Sleep apnoea as an independent risk factor for cardiovascular disease: current evidence, basic mechanisms and research priorities

W. T. McNicholas, M. R. Bonsignore, the Management Committee of EU COST ACTION B26


Considerable evidence is available in support of an independent association between obstructive sleep apnoea syndrome (OSAS) and cardiovascular disease, which is particularly strong for systemic arterial hypertension and growing for ischaemic heart disease, stroke, heart failure, atrial fibrillation and cardiac sudden death.

The pathogenesis of cardiovascular disease in OSAS is not completely understood but likely to be multifactorial, involving a diverse range of mechanisms including sympathetic nervous system overactivity, selective activation of inflammatory molecular pathways, endothelial dysfunction, abnormal coagulation and metabolic dysregulation, the latter particularly involving insulin resistance and disordered lipid metabolism.

The present report, which arose out of a European Union Cooperation in the field of Scientific and Technical Research (COST) action on OSAS (COST B26), reviews the current evidence for an independent association and proposes research priorities to identify the underlying mechanisms involved, with a view to identifying novel therapeutic strategies.

Large-scale collaborative studies of carefully defined patient populations with obstructive sleep apnoea syndrome, adequately controlled for potential confounders, are needed. Such studies carry the prospect of evaluating potential interactions between different basic mechanisms operating in obstructive sleep apnoea syndrome and cardiovascular disease, and interactions with other related disorders, such as obesity, diabetes and dyslipidaemia. Furthermore, translational studies involving cell culture and animal models linked to studies of obstructive sleep apnoea syndrome patients are necessary to integrate basic mechanisms with the clinical disorder.

Obstructive sleep apnoea syndrome (OSAS) is characterised by repeated episodes of upper airway obstruction during sleep, associated with increasing respiratory efforts, intermittent arterial oxygen desaturation, systemic and pulmonary arterial blood pressure surges and sleep disruption. The main symptoms of OSAS are nocturnal respiratory pauses interrupted by loud intermittent snoring and excessive daytime somnolence. According to the recently updated International Classification of Sleep Disorders published by the American Academy of Sleep Medicine, a diagnosis of OSAS can be made if the respiratory disturbance index (RDI) is ≥15, independent of occurrence of symptoms, or whenever an RDI >5 is associated with any of the following: 1) sleep attacks, excessive daytime sleepiness (EDS), unrefreshing sleep, fatigue or insomnia; 2) awakenings with a choking sensation; or 3) witnessed heavy snoring and/or breathing pauses referred by the partner 1. The definitions of sleep-related respiratory disturbances have been clarified in recent years, particularly apnoea, hypopnoea and respiratory effort-related arousals 13. The gold standard technique for the diagnosis of sleep apnoea and related disorders is overnight polysomnography, although increasing attention is being paid to the development of limited diagnostic systems for use in clinical practice, which assess cardiorespiratory variables during overnight studies. These latter systems require a lower level of logistical support than polysomnography and many are suitable for home-based sleep studies.

Current knowledge on the natural history of the disease is still limited 4, but the long-term consequences of OSAS appear relevant. Untreated OSAS increases the risk for car accidents 5, and worsens quality of life 6 and mood 7, 8. The major health risk in OSAS patients, however, is the strong association with acute cardiovascular events (i.e. stroke, myocardial infarction and nocturnal sudden death) and chronic conditions such as systemic hypertension, coronary artery disease and heart failure 9.

OSAS can be effectively treated by applying nasal continuous positive airway pressure (CPAP) during sleep 1012. However, although CPAP technology has improved considerably over the years, acceptance of CPAP therapy remains a problem in patients without excessive daytime somnolence. Other options are available such as oral appliances 13 and upper airway surgery 14 but their use is not firmly evidence based as prospective randomised trials on current therapeutic and preventive efficacy are still lacking 1315. Therefore, the choice of therapy and the impact of preventive strategies on the long-term health effects of OSAS remain open questions. From a public health perspective, epidemiological and clinical studies on OSAS are cost-effective investments, since CPAP treatment decreases the excess healthcare costs for cardiovascular disease 1618 and car accidents 19, 20 incurred by OSAS patients before diagnosis. Moreover, the cost effectiveness of early OSAS management increases with OSAS severity 21.

