Apneic Sleep, Insufficient Sleep and Hypertension

INTRODUCTION

Obstructive sleep apnea (OSA) is a sleep-related breathing disorder that is highly prevalent and associated with multiple adverse cardiovascular consequences, including systemic hypertension, pulmonary hypertension, atrial fibrillation, heart failure and stroke^1–4. Of these, the strongest association is that with systemic hypertension. Recent studies have also shown an association between short sleep duration and hypertension^{5, 6}. Autonomic mechanisms, specifically sympathetic activation, are thought to play a major role in the pathophysiology underlying the association between apneic and inadequate sleep and hypertension^7–9.

This review describes the normal physiology of blood pressure (BP) in sleep and the literature examining elevated BP in subjects with sleep-related breathing disorders and insufficient sleep. We will examine potential mechanisms underlying the relationship between disrupted sleep and hypertension, and the impact of their treatment on BP. The review is limited to studies of adult subjects.

NORMAL SLEEP AND BLOOD PRESSURE

Neural circulatory control of BP during sleep is sleep stage dependent. The changes in respiration, BP and heart rate (HR) during normal sleep and wakefulness are described below.

NREM sleep

In non-rapid eye movement (NREM) sleep, a decrease in central respiratory drive occurs. This results in a fall in minute ventilation leading to a rise in blood carbon dioxide (PaCO2) levels. There is a concomitant rise in parasympathetic tone with a decrease in sympathetic activity. These changes progressively increase from lighter to deeper stages of NREM, i.e. sleep stages N1 through N3. BP, HR, cardiac output and peripheral vascular resistance all decrease in NREM stages 1 through 3^{10, 11}. Thus, BP normally shows a “dipping” pattern at night¹².

REM sleep

In rapid eye movement (REM) sleep, there is a reduction in tone of the respiratory muscles except for the diaphragm¹³. Simultaneously, a decrease in sensitivity of the central chemoreceptors occurs¹⁴. Consequently, ventilation decreases to levels even lower than that seen during NREM sleep and PaCO2 levels start to rise. Sympathetic tone increases during tonic REM sleep, with intermittent surges in both sympathetic and parasympathetic activation during phasic REM, with marked BP and HR fluctuations, such that the overall BP and HR in REM sleep are similar to the levels seen during wakefulness¹⁰.

Wakefulness

An “arousal” from sleep is represented by a change in the electroencephalogram for <10 seconds, and if this change is prolonged for 10 or more seconds, it is termed as an “awakening”¹⁵. During arousals and awakenings from sleep, sudden increases in minute ventilation, HR and BP are noted^{16, 17}. Some degree of ventilatory “overshoot” occurs in that the increase in respiratory drive during arousals and awakenings is higher than that expected for the PaCO2 level at the time¹⁶. A rise in sympathetic activity and decline in parasympathetic tone occurs; the consequent rise in BP and HR during arousals and awakenings significantly exceeds baseline wakefulness levels¹⁷.

SLEEP-RELATED BREATHING DISORDERS AND HYPERTENSION

OSA is the most common sleep-related breathing disorder, characterized by repeated complete (apnea) or partial (hypopnea) upper airway inspiratory closure in sleep¹⁸. The severity of OSA is gauged by the number of disordered breathing events per hour of sleep or recording time on a sleep study (in-laboratory polysomnography or a home sleep apnea test). The events are accompanied by oxyhemoglobin desaturation and sometimes with arousals from sleep. Up to 24% of men and 9% of women between 30–60 years of age are estimated to have OSA in the general U.S. adult population, with some other cohorts showing higher prevalence rates of sleep disordered breathing (SDB) (e.g. 50% of men and 23% of women with moderate to severe SDB in a population-based study in Switzerland)^{1, 2}.

Mechanisms linking sleep apnea and hypertension

The chemoreflexes

Whereas in normal sleep, changes in BP are mainly sleep stage dependent, in subjects with SDB, these changes depend on the severity of disordered breathing events^{19, 20}. Recurrent episodes of cessation of breathing due to upper airway closure or narrowing with accompanying hypoxemia lead to chemoreflex-mediated sympathetic activation²¹. At the termination of the disordered breathing event, an increase in cardiac output occurs, but this is delivered into a constricted vasculature. As a result, intermittent rises in BP as high as 250/110 mm Hg may be noted in individuals with OSA, that are attenuated by continuous positive airway pressure (CPAP) treatment (Figure 1)^{21, 22}. These sudden increases in BP with disordered breathing events may be responsible for the non-dipping pattern of BP seen in subjects with untreated OSA^{23, 24}.

