Obstructive Sleep Apnea and Hypertension

Abstract

There has long been a recognized link between obstructive sleep apnea (OSA) and the cardiovascular system, no aspect of which has been more studied than blood pressure. Research in OSA has not only demonstrated dysregulation of homeostatic cardiovascular mechanisms but also has furthered our understanding of blood pressure regulatory control. Acute nocturnal blood pressure elevations associated with disordered breathing events have been reproduced from a number of observational studies, the accrual of which has also made an increasing argument for the importance of OSA in the pathogenesis of diurnal hypertension, as suggested by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7), which implicated OSA as a secondary cause of hypertension. Accumulating data from randomized controlled treatment trials in OSA, particularly with continuous positive airway pressure, though sometimes inconsistent, suggest a potential role in blood pressure reduction. Further research is needed to better clarify indications for OSA treatment as well as its role as an adjunct to other antihypertensive treatments.

In normal individuals, sleep is associated with reduced blood pressure (BP) when compared with wakefulness. Referred to as the “dipping” phenomenon, systolic and diastolic BP may decline as much as 10% to 15% during sleep. ¹ , ² Autonomic neural influences play a key role in this circadian variation in BP. In humans, wakefulness normally transitions to non‐rapid eye movement (NREM) sleep, which composes the majority of sleep time and compared with wakefulness is characterized by cardiovascular stability. With parasympathetic predominance via sympathetic neural withdrawal, NREM sleep is marked by a reduction in heart rate and BP. High baroreceptor gain maintains BP under tight control. Progression of NREM sleep from stages 1 through 4 is accompanied by further reduction in sympathetic neural traffic such that by stages 3 and 4 (slow wave), the output may be half that encountered in wakefulness. ³ , ⁴ Collectively, NREM sleep is associated with lower BP, bradycardia, and reductions in cardiac output and systemic vascular resistance. Accordingly, serum catecholamines are at their lowest levels during sleep.

The stability of NREM sleep is cyclically interrupted by rapid eye movement (REM) sleep, which occurs more frequently during the second half of the night. REM sleep is associated with transient surges in sympathetic neural output, heart rate, and BP, ⁴ the effects of which may be clinically important in those with preexisting cardiovascular disease. ⁵ REM sleep is pathophysiologically important in obstructive sleep apnea (OSA) for a number of reasons. Due to widespread skeletal muscle atonia associated with REM sleep, the upper airway is most vulnerable to collapse during this period. Second, apneas occurring during REM sleep, because of disadvantageous lung mechanics, are associated with more profound oxyhemoglobin desaturations than those that occur during NREM sleep.

INTERACTIONS BETWEEN SLEEP DURATION AND HYPERTENSION

Spurred in part by striking evidence of metabolic dysregulation resulting from short‐term acute sleep deprivation experiments in healthy individuals, ⁶ there has been increasing interest in the downstream effects of short sleep duration on systemic health and disease. Central to this theme is the recognition that members of western society lead progressively busier lives, obtain less sleep, and have increasing rates of obesity, hypertension, and cardiovascular disease. ⁷ , ⁸ It is estimated that more than two‐thirds of the US population obtains <8 hours of sleep per week‐night ⁹ and that the nightly sleep duration has dwindled by nearly 2 hours over the course of the past century.

While it is difficult to extrapolate data from controlled laboratory‐based sleep deprivation experiments in healthy volunteers to a larger free‐living population, recent longitudinal data from population‐based cohorts may help bridge this gap. The National Health and Nutrition Examination Survey I (NHANES I) showed that over 10 years of follow‐up in a middle‐aged population, self‐reported sleep duration of <7 hours per night was associated with higher body mass index and increased likelihood of obesity, as compared with those reporting ≥7 hours of sleep. ¹⁰ While it is conceivable that the development of hypertension may be mediated in large part through comorbid obesity, there are experimental and observational data implicating sleep debt as an independent risk factor for hypertension, a hypothesis that has solid biologic plausibility. The cumulative BP “load” and exposure to an activated sympathetic nervous system is increased as time awake is prolonged, while the protective effects of reduced BP and sympathetic withdrawal during sleep are truncated. ¹¹

Initial reports from longitudinal population‐based studies have suggested important effects of sleep duration on the prevalence of hypertension. Gangwisch and colleagues ¹² utilized NHANES I data to show significant increases in incident hypertension over a 10‐year period in those with self‐reported sleep durations ≤5 hours per night. This relationship was minimally attenuated after controlling for obesity and diabetes. In the cross‐sectional Sleep Heart Health Study (SHHS), ¹³ the highest odds of hypertension were seen in those reporting <6 hours' sleep per night. It should be noted that reporting >8 to 9 hours per night also conferred a higher risk of hypertension, although risk was lower than in those with the shortest sleep duration. However, more recent data from the Rotterdam Study of elderly individuals (older than 58 years), which included further validation of the sleep self‐report with actigraphy, showed no effect of sleep duration on hypertension. ¹⁴ All of these studies are limited by their methodology and reliance on self‐report; as such, larger and more rigorous interventional controlled trials may be needed to further clarify the role of sleep duration on the development of hypertension.