OSAS is a highly prevalent disease in the population, affecting ≥4% of males and 2% of females 22. Population studies have shown that both sleep apnoea and daytime somnolence are frequent in the general population and increase with age 23, 24. The typical OSAS patient is male, middle-aged and obese. Obesity unfavourably affects respiratory function and may promote the collapse of upper airways during sleep 25; in addition, when body weight increases, the frequency of respiratory events during sleep also rises 26. In the USA, the current epidemic of obesity has been estimated to account for up to 40% of cases of sleep-disordered breathing (SDB) in adults 27; whether such figures might also be applicable in non-USA populations is unclear. For example, the correlation between obesity and the apnoea/hypopnoea index (AHI) is weaker in Asian than in Caucasian subjects, suggesting that racial factors may affect OSAS prevalence through craniofacial structure 28, 29. In Europe, the prevalence of obesity is lower than in the USA, with a higher proportion of relatively lean OSAS patients referred to sleep clinics. In summary, while the association of OSAS and obesity is unquestionable, the pathophysiology and health consequences of mild-to-severe OSAS in nonobese patients are still incompletely defined.

Besides decreasing lung function, obesity increases cardiovascular risk, making it hard to assess the independent role of OSAS on cardiovascular morbidity and mortality. The inter-relationships between obesity and OSAS are complex and possibly bidirectional 30, and some authors believe that OSAS should be considered as one manifestation of the metabolic syndrome (i.e. a cluster of cardiovascular risk factors) rather than a “local” disease principally affecting the upper airways 31.

From a clinical viewpoint, a pragmatic approach to treating OSAS and preventing cardiovascular events would be desirable. Even if OSAS exerted no independent effect on the prevalence or severity of cardiovascular disease, the cluster of risk factors in OSAS patients is often so impressive that it would warrant active intervention, even without knowing the precise relative role played by each factor. In other words, given the frequent association of OSAS and other major risk factors, the best approach could be to aggressively address not only respiratory disturbances during sleep but also the metabolic, inflammatory and cardiovascular problems of OSAS patients. This would require a highly organised and motivated effort, by both medical and social institutions.

European research has contributed to pivotal, top-quality research on SDB since the 1960s, but current clinical and basic research investments on OSAS in the European Union (EU) are very disappointing. The impact of low institutional funding is worsened by scarce investments by the pharmaceutical industry, since OSA treatment is not “pill based”. In addition, the multidisciplinary nature of the disease probably contributes to dispersion of funding among different medical specialties (neurology, respiratory medicine, cardiology, endocrinology and obesity centres, otolaryngology and dentistry) without efficient coordination at the European level. Conversely, in North America, at least four large population studies (Wisconsin Sleep Cohort, Sleep Heart Health Study, Pennsylvania Sleep Cohort and Cleveland Family Study) have provided cross-sectional and longitudinal data to assess the natural history of OSAS and collect long-term results on cardiovascular morbidity and mortality 22, 3236. Differences in lifestyle and prevalence of obesity between Europe and North America 37 highlight the necessity to develop research programmes specifically focused on the European population.

Unfortunately, the current status of European research on OSAS reveals the same problems recently highlighted for diabetes, another major disease for which European research is not sufficiently supported, despite the high social and health costs. The problems for diabetes are that funding resources are insufficient, effective central coordination is lacking, and the efficiency of investments in Europe is poor compared with the USA 38. A European initiative for diabetes (EURADIA) has been jointly developed by scientific societies and the pharmaceutical industry to increase the chances of reversing the European lack of competitiveness in this research area.