Figure 1.

Recording of sympathetic nerve activity, respiration and intra-arterial BP in the same subject with OSA- while awake, in REM sleep and with CPAP treatment, showing elimination of apnea, decreased SNA and prevention of BP surges with CPAP (reproduced with permission from²¹)

Animal studies have demonstrated an increase in BP with repeated upper airway closure and intermittent hypoxia^25–28. Studies of human subjects have demonstrated increased sympathetic activity and mean arterial BP in acute as well as chronic intermittent hypoxic conditions^29–34. It is thought that hypoxia is the main stimulus for the increased sympathetic response seen in sleep apneics, but hypercapnia and the apnea itself are also implicated^{19, 35–37}. Additionally, in subjects with OSA, repeated arousals from sleep at the end of disordered breathing events may contribute to persistent sympathetic overactivity and to the development of systemic hypertension over time^38,39.

Higher sympathetic activity that is seen during sleep in subjects with OSA may eventually be noted during wakefulness as well. Cortical autonomic network grey matter changes relating directly to the intensity of sympathetic discharges during wakefulness have been described recently⁴⁰. Chronic intermittent hypoxia-induced changes in sympathetic regulation appear to precede the increase in BP in animal models of sleep apnea⁴¹. In human subjects, the sympathetic response to hypoxia has been shown to be independently linked with higher daytime and nighttime BP in subjects with OSA⁴², but it is not completely clear that the sympathetic dysregulation precedes the increase in BP¹⁹.

The chemoreflexes, including the central and peripheral chemoreceptors, play a key role in the sympathetic activation seen in OSA. The peripheral chemoreceptors located in the carotid body, respond primarily to hypoxemia^{19, 25, 43–46}. The central chemoreceptors, on the other hand, are located in the brainstem, and respond predominantly to hypercapnia^{19, 46}. Both peripheral and central chemoreceptors elicit hyperventilation and increased sympathetic outflow to peripheral blood vessels. Bradycardia, mediated by increased cardiac vagal drive, is triggered by the peripheral chemoreflex response to hypoxemia. Increased sensitivity of the peripheral chemoreceptors^{26, 47–49} and their tonic activation^{50, 51,42, 52} even during periods of normoxia in subjects with OSA, may account for the heightened sympathetic drive observed during wakefulness. Importantly, parasympathetic (vagal) activation to the heart, induced by the peripheral chemoreflex, may elicit marked bradyarrhythmias, including heart block and sinus arrest⁵³.

It appears that the responses of the peripheral chemoreceptors to hypoxemia in OSA are independent of the effects of comorbid hypertension as they are seen in both hypertensive and normotensive subjects^{52, 54}. Such responses appear to be independent of the effects of obesity as well¹⁹. Central chemoreceptor activation elicited by hypercapnia is accentuated in those with obesity (although not all reports are consistent), whereas it is mainly the peripheral chemoreceptor response to hypoxemia that is enhanced in subjects with OSA^{47, 54, 55}. Resting muscle sympathetic nerve activity, measured during normal awake normoxic breathing, is higher in subjects with OSA compared to obese controls^{56, 57}. Sympathetic activation in response to apnea and hypoxemia is also increased in subjects with OSA versus weight-matched controls, to a degree similar to that observed in subjects with early hypertension^{31, 36, 50, 58, 59}.

There appears to be a dose-response relationship between severity of OSA and hypertension^60–62. Furthermore, the effects on BP may be more prominent in OSA subjects with excessive daytime sleepiness, presumably due to frequent arousals leading to sympathetic activation independent of the effects of hypoxemia⁶³. In addition to the severity of OSA and presence of daytime sleepiness, the temporal distribution of apneas during sleep, variability in severity of apnea on different nights, as well as the degree of desaturation and/or autonomic activation may influence an individual’s risk of developing hypertension¹⁴.