DIPPERS VS NONDIPPERS

Studies comparing dippers with nondippers have been limited to some degree by differing methods of measuring ambulatory BP as well as by varied applied definitions. However, population‐based studies suggest heightened cardiovascular risk in nondippers, who have been shown to have a higher risk of stroke ¹⁵ and incident heart failure, ¹⁶ and that nondipping is a risk factor for progression of renal disease. ¹⁷ Of importance, these conditions are commonly comorbid with OSA. There is a high prevalence of OSA in hypertensive nondippers, ¹⁸ while studies of nocturnal BP in those with sleep apnea demonstrate less BP dipping at night, factors which may confer added cardiovascular risk. ¹⁹ While nondippers may be more resistant to antihypertensive drug therapy, a small study has suggested that nondippers may be more sensitive to the BP‐reducing effects of nocturnal continuous positive airway pressure (CPAP). ²⁰

OVERVIEW OF OSA

OSA is characterized by repetitive episodes of upper airway narrowing or occlusion, occurring up to several hundred times per night in severe cases. Upper airway events result in oxyhemoglobin desaturation and terminate in CNS arousal, resulting in sleep fragmentation—a key mechanism in the most important symptom of OSA, excessive daytime sleepiness—which emerging data suggest may be an important mediator in the link between hypertension, OSA, and the antihypertensive effects of OSA treatment. OSA is an exceedingly common disorder, affecting as much as a quarter of the middle‐ to older‐aged population, the most important risk factors for which are obesity, male sex, and advancing age. ²¹ Since these conditions also dominate the risk profile of hypertension, ²² , ²³ it can be difficult to determine the independent effects of OSA on the development of hypertension. The most immediately effective and well‐studied treatment for OSA is CPAP, which maintains upper airway patency during sleep, promotes sleep continuity, and significantly improves subjective and objective measures of daytime sleepiness. ²⁴ , ²⁵ Weight loss is also a very effective treatment option for OSA, although its impact on reducing BP purely through attenuation of sleep‐disordered breathing is not well understood.

OSA AND HYPERTENSION: ACUTE PATHOPHYSIOLOGIC MECHANISMS

The repetitive episodes of upper airway narrowing or occlusion described above give rise to acute stressors such as hypoxemia and reoxygenation, swings in intrathoracic pressure, and CNS arousals. These mechanisms, among others, are associated with well‐recognized acute increases in peripheral vasoconstriction and attendant rises in BP during sleep.

Episodic deoxygenation drives some important aspects of the acute (and probably chronic) pathophysiology of OSA. Hypoxemia stimulates the peripheral arterial chemoreceptors, namely the carotid bodies, which are important in mediating the response to OSA‐associated hypoxemia. Carotid body afferents that relay in the brain stem elicit reflex increases in sympathetic efferent traffic during hypoxemic stimulation, as demonstrated by direct peripheral intraneural electrode recordings. ²⁶ , ²⁷ Stimulation of respiratory centers within the brain stem, in a normal response, increases respiratory muscle output and minute ventilation. Lung inflation, by way of stimulation of parenchymal vagal mechanoreceptors, serves to temper sympathetic outflow. The lack of lung inflation during apnea results in disinhibition of sympathetic neural activity and, therefore, a potentiated sympathetic response to hypoxemia. ⁴ , ²⁸ , ²⁹

Those with OSA appear to have an exaggerated peripheral chemoreflex response to hypoxemia, as demonstrated by augmented ventilatory and autonomic drive when compared with nonapneic controls. ³⁰ Chemoreflex activation results in increased sympathetic traffic to the peripheral vasculature, with a consequent acute rise in arterial BP. ³⁰ , ³¹ This heightened chemoreflex sensitivity may contribute to enhanced sympathetic tone in OSA, even during wakefulness under conditions of normoxia.

Homeostatic mechanisms, which under normal conditions temper increased sympathetic drive, are disrupted in OSA. In addition to the loss of vagal mechanisms related to reduced lung inflation during apnea, baroreflex dysfunction has also been identified in OSA. Originating in major blood vessels, such as the carotid sinus and aorta, and mediated through the CNS, the baroreflexes also serve to buffer ventilatory, pressor, and sympathetic responses to peripheral chemoreflex excitation. ²⁷ , ³² Preexisting hypertension, often seen in association with OSA, may result in impaired baroreflex function and may thus contribute indirectly to augmentation of the chemoreflex‐mediated sympathetic response. ³³

Repetitive breathing events in OSA commonly terminate in CNS arousal, with restoration of upper airway patency and resumption of ventilation. Each arousal from sleep is accompanied by acute increases in sympathetic neural output, ³⁴ which may contribute to the autonomic dysregulation characteristic of OSA.