A similar joint effort of the European Scientific Community dealing with sleep medicine resulted in the proposal of an Action on OSAS through the EU Cooperation in Scientific and Technological research (COST) programme. The most important feature of the COST programme is the “bottom-up” approach, which reflects demands directly raised by the scientific community and helps create and strengthen scientific relationships and disseminate information on the Internet 39. The Action B26 on OSAS was approved in March 2005 and started on May 31, 2005 40. The first objective, jointly developed with the European Respiratory Society, has been to review the current state of the art in order to indicate strategically relevant directions for OSAS research in Europe. Additional connections have been established or are planned with other scientific societies focusing on sleep, cardiovascular disease and hypertension to foster multidisciplinary discussion on OSAS research and aim at providing effective prevention and treatment of OSAS and its consequences.

The programme of the COST B26 Action is to promote the integration of European research on OSAS to better understand: 1) the genetic and cell biological mechanisms by which intermittent hypoxia is detrimental for cardiovascular health; and 2) the mechanisms responsible for highly variable levels of EDS in OSAS patients. In addition, the Action will address genetic and epidemiological issues applied to the European population as useful steps to better identify patients at particularly high cardiovascular risk, and to develop further alternatives to CPAP therapy.

The current review discusses: 1) the epidemiologic and clinical evidence of OSAS associated cardiovascular involvement; and 2) the current evidence on pathogenetic mechanisms that affect subjects with untreated OSAS, in order to identify “hot” topics and propose priorities for future research.


Early concerns about possible consequences of SDB in patients with severe OSAS focused on the possibility that nocturnal hypoxaemia might cause pulmonary hypertension and cor pulmonale. However, pulmonary arterial pressure is only modestly elevated during sleep and wakefulness in subjects with uncomplicated OSAS 41. Rather, the systemic circulation was soon recognised as a major target of OSAS. As early as 1980, the San Marino epidemiological study by Lugaresi et al. 42 highlighted the association of systemic hypertension and snoring in the general population. Ten years later, a high prevalence of cardiovascular disease (systemic hypertension, coronary artery disease and cerebrovascular disease) in OSAS patients at diagnosis, and a dose–response effect between cardiovascular involvement and OSAS severity, was reported by the Stanford group 43.

Nowadays, evidence for an association of OSAS with several aspects of cardiovascular disease and/or risk is available. The literature on the different aspects of this topic is extensive, and a full discussion of all relevant papers is beyond the scope of the current review. Reference to extensive reviews, which the interested reader is referred to, is provided at the beginning of each of the following sections, while the most recent papers on each topic are reported in detail.

Cardiovascular morbidity and mortality

According to uncontrolled studies, untreated moderate-to-severe OSAS was associated with increased rates of nonfatal cardiovascular events after a relatively short follow-up 43, 44. At least four longitudinal studies have confirmed increased cardiovascular morbidity in OSAS patients 4548, but the concomitant occurrence of other cardiovascular risk factors often limited the assessment of an independent pathogenetic role for OSAS 49. A prospective 5-yr study analysed cardiovascular outcomes in 400 subjects according to snoring status and the presence of cardiovascular risk factors, and reported a significant increase in the number of fatal and nonfatal cardiovascular events in subjects positive for both snoring and other well-established risk factors present at study entry. Conversely, either simple snorer status or the isolated occurrence of cardiovascular risk factors increased the risk only slightly 45. In consecutive subjects with suspected OSAS and no clinical evidence of cardiovascular disease, OSAS patients who refused treatment had a higher incidence of cardiovascular events in the 7 yrs following diagnosis compared with non-OSAS subjects and patients compliant with CPAP therapy 46. Overall, the bulk of available data indicate a high risk for cardiovascular events in untreated OSAS.

However, the association of OSAS and cardiovascular disease does not necessarily imply a cause–effect relationship. OSAS patients are often obese and show evidence of the metabolic syndrome (i.e. at least three of the following cardiovascular risk factors: central obesity, systemic hypertension, low high-density lipoprotein cholesterol (HDL-C), high triglycerides, and impaired glucose tolerance). Such a clinical context makes it difficult to assess the independent effects of OSAS on cardiovascular risk 49, 50.