The baroreflex

Subjects with OSA have been noted to have blunted baroreceptor sensitivity and impaired baroreflex responses^{10, 64}. Although the baroreflexes are thought to be only marginally involved in the acute autonomic response to apnea and hypoxemia, they likely play a significant role in the suppression of sympathetic activity following termination of the disordered breathing event^{10, 64, 21}. In the long term, repeated increases in BP in subjects with OSA may result in resetting of the baroreceptors to a higher baseline BP level. This may contribute further to a “permissive” sympathetic activation and BP elevation^{65, 66, 43, 67, 68}.

Central Control

An increase in the central neural response to chronic intermittent hypoxia may contribute to the exaggerated acute sympathetic activity in OSA^{69, 70}. There may be a feedback loop wherein increased levels of serum catecholamines in subjects with OSA stimulate central sympathetic nuclei, leading to further increases in sympathetic tone^{71, 72}.

Cardiovascular Variability

Changes in cardiovascular variability, including significant increases in variability in BP that occurs in a dose-response fashion in subjects with OSA, reflects the autonomic dysfunction seen in these subjects^73–75. Spectral analyses show increased low to high frequency band power suggesting decreased baroreflex efficiency and sympathetic overactivity in subjects with OSA^{73, 76}, in contrast to that seen in normal sleep⁷⁷, and may result in impaired HR regulation and heightened vasomotor tone leading to hypertension. CPAP treatment has been shown to affect some measures of cardiovascular variability, both in the short- and long-term^{78, 79}.

Vascular mechanisms

Increases in endothelin release in the nighttime, endothelial cell dysfunction and apoptosis are seen in the setting of OSA, with a reversal of these effects noted with CPAP^{71, 80–84}. Statin medication has been shown to prevent endothelial dysfunction and hypertension resulting from intermittent hypoxia in rats; further studies are required to explore whether these effects occur in humans⁸⁵.

Renin-Angiotensin-Aldosterone System

Sympathetic activation results in the release of renin in the kidney which leads to increased production of angiotensin II. This in turn increases sympathetic tone by acting through central and peripheral mechanisms and can lead to the development of hypertension^86–88. Plasma aldosterone concentration and 24-hour urine aldosterone levels were found to positively correlate with the apnea-hypopnea index (AHI) in one study⁸⁹. Angiotensin production in response to hypoxemia may be increased at the tissue level as well in subjects with OSA¹⁹. Activation of the renin-angiotensin-aldosterone system can conceivably potentiate parapharyngeal edema and rostral fluid shifts in sleep, increasing upper airway obstruction and worsening OSA severity⁹⁰.

The renin-angiotensin-aldosterone system may be an important mechanism underlying resistant hypertension in subjects with OSA. A recent randomized proof-of-concept study of renal denervation in subjects with OSA and resistant hypertension demonstrated a reduction in AHI by 8 events per hour in those who underwent the procedure versus an AHI reduction of only 1 per hour in controls⁹¹, suggesting a potential interaction between renal sympathetics and OSA severity.

Aortic stiffness

Aortic stiffness, measured most commonly by the carotid-femoral pulse wave velocity, is a marker of adverse cardiovascular outcomes including hypertension⁹². Aortic stiffness has been associated with OSA as well as sleep deprivation and it is possible that this could play a role in the development of hypertension^93–95.

HPA axis

OSA has been associated with hypothalamus-pituitary-adrenal (HPA) axis activation, although not all studies are consistent⁹⁶. HPA axis activation ostensibly leads to an increase in vasopressin-induced increased BP. Acute and subacute sleep deprivation are associated with higher serum cortisol levels that can also increase BP⁹⁷.

Other factors

Oxidative stress and inflammation are other factors that may play a role in the relationship between OSA and hypertension^{21, 98, 99}. Increased inflammation in addition to cerebral white matter changes are thought to play a role in the adverse cardiovascular events seen in subjects with OSA and non-dipping of BP at night¹⁰⁰.

Moreover, disordered breathing events themselves, generating negative intrathoracic pressures of up to 40 cm H2O, can lead to increased left ventricular transmural pressures²¹. Finally, impairment of exercise tolerance may be a contributory factor to hypertension or other adverse cardiovascular outcomes in subjects with OSA; this area needs further study^{101, 102}.