CARDIOVASCULAR DYSREGULATION IN OSA

It is likely that there are cumulative effects related to the repetitive acute perturbations of OSA that over time may be important in the pathogenesis of chronic conditions such as hypertension. Disturbed neural circulatory control during daytime wakefulness in patients with OSA, even in the absence of overt vascular disease, is evident by the finding of heightened sympathetic drive by measurement of muscle sympathetic nerve activity, even in the absence of hypoxemia. ³¹ This finding may reasonably be attributable to increased tonic chemoreflex drive. ³⁵

Abnormalities in heart rate variability, a noninvasive method for assessing cardiac autonomic function, are present in OSA. ³⁶ Reduced heart rate variability seen in some studies of OSA may act as a marker for future cardiovascular disease. ³⁷ The vascular endothelium, a biologically active system, may also be dysfunctional in OSA. Whether this results from OSA per se is not entirely clear, but population‐based studies suggest that endothelial dysfunction may be an important marker of cardiovascular risk. The small vessel dilatory response to vasoactive substances such as acetylcholine, which represents resistance vessel endothelial function, is blunted in sleep apnea, ³⁸ , ³⁹ although these findings are not evident in all studies. ⁴⁰ Whether large or conduit vessel endothelial function is also attenuated in OSA is unclear. Levels of serum endothelin, a potent vasoconstrictor, may also be elevated in patients with OSA compared with controls. ⁴¹

Other features of OSA may indirectly increase the risk of hypertension, not least of which is the striking prevalence of overweight and obesity, primary risk factors in this patient population. While excess body weight independently puts persons with OSA at risk for diabetes, other mechanisms in OSA may contribute to glucose intolerance. These include increased sympathetic tone, repetitive hypoxemia, and sleep debt. Both clinic‐ and population‐based ⁴² , ⁴³ studies, mostly in men with severe OSA, have supported a relationship independent of obesity. ⁴⁴ OSA treatment trials have shown mild and short‐lived results in the reversal of glucose intolerance. ⁴⁵ Heightened inflammation as demonstrated by elevation of biomarkers such as C‐reactive protein may also modulate vascular risk in OSA, a theory further supported by research at the cellular level. Up‐regulation of leukocyte adhesion factors in OSA ⁴⁶ , ⁴⁷ could predispose to endothelial injury and vascular events. These pathways could be mediated through neutrophil‐derived oxidative stress ⁴⁸ and abnormalities in coagulation markers in OSA patients. ⁴⁹ Finally, there is evidence to support the role of reduced levels of the potent vasodilator nitric oxide in the mediation of vascular disease and blood pressure regulation in OSA. Ip and colleagues ⁵⁰ found significant correlations between reduced nitrite/nitrate levels and severity of OSA, with significant increases in these levels following overnight application of CPAP.

SYSTEMIC HYPERTENSION, OSA, AND TREATMENT

Despite the strong association between OSA and hypertension noted above, it is important to acknowledge that because both conditions are exceedingly common and share similar risk factors (obesity, male sex, older age), establishing an independent causative role for OSA in hypertension mandates high‐level methodologies with a minimum of bias. Also, because OSA severity continues to be characterized by the apnea‐hypopnea index (AHI), the relative effects of potentially important non‐frequency‐based parameters, such as duration and degree of hypoxemia, may often go unrecognized.

Early case‐control and cross‐sectional reports linking OSA and systemic hypertension, limited by study design and potential confounding effects of comorbid variables, have been augmented with more comprehensive longitudinal population‐based studies. ⁵¹ , ⁵² The prospective Wisconsin Sleep Cohort provided some of the first persuasive evidence implicating OSA as a possible causal factor in hypertension. ⁵¹ Specifically, the presence of hypertension 4 years after initial assessment was found to be dependent on the severity of OSA at baseline. While it is notable that the study did not specifically identify patients free of hypertension at baseline and could not determine the incident risk imparted by OSA, ⁵³ collective available data currently implicate OSA in the pathophysiology of sustained daytime hypertension. As will be discussed, the role of OSA treatment in the reversal of these hypertensive effects have been less clear‐cut.