Two recent studies assessed cardiovascular prognosis in OSAS over 10 yrs after diagnosis 47, 48. Patients were free to accept or refuse CPAP treatment, but neither study was randomised due to the ethical unacceptability of withholding effective CPAP therapy for such a long time. The study by Marin et al. 47 showed that long-term cardiovascular morbidity and mortality increased only in patients with untreated severe OSAS, whereas simple snorers, OSAS patients with mild disease or patients with severe OSAS who accepted CPAP treatment showed morbidity and mortality figures very similar to those obtained in the general population. The study by Doherty et al. 48 instead, suggested that untreated OSAS may increase the severity rather than the prevalence of cardiovascular disease. Indeed, the incidence of hypertension, ischaemic heart disease and other cardiovascular disorders during follow-up was not significantly different in treated and untreated patients, irrespective of acceptance or refusal of CPAP treatment. However, only untreated patients showed excess cardiovascular mortality during follow-up 48. Finally, an increased risk for death or stroke and a dose–effect relationship between OSAS severity and risk were reported in a large patient series followed for a median of 3.4 yrs. Unfortunately, the short duration of follow-up and the small number of observed events did not allow the specific assessment of the effects of therapy 51.

Data on cardiovascular mortality further indicate the need to effectively treat severe OSAS in order to avoid potentially catastrophic cardiovascular consequences, especially in young patients. The first observational reports on cardiovascular mortality in OSAS 52, 53 were confirmed by later studies. Before the introduction of CPAP into clinical practice, patients with mild-to-moderate OSAS who had been conservatively treated and encouraged to lose weight showed a higher 5–8-yr cardiovascular morbidity and mortality compared with patients who had been tracheostomised because of very severe OSAS. In the latter group, survival was similar to that of the general population 52, 53. The cardiovascular effects of OSAS appear particularly dangerous in young subjects 53, as recently confirmed after long-term follow-up of almost 15,000 patients 54. The patient cohort analysed by Marin et al. 47 also showed high mortality rates in the youngest age groups. Conversely, CPAP-treated patients who were compliant with treatment had mortality rates similar to those recorded in the general population 55, whereas excess cardiovascular fatal events occurred in patients aged <50 yrs who refused CPAP treatment 55, 56. Indeed, the major determinant of long-term outcome was compliance with CPAP treatment rather than severity of OSAS at diagnosis 57.

OSAS may also be involved in the pathogenesis of nocturnal sudden death. In 1991, Seppälä et al. 58 reported that increasing snoring severity was associated with increased risk for nocturnal sudden death, but the data were only indicative of a possible association of OSAS and nocturnal events since they were not based on polysomnographic recordings. In patients with polysomnographic diagnosis of OSAS, Gami et al. 59 found that the risk of nocturnal (between 00:00–06:00 hrs) sudden death increased with OSAS severity, whereas cardiovascular events mostly occurred between 06:00–12:00 hrs in patients with no sleep apnoea or subjects from the general population. The role of CPAP treatment could not be assessed in this study, but the authors proposed that the good compliance with treatment expected in patients with severe OSAS suggests that CPAP probably had little effect on the results. This proposal is at variance with the significant protection reported by CPAP treatment in several other observational studies, and deserves prospective and carefully designed long-term studies before definitive conclusions can be drawn.


There is strong evidence that OSAS is an independent risk factor for systemic hypertension 9, 6062. Careful case–control studies have confirmed the association of sleep apnoea and increased blood pressure independent of confounders such as obesity 63. Also, data in patient and population samples extensively support a role for OSAS in the pathogenesis of hypertension 46, 6470. In longitudinal population studies, SDB increased the risk for increased blood pressure at follow-up 71, 72. Prospective randomised trials have shown that OSAS treatment decreases blood pressure 73, 74, especially in hypertensive subjects 7577.