In summary, chemoreflex activity is abnormal in OSA and likely contributes to the development of hypertension over time. Other mechanisms such as impaired baroreflex, vascular dysfunction and activation of the renin-angiotensin-aldosterone system may also be implicated in OSA-induced BP elevation. Recognition of the role of the chemoreflexes and other mechanisms by which BP is increased in OSA may allow us to alter the development and progression of hypertension in this patient population. Investigation into novel methods of quantification and manipulation of the chemoreflex response may help individualize prevention and treatment strategies in patients with OSA.

Obstructive sleep apnea and systemic hypertension

OSA is seen in approximately 50% of those with hypertension^{59, 64}. Conversely, 50% with hypertension have comorbid OSA¹⁰³. The high prevalence of resistant hypertension in OSA may be partially explained by shared risk factors such as obesity¹⁰⁴. Large epidemiological studies have shown OSA to be independently associated with prevalent as well as incident hypertension. OSA increases the risk of baseline and future hypertension up to 12 years later in a dose-response manner^{60–62, 105}. REM-related OSA, a subtype of OSA where obstructive sleep disordered breathing events are clustered in REM sleep, has also been shown to be significantly associated with hypertension after accounting for confounders^{106, 107}. As REM sleep mainly occurs in the second half of the night, these findings reflect the importance of attempting to treat OSA during the entire sleep period. There appears be a stronger association between hypertension and OSA severity in those with subjective and objective excessive daytime sleepiness^108–110. Subjects with excessive daytime sleepiness appear to have more severe nocturnal desaturation and higher diastolic BP than those without sleepiness¹¹¹.

Obstructive sleep apnea and non-dipping pattern of blood pressure

The prevalence of non-dipping of BP in patients with OSA is high, between 50–80%^{112, 113}. Non-dipping of BP is a negative prognostic indicator, the presence of which is associated with increased cardiovascular morbidity, left ventricular hypertrophy and microalbuminuria, and may contribute to the adverse cardiovascular outcomes seen in subjects with OSA^114–116. Similarly, nocturnal BP may be a stronger predictor of adverse cardiovascular consequences compared to daytime BP. During sleep, the renal afferent arterial tone is lower than that during wakefulness, allowing for more direct transmission of elevated BP to the glomerulus. Age appears to exert a differential effect, such that the severity of respiratory abnormalities in subjects with OSA may be related to non-dipping status in younger individuals while the severity of sleep disturbance may be related to non-dipping status in older individuals¹¹⁷.

Obstructive sleep apnea and resistant hypertension

OSA is a common cause of resistant hypertension as well as refractory hypertension^118–122. The 7th Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-7) listed OSA as one of the secondary causes of hypertension¹²³.

OSA is seen in a higher proportion (approximately 70%) of those with resistant hypertension^{19, 64, 86, 122, 124, 125}. In one study, more than half of subjects with resistant hypertension were noted to be at high risk for OSA based on the Berlin questionnaire¹²⁵.

Conversely, those with severe OSA are at increased risk of having resistant hypertension. In a study of subjects with cardiovascular disease, those with untreated severe OSA had four-fold higher odds of having resistant hypertension compared to those with moderate OSA¹²⁶.

Of particular concern is the role of OSA in the development of resistant hypertension in minority populations, especially in African-Americans, in whom there is a high prevalence of undiagnosed OSA, and in whom severe OSA is accompanied by a greater than 3-fold increase in risk of having resistant hypertension¹²⁷.

Effects of treatment of sleep apnea on hypertension

CPAP has been shown to decrease sympathetic activity as well as BP in sleep^{21, 128, 129}. The decrease in sympathetic activity seems to be greater in those with severe OSA and coexisting diabetes mellitus¹³⁰. Reduction in BP is maximal in those with those with the longest hours of CPAP usage and appears to be sustained over time^{47, 131–133}. Fixed CPAP may be better at controlling diastolic BP than auto-titrating devices according to a recent study¹³⁴.

CPAP has been shown to decrease 24-hour systolic and diastolic BP in the range of 2–4 mm Hg in a number of studies, in individuals with OSA and prevalent pre-hypertension as well as hypertension^{128, 130, 135–140,109}. Oral appliance treatment has also been shown to result in a similar decrease in 24-hour BP in subjects with OSA^{141, 142}. Studies on the effects of CPAP on incident hypertension, however, are conflicting^{105, 143}. The recent SAVE trial did not show a beneficial effect of CPAP on the composite end-point of major adverse cardiovascular events in subjects with existing cardiovascular disease but it should be noted that mean nightly usage of CPAP was 3.3 hours in this study which would be considered clinically inadequate^{140, 144}. In a meta-analysis of trials evaluating the effects of CPAP on cardiovascular outcomes, the results of which were largely driven by the SAVE trial, a borderline statistically significant response (p=0.05) was noted with regard to reduction in BP^{145, 146}.