The most effective form of therapy for OSA, CPAP, has been shown to acutely attenuate sympathetic drive and nocturnal BP in patients with OSA. ³¹ , ⁵⁴ , ⁵⁵ However, the data regarding effects on daytime BP have been more difficult to interpret. A number of observational studies, often uncontrolled and in highly select populations, have suggested improvements in daytime BP control with the use of CPAP. Because of these shortcomings and an apparent true placebo effect realized in measurement of BP, a number of randomized placebo‐controlled studies have been performed, yielding variable and sometimes inconsistent results. The generalizability of the studies is somewhat limited because they comprise small sample sizes and because the majority of participants are normotensive at baseline. However, some findings are worth mentioning. With the largest study to date, Pepperell and colleagues ⁵⁶ found a small but significant reduction in BP in a largely normotensive cohort over only 4 weeks of therapy. Becker and associates, ⁵⁷ in a controlled trial comparing therapeutic with subtherapeutic CPAP, found fairly dramatic reductions in mean BP (9.9±11.4 mm Hg) in a small cohort with severe OSA (mean AHI >60/hr) treated for >60 days, the longest trial to date. Potential limitations of the study include a high dropout rate and the fact that about two‐thirds of participants were treated with various antihypertensive medications. A point that does stand out, suggesting the importance of treatment dose effect, is that subtherapeutic CPAP reduced the AHI by 50% but did not result in any reduction in BP.

While excessive daytime sleepiness is a common and potentially dangerous sequela of OSA, it is not a universal symptom. There is burgeoning evidence to suggest that sleepiness may be an important mediator of some of the systemic effects of OSA. That is, in the absence of sleepiness, even severe OSA as quantified by the AHI does not always translate to reductions in BP after CPAP treatment, regardless of whether normotension or hypertension exists at baseline. In a randomized controlled trial, Barbe and colleagues ⁵⁸ showed that in normotensive patients with severe sleep apnea by AHI criteria but no daytime sleepiness, CPAP treatment imparted no reductions in BP. Similar findings were very recently reported by Robinson and coworkers ⁵⁹ in a cohort of hypertensive patients with sleep apnea. Even mild subjective sleepiness confers some BP benefit with the use of CPAP. ⁶⁰

Finally, a randomized trial comparing ambulatory BP in patients with moderately severe sleep apnea following treatment with either therapeutic CPAP, sham CPAP, or supplemental oxygen was recently completed. ⁶¹ While use of therapeutic CPAP resulted in BP reductions, supplemental oxygen, despite normalizing oxygen saturation, did not (Figure 1). This finding suggests that hypoxia‐mediated mechanisms may not fully explain the acute and chronic effects of sleep apnea on the vasculature. It may well be that CNS arousals, which are attenuated if not abolished with CPAP therapy, are just as important, perhaps through effects on sympathetic output or hemodynamics.

Figure 1.

Daytime and nighttime changes in blood pressure (BP) associated with placebo, continuous positive airway pressure (CPAP), and supplemental oxygen. Note significant reductions in arterial pressures with CPAP but not with placebo or supplemental oxygen. SBP indicates systolic blood pressure; MAP, mean arterial pressure; DBP, diastolic blood pressure. From Norman et al. ⁶¹

To help tie some of these findings together, a meta‐analysis of 12 randomized controlled trials of BP lowering with CPAP treatment in OSA was recently published ⁶² and the results confirmed by a subsequent paper. ⁶³ While confirming heterogeneity of study design and population, the pooled effect of CPAP treatment in those with OSA (both normotensive and hypertensive participants) was a net reduction in mean BP of 1.5 to 2 mm Hg (Figure 2). The results also suggested a greater antihypertensive effect in those with hypertension and daytime sleepiness at baseline.

Figure 2.

Meta‐analysis for the net change in 24‐hour ambulatory mean blood pressure (MBP) seen across 12 placebo‐controlled continuous positive airway pressure (CPAP) trials in obstructive sleep apnea. While there was moderate heterogeneity among the studies (several had negative outcomes, as represented by the horizontal line crossing over the vertical zero value), the pooled effect was a net reduction in MBP of 1.69 mm Hg associated with CPAP therapy. CI indicates confidence interval. From Haentjens et al. ⁶²

Because chronic conditions such as OSA‐associated hypertension could reasonably lead to vascular remodeling and other structural cardiovascular changes, it is entirely feasible that short‐term controlled studies may fail to disclose the true effects of faithful CPAP therapy on hypertension and its consequences. Furthermore, given the prevalence of hypertension and its effects on the development of other cardiovascular disease, including heart failure and stroke, the results of small changes in BP and decreases in nocturnal BP may be far‐reaching.

Disclosures:

Dr Somers is supported by NIH Grants HL61560, HL65176, HL73211, and M01‐RR00585 and the Mayo Clinic College of Medicine. Dr Caples is supported by the Mayo Foundation, the Annenberg Foundation, and Restore Medical and is a coinvestigator with Dr Somers on a grant from the ResMed Foundation. Dr Somers is a consultant for Respironics and Cardiac Concepts.