The role of intermittent hypoxia in the pathogenesis of systemic hypertension is unclear. A causal relationship between intermittent hypoxia during sleep, systemic hypertension and cardiac hypertrophy has been convincingly demonstrated in a chronic dog model of OSAS 78, 79 and in rats exposed to intermittent hypoxia 80. However, a recent randomised controlled study in OSAS patients showed decreased blood pressure after effective CPAP treatment, but not after sham-CPAP together with supplemental nocturnal oxygen for 2 weeks 81. Therefore, while the role of OSAS in increasing blood pressure can be considered proven, the independent role of intermittent hypoxia on blood pressure regulation in humans still awaits confirmation.

Systemic hypertension in OSAS is a major clinical problem, especially as it is greatly underdiagnosed. In one study, the prevalence of hypertension was 67% in newly diagnosed OSAS patients not known to be hypertensive before undergoing assessment for SDB 82. Notably, the role of OSAS as a cause of underdiagnosed hypertension has been recently acknowledged 83. The isolated increase in diastolic blood pressure might be the earliest hypertensive change associated with OSAS 82, 84, whereas isolated systolic hypertension was uncommon in subjects with SDB 85, 86. Instead, a high prevalence of systolic hypertension was found in patients with OSAS and chronic heart failure 87. Most OSAS patients, however, show elevated systolic and diastolic blood pressure values 85, 86, as well as increased blood pressure variability during sleep 88 and decreased baroreflex sensitivity 82, 89. Overall, hypertension, increased blood pressure variability and decreased effectiveness of cardiovascular control mechanisms may all contribute to the increased cardiovascular risk of OSAS.

Assessing the independent pathogenetic role of OSAS in left ventricular hypertrophy is complicated by multifactorial influences potentially affecting cardiac structure, such as hypertension, obesity, intermittent hypoxia and mechanical changes associated with apnoeas 90, 91. There is agreement, however, that patients with moderate-to-severe OSAS show an increased prevalence of left ventricular hypertrophy 82, 9095 and diastolic dysfunction 96, 97, both being reversed by CPAP treatment.

The relationship between SDB and systemic blood pressure appears variable in paediatric OSA 98100 and in population studies of snoring children 100, 101. Children with OSAS showed left ventricular diastolic dysfunction with increasing apnoea severity 102, 103, autonomic dysfunction 104, increased circulating inflammatory markers 105107 and adhesion molecules 106, suggesting similar potential pathogenetic mechanisms of cardiovascular damage in adult and paediatric SDB. More data are clearly needed to better understand the immediate and long-term cardiovascular impact of sleep apnoea in children 108, 109.

Coronary artery disease

The association of obstructive sleep apnoeas and coronary artery disease (CAD) has been documented in patient series selected for either disease 110112 and in a cross-sectional analysis of the Sleep Heart Health Study cohort data 113. Reviews on the relationship between ischaemic heart disease and sleep apnoea are available 114116.

Untreated OSAS worsens the prognosis of patients with CAD. In addition to the increased incidence of nocturnal sudden death in patients with OSAS compared with the general population 58, 59, patients with known CAD and an RDI >10 events·h-1 were much more likely to experience cardiovascular death over a 5-yr period than those with low RDI (37.5 versus 9.3%, respectively) after controlling for age, weight and smoking 116. In longitudinal studies, untreated SDB significantly worsened the prognosis of patients with documented CAD 117, whereas patients with CAD and OSAS receiving CPAP had a better clinical course compared with those who refused treatment 118. An increased incidence of CAD in patients clinically free of coronary symptoms at the time of OSAS diagnosis, and a protective effect of CPAP treatment, have been recently reported by a prospective 7-yr follow-up study in the Gothenburg Sleep Cohort 119. Moreover, the frequency of nocturnal desaturations correlated with the extent of coronary lesions and explained 13.4% of their variance, suggesting a pathogenetic role of OSAS in coronary atherosclerosis 120. Some studies, however, did not confirm that OSAS worsened the prognosis of patients with CAD, as 10-yr survival did not differ between apnoeic and nonapnoeic subjects in a small series of 50 coronary patients 121.