It is possible that CPAP treatment and antihypertensive medications exert a synergistic effect on decreasing BP in subjects with OSA. The use of angiotensin II receptor antagonist medications have been shown to result in a more significant fall in BP, particularly nocturnal BP, in hypertensive subjects without OSA versus those with coexisting OSA¹⁴⁷. In this study, the use of adjunct CPAP therapy for four or more hours per night for 6 weeks led to a significant reduction in systolic BP by a mean of 4.7 mm Hg in those who were adherent to the device. Similarly, CPAP treatment along with weight loss has been shown to result in a greater reduction in BP compared to either intervention alone¹⁴⁸.

Recent meta-analyses have shown a reduction of 5–7 mm Hg in systolic BP and 3–6 mm Hg in diastolic BP with the use of CPAP in subjects with OSA and resistant hypertension^149–151. One of the randomized controlled trials included in these meta-analyses was a multi-center study of subjects with resistant hypertension and OSA that showed that those who were treated with CPAP had a significant decrease in 24-hour mean and diastolic BP in the long-term, i.e. at 12 weeks follow-up; 72% of subjects in the CPAP group used their device for 4 or more hours per night in this study (Figure 2)¹⁵². An improvement in patient-reported outcomes such as Epworth Sleepiness Scale, Fatigue Severity Scale and Patient Health Questionnaire-9 scores has also been reported in this population¹⁵³. Furthermore, a higher proportion of subjects with resistant hypertension demonstrate a dipping pattern after treatment with CPAP^{152, 154, 155}. There are conflicting data regarding the impact of timing of antihypertensive medication on reduction of nocturnal BP and non-dipping pattern in BP in OSA subjects treated with CPAP^{156, 157}.

Figure 2.

Correlation between CPAP use and BP at the end of the follow-up period (reproduced with permission from¹⁵²)

Abbreviations: BP=blood pressure, CPAP=Continuous positive airway pressure, OSA=obstructive sleep apnea, REM=rapid eye movement sleep, RESP=respiration, SNA=sympathetic nerve activity

Sleepiness may be an important marker of OSA pathophysiology and response to treatment. It is possible that hypertensive subjects with OSA who are sleepy experience the greatest reductions in BP post-CPAP treatment¹³⁸ compared to those who are not sleepy^{143, 158, 159}. One recent study noted that a cluster of plasma micro-ribonucleic acid predicted a decrease in BP post-CPAP treatment¹⁶⁰. Such studies, if replicated, may have implications with regard to individualization of treatment in subjects with OSA to help reduce the risk of hypertension and downstream negative consequences.

Thus, there appears to be a dose-response relationship between OSA and both prevalent and incident hypertension in large epidemiologic studies, particularly in those with REM-related OSA or excessive sleepiness. A link between untreated OSA and nondipping BP at night as well as resistant hypertension is also evident. CPAP may have a beneficial, albeit modest, effect in lowering BP in those with hypertension. Further large-scale prospective and randomized controlled trials are required to conclusively demonstrate the relationship between OSA and hypertension and clarify the effects of adequate CPAP therapy, particularly in preventing the development of hypertension. Studies are also needed to help determine which subsets of patients with OSA and hypertension experience the greatest reduction in BP with appropriate OSA treatment and the comparative effects of different treatment modalities on BP. Renal denervation has recently emerged as an intriguing potential therapeutic option for resistant hypertension in patients with comorbid OSA^{91, 161}. Last, defining the role of OSA (often undiagnosed) in mediating uncontrolled and resistant hypertension in African-Americans is a high investigative priority, as is the potential contribution of OSA to racial disparities in hypertension prevalence and in risk of cardiovascular events¹²⁷.

