is beneficial for restoring normal breathing and oxygen intake over the short term, it has long-term deleterious effects on vascular tone and blood pressure, among other effects (Caples et al., 2005). These early events—which are mediated by a variety of chemoreceptors in the carotid body and brainstem—trigger pathophysiological changes that occur not only during the obstructive apneas, but also extend into wakeful states during the day. For example, during daytime wakefulness, people with OSA have higher sympathetic activity (Somers, et al., 1995) and heightened chemoreflex sensitivity, which in turn generates an increased ventilatory response (Narkiewicz et al., 1999). The full pathophysiology of OSA remains somewhat elusive, although research is piecing together the relationships between OSA and a range of the previously described long-term health effects. The etiology of central sleep apnea, although also not well understood, is hypothesized to result from instability of respiratory control centers (White, 2005).
There are a number of risk factors for OSA, including:
Obesity, male gender, and increasing age (Table 3-1) (Young et al., 1993). It is unclear how incidence changes with older age; some data suggest that snoring and OSA may decline after age 65 years (Young et al., 1993); however, other studies show very high prevalence rates of OSA in elderly individuals (Bliwise et al., 1988; Ancoli-Israel et al., 1993; Foley et al., 2003). The pathophysiological roles of these risk factors are not well understood, although evidence suggests that fat deposition in the upper airways, which is more likely in males, contributes to the physical narrow ing that causes OSA (Robinson et al., 2004a). Menopause also increases the risk of OSA (Bixler et al., 2001; Young et al., 2003), possibly through lower levels of progestational hormones that influence the respiratory system through changes in body fat distribution (Vgontzas and Kales, 1999). However, recent studies suggest that there may be a referral bias that results in a lower apparent rate of sleep apnea in females than in males (Kapsimalis and Kryger, 2002; Shepertycky et al., 2005). Epidemiological evidence suggests that hormone replacement therapy lessens the risk of OSA (Shahar et al., 2003). In children, the main risk factor for OSA is tonsillar hypertrophy, although OSA may also occur in children with congenital and neuromuscular disorders and in children who were born prematurely (Rosen et al., 2003). Asthma, a common childhood respiratory illness, is also associated with OSA in children (Sulit et al., 2005).
In adolescents, risk factors may be more similar to those seen in adults and include obesity (Redline et al., 1999). Being a minority is a risk factor for both increased prevalence and severity of sleep-disordered breathing in both children and adults (Rosen et al., 1992; Ancoli-Israel et al., 1995; Rosen et al., 2003). The prevalence of sleep-disordered breathing in the United States is approximately three times higher in middle-aged members of minority groups compared to non-Hispanic whites (Kripke et al., 1997). African American children are at increased risk, even after adjusting for obesity or respiratory problems (Redline et al., 1999; Rosen et al., 2003). Familial and probably genetic factors strongly contribute to OSA (Buxbaum et al., 2002; Palmer LJ et al., 2003; Palmer et al., 2004).
Patients with cardiovascular disease and diabetes are also at higher risk for developing both OSA and central sleep apnea (Sin et al., 1999).
Patients with impaired baroreflexes (e.g., patients with hypertension or heart failure and premature infants) may be especially susceptible to excessive autonomic responses to chemoreflex stimulation during periods of apnea. In these patient groups, bradyarrhythmias, hypoxia, hypoperfusion, and sympathetic activation during apnea may predispose to sudden death (Somers et al., 1988; 1992).
Risk Factors for Obstructive Sleep Apnea.
Sleep-Disordered Breathing May Affect Mortality
Limited evidence suggests that sleep-disordered breathing may affect an individual’s mortality (Young et al., 2002a,b; Lavie et al., 2005). Studies of patients at sleep clinics tend to show an association between sleep apnea and mortality (He et al., 1988), but several well-designed, population-based studies failed to find an association (Ancoli-Israel et al., 1996; Lindberg et al., 1998; Kripke et al., 2002), except in one subgroup of patients below age 60 with both snoring and excessive daytime sleepiness. The subgroup experienced twice the risk of mortality (Lindberg et al., 1998). A recent observa tional study of a large cohort of sleep apnea patients (n = 403), snorers, and healthy controls who had been followed for an average of 10 years, found a threefold higher risk of fatal cardiovascular events with severe OSA (Marin et al., 2005). An observational follow-up study of the long-term effects of CPAP therapy on mortality found that compared to individuals that began receiving CPAP therapy for at least 5 years (n = 107), individuals that were untreated with CPAP (n = 61) were more likely to die from cardiovascular disease (14.8 percent versus 1.9 percent, log rank test, P = .009) (Yaggi et al., 2005; Doherty et al., 2005).
In adults, OSA is most effectively treated with CPAP and weight loss (Strollo et al., 2005; Grunstein, 2005a). Evidence of CPAP’s efficacy for alleviating daytime sleepiness comes from randomized controlled trials and meta-analysis (Patel et al., 2003). The problem is that many patients are noncompliant with CPAP (see Chapter 6). Other options, although less effective, include a variety of dental appliances (Ferguson and Lowe, 2005) or surgery (e.g., uvulopalatopharyngoplasty) (Powell et al., 2005). In children, the first-line treatment for most cases of OSA is adenotonsillectomy, according to clinical practice guidelines developed by the American Academy of Pediatrics (Marcus et al., 2002). Children who are not good candidates for this procedure can benefit from CPAP. Central apnea treatment is tailored to the cause of the ventilatory instability. Commonly used treatments include oxygen, CPAP, and acetazolamide, a drug that acts as a respiratory stimulant (White, 2005).