OSAS may precipitate angina or myocardial ischaemia during sleep in patients with coronary lesions 115. Given the high prevalence of both OSAS and CAD, the frequency of nocturnal anginal symptoms was found to be surprisingly low in OSAS patients 122. Nocturnal angina occurred in <1% of 4,000 patients studied by polysomnography and eight-channel electrocardiogram monitoring 122. In addition, the symptomatic association of OSAS and CAD has been documented only by case reports or small patient series 115. In summary, available data in patients with OSAS and nocturnal angina agree that the extent of coronary lesions is variable and angina symptoms are resolved with CPAP treatment 115. Serum cardiac troponin T, a marker of myocardial damage, did not show evidence of injury in a small series of CAD patients with OSAS at baseline or during acute CPAP application 123. It is likely that some still unknown additional predisposing factors are necessary for OSAS to trigger significant coronary insufficiency during sleep. Experimental data in rats indicate that daily exposure to intermittent hypoxia (IH) for 8 h·day-1 for 35 days increased infarct size after ischaemia–reperfusion, in the absence of significant changes in arterial blood pressure in vivo or evidence of vascular/endothelial dysfunction 124. However, the modality of exposure to IH may significantly affect the myocardial response to ischaemia–reperfusion 125, suggesting that studies at the molecular level are necessary to better understand clinical data obtained in humans.

Asymptomatic electrocardiographic changes may occur during sleep and indicate detrimental effects of obstructive apnoeas on the heart. In a group of 23 OSAS patients without symptoms or history of CAD, about one-third showed asymptomatic ST-segment depression during sleep, but only one patient was also positive at the exercise stress test, suggesting a low prevalence of symptomatic coronary artery disease in OSAS 126. Indeed, isolated apnoea-associated ST-segment changes could be nonspecific (i.e. caused by post-apnoeic hyperventilation). Conversely, up to one-third of OSA patients with angiographically proven coronary lesions showed ST-segment changes during sleep, correlated with the severity of hypoxaemia 127, 128 or measures of sleep fragmentation 129. In a large clinical series of CAD patients, episodes of ST-segment depression occurred in about one-third of the sample, but a tight temporal relationship between apnoeas and ECG changes could not be demonstrated 130, supporting the interpretation that additional factors are likely to be involved in the pathogenesis of symptomatic coronary events in OSAS patients 131. At present, the clinical meaning and relevance of the described changes remain undefined 132.

In summary, epidemiological and clinical studies show that the association of OSAS and CAD is frequent and possibly relevant for prognosis. The mechanism(s) by which OSAS exerts its detrimental effect remain to be established and future studies should actively pursue the identification of OSAS patients at high cardiovascular risk.


Epidemiological data suggest a strong relationship between OSAS and acute cerebrovascular events. The prevalence of SDB in patients with acute stroke ranges 44–72% in different series, and after the acute phase still remains higher than in the general population 133, 134. The evidence for an association between stroke or transient ischaemic attacks and a history of heavy snoring or polysomnographic evidence of obstructive sleep apnoeas has been recently summarised in patients with acute stroke 133, 134. Moreover, the potential value of obtaining a polysomnographic recording 135 and a cost-effectiveness analysis of OSAS diagnosis and treatment 136, 137 has also been carried out in patients with acute stroke.

Prospective data from the Wisconsin Sleep Cohort clearly indicate that an AHI ≥20 was associated with a four-fold increase in the risk of stroke during a 4-yr follow-up study, supporting the view that SDB precedes, and does increase the risk for, the occurrence of stroke after OSAS diagnosis 138. In the Caerphilly Cohort (UK), a high risk for stroke over a 10-yr follow-up was found in males reporting more than one of the following symptoms: snoring, witnessed apnoeas, daytime sleepiness, insomnia, and restless legs 139. Finally, increasing incidence of stroke and death with OSAS severity was found in a cohort of OSAS patients after a median follow-up of 3.4 yrs 51.