INSUFFICIENT SLEEP, INSOMNIA AND HYPERTENSION

Inadequate Sleep Duration

According to current statistics less than two thirds of US adults sleeps habitually >7 hours¹⁶². A persisting state of sleep debt is thus afflicting roughly one-third of Americans, with comparable estimates worldwide, and the detrimental public health consequences are increasingly recognized. Chronic sleep deficiency is an independent contributor to morbidity and mortality, with especially abundant evidence substantiating a linkage between short sleep and risk of hypertension^163–166. However, excessive sleep duration may be similarly harmful, as suggested by the vast majority of epidemiological data depicting a U-shaped relation between sleep time and elevated BP. A contemporary examination of over 700,000 national survey respondents indeed found that the prevalence of hypertension progressively increased when departing from the reference category of 7 hours¹⁶⁶. Longitudinal investigations assessing incidence of hypertension reaffirm the predictive significance of short sleep duration, while the impact of long sleep for future hypertension is generally weaker^{163, 165, 167}. In a 8–10 year follow-up analysis of the National Health and Nutrition Examination Survey (NHANES) I cohort, fully-adjusted models showed that individuals reporting ≤5 hours of sleep were 32% more likely to develop hypertension than those sleeping 7–8 hours¹⁶³. Actigraphy-derived objective short sleep duration predicted BP values and incident hypertension at five years in participants from the Coronary Artery Risk Development in Young Adults (CARDIA) study¹⁶⁷. The results also showed that each hour of decrease in sleep length yielded a 37% increased risk of elevated BP.

The association between short sleep and elevated BP features a demographic patterning, with greater vulnerability seen in women than in men¹⁶⁵, and in younger and middle-age individuals compared to older subjects^{163, 168}. A moderating role of race/ethnicity is also evident, as the relationship between risk of hypertension and insufficient sleep is more robust in Blacks compared to Whites^{167, 169}. It is noteworthy that the co-occurrence of short sleep duration and overt sleep disorders confers incremental risks beyond those bestowed by the sleep disorder alone. In the OSA population, a dose-response relationship between sleep duration and comorbid hypertension has been documented, with the risk of high BP increasing as sleep time decreases¹⁷⁰.

While there is a lack of deterministic data in support of a pressor effect of long sleep duration, laboratory-based studies of total or partial sleep deprivation have revealed that BP elevation following sleep curtailment is accompanied by altered BP-regulatory mechanisms^171–173. Sympathetic activation secondary to sleep loss is indicated by increased plasma norepinephrine and heart rate variability indices reflecting enhanced cardiac sympathetic drive¹⁷⁴. Conversely, muscle sympathetic firing is suppressed in response to sustained wakefulness^{171, 172}, whereas findings on heart rate changes are conflicting^{171–173, 175}. Sleep deprivation also causes arterial baroreflex resetting to higher BP levels¹⁷¹, and functional and structural vascular deterioration, as reflected by blunted endothelium-dependent vasodilation^{175, 176} and increased arterial stiffness¹⁷⁷. Furthermore, given the multi-system adverse impact of sleep deficiency, it is plausible that other factors may be implicated in any associated BP-elevation, including impaired insulin sensitivity, inflammation, and HPA axis stimulation (as reviewed elsewhere¹⁷⁸).

Thus, there is a large body of evidence comprising cross-sectional and longitudinal studies that documents an elevated risk of hypertension in those with insufficient sleep. Short sleep duration appears to confer an increased risk of hypertension beyond that imparted by concurrent sleep disorders such as OSA. Aberrant autonomic, vascular and metabolic responses may contribute to this risk; however, this area requires further study. As a result of the accumulating literature on the clinical repercussions of inadequate sleep, the leading sleep societies have recently issued age-appropriate recommendations for healthy sleep duration^{179, 180}. Research investigating the health advantages of sleep normalization in short or long sleepers is warranted. In particular, randomized controlled trials investigating the impact of increasing sleep duration on BP are needed.

Insomnia

Insomnia is the most frequent sleep disturbance and is generally regarded as a difficulty in initiating or maintaining sleep, or early awakening, which results in non-refreshing sleep and impaired daytime functioning. Sporadic insomnia symptoms are reported by about 30% of the general population, and 6–10% of individuals meet formal diagnostic criteria for chronic insomnia disorder¹⁸¹. While the links between insomnia and mood and cognitive disturbances have long been known, growing research is unveiling the medical complications of insomnia. Prevalence of hypertension was found to be higher in those with insomnia than in those without insomnia (14.9–43.1% vs 12.8–18.7%) in a number of population-based studies^{182, 183}, but others have not confirmed these findings¹⁸⁴. Data on susceptibility to future hypertension are similarly inconsistent^{185, 186}. Such discrepancy can be partially attributed to the various definitions and severities of insomnia used, which critically influence its predictive relevance.

As mentioned above, short sleep duration in the setting of a primary sleep disorder exacerbates the unfavorable risk profile associated with the latter. Accordingly, insomnia with objective short sleep duration is the most unfavorable of all insomnia phenotypes¹⁸⁷. Using the Penn State Cohort dataset, Vgontzas and colleagues found that odds ratios for hypertension were five-fold higher in insomniacs sleeping <5 hours (as quantified from PSG) than in normal sleepers, while insomnia with normal sleep duration did not portend significantly heightened risk¹⁸⁸. Comparable positive results were obtained when the same cohort was examined for the joint effect of insomnia and short sleep on hypertension incidence¹⁸⁹. Notably, this relationship is apparent only when considering objectively quantified sleep time¹⁹⁰.

With regard to the mechanisms mediating hypertension risk in insomnia, several studies have found evidence suggestive of abnormal neural circulatory control; compared to good sleepers, insomniacs exhibit attenuated baroreflex sensitivity and increased sympathetic neural cardiovascular reactivity to stress¹⁹¹. Higher circulating norepinephrine¹⁹², increased heart rate and decreased pre-ejection period^{193, 194} in insomniacs relative to controls are also consistent with sympathetic upregulation, while increased cortisol suggests enhanced HPA axis activity¹⁹⁵. Importantly, this pattern of derangement supports the theory that a state of hyperarousal is implicated in insomnia onset and progression¹⁹⁶. The contribution of physiological hyperarousal to BP elevation in insomnia is further corroborated by a study from Lanfranchi and colleagues, showing that the abnormal 24-hour BP profile, with blunted nocturnal dipping, displayed by insomniacs is associated with EEG measures of cortical activation during sleep¹⁹⁷. Also, a case-control study found that insomnia with physiological hyperarousal (indicated by long sleep latency on the multiple sleep latency test, MSLT) was associated with greater risk of hypertension in a dose-dependent fashion¹⁹⁸. Also of great interest are emerging data implicating insomnia in progression of renal dysfunction¹⁹⁹.

Thus, cross-sectional data suggest an association between insomnia and prevalent hypertension and there is evidence of a link between insomnia with objectively measured short sleep duration and incident hypertension from longitudinal studies. Available evidence suggests that a state of physiologic hyperarousal with sympathetic activation contributes to such risk but further studies are needed to confirm this and expand on pathophysiological pathways. Initial data suggestive of ancillary antihypertensive benefits of pharmacological and behavioral treatments of insomnia have been reported^{200, 201}, although the lack of replication of findings in a targeted trial²⁰² prevents definitive conclusions and speaks to the need for more systematic, adequately designed research studies investigating the effects of sedative hypnotic medications as well as cognitive behavioral therapy for insomnia on BP. Finally, whether effective treatment of insomnia can attenuate or reverse renal impairment, thus resulting in improved BP control, awaits further investigation^{199, 203}.

Shift Work

A crucial constituent of the sleep regulation machinery, along with the homeostatic drive, is the circadian system. The endogenous biological clock, entrained to the 24-hour day-night cycle primarily by light, contributes to modulate sleep timing and duration through melatonin secretion. As most biological functions and behaviors also express circadian oscillations, it follows that distortion of biological rhythms has a broad, negative impact on health, as manifested in shift work.

Shift workers, who represent more than 28% of American workers²⁰⁴, experience a mismatch between behavioral rhythms, such as rest/activity, sleep/wake and fasting/feeding cycles, and biological rhythms, with the latter continuing to be synchronized to the external dark/light cycle. Such misalignment adversely affects risk status, and may raise susceptibility to hypertension. Relative to day-workers, those working alternative shifts have been found to be more likely to have high BP^{205, 206} and exhibit aberrant 24-hour BP pattern and dampened dipping during sleep²⁰⁷, although there are reports of no association^{208, 209}. More robust data indicate that shift work predisposes to new-onset hypertension and to progression towards more severe stages^210–212. Interestingly, the impact of circadian disruption is long-lasting, as this excessive risk of hypertension in shift workers persists even several years after retirement²¹³. By reviewing cross-sectional and longitudinal studies on this subject, a recently published meta-analysis estimated the pooled adjusted odd ratios for prevalent and incident hypertension in shift workers to be 1.10 (1.00–1.20) and 1.31 (1.07–1.60), respectively²¹⁴. However, when segregating rotating or fixed night work schedules, it emerged that these associations were substantially stronger and significant only among rotating shift workers. This conceivably reflects more deleterious consequences of inconsistent activity and sleep schedules, and related exposure to light, which impede re-entrainment and foster a chronic circadian mismatch.