Some uncertainties on the role of OSAS in the pathogenesis of stroke are due to the fact that SDB can precede but also follow occurrence of stroke 140. A pathogenetic involvement of OSAS in cerebrovascular disease is suggested by the direct relationship found between the severity of nocturnal oxygen desaturation and intima-media thickness and/or the occurrence of atherosclerotic plaques in the carotid arteries of OSAS patients, independent of occurrence of hypertension 141, 142. Moreover, central but not obstructive sleep apnoeas decrease during recovery after a stroke or a transient ischaemic attack, suggesting that obstructive events are most likely to pre-date the cerebrovascular event 143. However, an independent association of OSAS with transient ischaemic attacks has been reported in some but not all studies 143145, and any temporal relationship between OSAS and cerebrovascular events should still be considered as undefined.

Given the established association of OSAS and hypertension, it is quite surprising that little is known on the time course of blood pressure in patients with stroke, according to the presence or absence of SDB. According to a recent report, the mean 24-h blood pressure level in patients with acute stroke positively correlated with severity of sleep apnoea, while the nondipping status (i.e. a nocturnal reduction <10% of the daytime blood pressure value) was associated with the severity of stroke 146. These data refer to a time period when central events are prevalent, and no data are available on circadian blood pressure profile in patients with OSAS after stroke.

OSAS could worsen prognosis after a stroke episode, as suggested by the high mortality of patients with an AHI >30 after stroke 147, and the improvement in 18-month survival of those OSAS patients (AHI ≥20) who were able to tolerate CPAP treatment after stroke 148. In addition, sleep apnoea may increase the risk of recurrence of ischaemic stroke 149. Patients with stroke and sleep apnoea show more severe functional impairment and longer hospitalisation during rehabilitation compared with patients without SDB 150.

Recently, Bassetti et al. 151 reported follow-up data in a large series of stroke patients. The study found that SDB (AHI ≥30): 1) was associated with diabetes, night-time stroke onset and macro-angiopathy as a cause of stroke; 2) improved after the acute phase; 3) was associated with increased post-stroke mortality; and 4) could be treated with long-term CPAP in ∼10% of patients. This patient sample cannot be considered as representative of an unselected population of patients with stroke, possibly explaining some differences compared with other studies. However, the observations provided could serve to design a large randomised trial on the effects of treatment of OSAS in patients with cerebrovascular disease 151.

Careful case–control studies have tested whether OSAS might be associated with silent cerebrovascular disease, with mixed results 152, 153. Brainstem white matter disease was hypothesised to be more common in OSAS than non-OSAS patients, but it correlated with arousal frequency only, in almost 800 elderly participants of the Sleep Heart Health Study 154, and the clinical significance of this remains unclear. Conversely, altered metabolism in the frontal white matter was shown in OSAS patients with and without clinically evident cardiovascular involvement, suggesting that OSAS affects brain function independent of vascular damage 155.

In summary, evidence for the association of SDB with cerebrovascular disease mostly relates to the pathophysiology of acute events. Current knowledge is still limited on the role of long-term SDB and on possible modifications of the natural history of stroke by CPAP treatment in patients with both SDB and stroke. Finally, even though not necessarily implying a stroke, OSAS might predispose to death during sleep in critically ill patients by increasing arousal threshold and lengthening apnoeic episodes causing cerebral hypoxaemia 156.

Heart failure

The relationship between OSAS and heart failure (HF) is complex. On the one hand, obstructive sleep apnoeas negatively affect cardiac function acutely and may cause cardiac remodelling in the long term, thus possibly contributing to the pathogenesis of HF. On the other hand, HF could contribute to the pathogenesis of SDB in the form of either obstructive or central apnoeas 157160. Assessing the impact of the association of OSAS and HF is clinically relevant, due to the rising prevalence of HF and the availability of effective treatment for OSAS. However, whether CPAP treatment may favourably affect long-term prognosis and quality of life in patients with OSAS and HF is still unknown as available data have only reported a 1–3-month follow-up 161.