Observational data indicate that shift workers have abnormal diurnal profiles of catecholamine excretion in comparison to day-workers²¹⁵. Furthermore, pre-post studies have found defective flow-mediated vasodilation²¹⁶, sympathetic cardiac predominance and increased inflammation²¹⁷ after one night of shift work. Additional insights on the pathways linking shift work and hypertension risk arise from experimental manipulations of circadian rhythms designed to mimic shift work. Mismatching sleep/wake and fasting/feeding behavioral cycles from internal circadian oscillations causes BP elevation, with possible mechanistic implications of increased inflammatory markers and augmented insulin and glucose levels^218–221. A mediating role of sleep curtailment and related pathophysiology is also plausible, as sleep is largely disrupted in shift workers and a degree of sleep loss is evident also during controlled conditions of experimentally-induced circadian misalignment. In this respect, an in-laboratory protocol simulating shift work compared sleep restriction with circadian misalignment versus sleep restriction with circadian alignment and found greater heart rate acceleration and catecholamine excretion in the former group²²², therefore suggesting that sympathoexcitation may contribute to the heightened risk of hypertension in shift workers.

In terms of remediation strategies, timed exposure to bright light and exogenous melatonin supplementation have been proven effective in resetting the master clock, thus favoring circadian adaptation and synchronization between biological and behavioral rhythms^{223, 224}. The possible protective effect of such tactics against hypertension risk in shift workers needs further investigation. Other strategies, such as assigning habitual early risers (larks) to the early morning shift, and habitual late sleepers (owls) to the late night shift, also merit exploration.

In summary, observational data suggest that shift workers are at risk of new-onset and severe hypertension even after retirement from shift work; the effects on BP may also be more prominent in rotating versus night shift workers. Multiple pathophysiological mechanisms such as sympathetic overactivity, inflammation, metabolic dysregulation and sleep deficiency have been implicated in the relationship between shift work and hypertension but additional work is needed to clearly elucidate the individual contribution of each of these factors. Studies investigating the effects of melatonin, bright light and other strategies employed in the treatment of circadian sleep-wake phase disorders on hypertension risk are required.

CONCLUSIONS

During normal physiologic sleep, there is fluctuation in autonomic activity, with intermittent BP increases during NREM sleep with arousals and with increased sympathetic activity and surges in BP in REM sleep. There appears to be an association between disrupted sleep and hypertension. Most of the literature has examined the link between OSA and hypertension but there is a growing body of literature evaluating other sleep disorders, in particular insomnia and insufficient sleep, in relation to hypertension.

OSA and hypertension are commonly comorbid conditions with shared risk factors. OSA is associated with prevalent and incident BP in a dose-response fashion, and treatment of OSA has been shown to lead to a modest reduction in 24-hour systolic and diastolic BP in those with hypertension. Sympathetic overactivity predominantly resulting from increased peripheral chemoreflex sensitivity to chronic intermittent hypoxia, increased sympathetic activity in relation to arousals, impaired baroreflexes, activation of the renin-angiotensin-aldosterone system, oxidative stress and inflammation are some of the factors that might explain the underlying pathophysiological mechanisms linking OSA and hypertension. OSA is associated with non-dipping of BP at night and is one of the most common causes of resistant hypertension; treatment of OSA may reverse non-dipping and facilitate BP control in resistant hypertension. There is some literature to suggest that the effects of CPAP treatment of OSA on BP may be especially pronounced in those with excessive daytime sleepiness.

The widespread prevalence of insufficient sleep and insomnia, along with their consequences in terms of heightened hypertension risk, justify the urgency of developing and implementing preemptive and corrective public health interventions. Normalization of sleep times and optimization of sleep quality may yield significant benefits for both maintenance of health and prevention of disease.