Relatively few studies are available on the association of OSAS and HF. In studies of patients with HF, the prevalence ranged 11–53% 162165. Cross-sectional data from the Sleep Heart Health Study showed a strong association of SDB with HF (odds ratio for upper versus lower quartile 2.38, confidence interval 1.22–4.62) 113.

Obstructive sleep apnoeas acutely affect cardiac function. Intrathoracic pressure swings during respiratory efforts increase venous return and ventricular afterload, while sympathetic activation secondary to hypoxia and arousal increases blood pressure and myocardial oxygen consumption at the end of apnoeas. The increase in cardiac load during obstructive apnoeas also occurs in the normal heart as shown in animal models 166, but its consequences are particularly striking in the failing heart. Accordingly, both the Mueller manoeuvre and obstructive apnoeas have been shown to acutely depress cardiac function especially in patients with heart failure 167, 168; these effects were acutely reversed by CPAP application 168.

In the long term, OSAS favours the development of systolic dysfunction 166, 169. In randomised controlled studies in patients with HF and OSAS, CPAP treatment for a few weeks increased left ventricular ejection fraction and decreased blood pressure and sympathetic activation, strongly suggesting a pathogenic role of OSAS in worsening cardiac function 170172. However, whether the positive effects of CPAP on the cardiovascular system may also imply an improvement in long-term prognosis of patients with OSAS and HF is still unknown 173. An interesting point, critically discussed by Arzt et al. 161, is the clinical relevance of EDS in the therapeutic approach of patients with HF and OSAS. An improved quality of life after CPAP treatment has only been documented in OSAS patients with EDS 171. More data on patients with HF and OSAS but no EDS are needed to improve clinical decision making for the treatment of OSAS. In addition, interpreting clinical data is complicated by the increased prevalence of HF with the progressive ageing of the population and by the rapid evolution in the treatment of HF. At present, the impact of OSAS on long-term survival appears small in patients with severe HF 173 or in elderly subjects 35, 174. Therefore, large clinical studies in well-characterised samples are necessary to obtain conclusive information on the prognostic role, if any, of OSAS in HF.

OSAS may contribute to the pathophysiology of HF by potentiating sympathetic activation. Patients with HF and sleep apnoea show higher daytime muscle sympathetic nerve activity (MSNA) compared with patients with HF but no sleep apnoea 175. CPAP treatment in patients with HF and OSAS decreased daytime MSNA, systolic blood pressure and heart rate 176, suggesting a significant contribution of OSAS to increased central sympathetic outflow. However, other studies on noradrenaline spillover in HF patients without apnoeas, with OSAS or with central sleep apnoeas (CSA), respectively, found similar spillover rates in the first two groups, while increased spillover in patients with CSA was entirely accounted for by a more severe cardiac disease 177. While differences in patient characteristics or methodology may account for different results among studies, this area deserves further investigation because of its potential clinical impact on the treatment of HF.


Since the earliest clinical observations, OSAS has been recognised as a potential cause of arrhythmias during sleep 9, 178181. Bradyarrhythmic episodes occur in OSAS patients, possibly reflecting reflex parasympathetic activity evoked by apnoeas; Guilleminault et al. 182 reported bradyarrhythmias in 18% of OSAS patients and Becker et al. 180 reported heart block episodes in 20% of patients with OSAS, especially when OSAS was severe. Subsequent studies from the same group confirmed these findings and underlined the trend for bradyarrhythmias to occur in REM sleep and disappear during CPAP treatment 183, 184. More recently, recordings obtained for several consecutive weeks have documented bradyarrhythmia in 47% of patients with untreated OSAS with a large intra-individual variability accounting for the higher prevalence compared with previous studies 185 and possibly for some negative results by previous investigations based on single-night Holter recordings 186. Furthermore, CPAP therapy abolished pathological dysrhythmias in OSAS patients within a period of 24–48 h 187; long-term recordings also documented disappearance of arrhythmias after a few weeks of CPAP treatment 184. A high prevalence of OSAS was found in patients in whom Holter recordings showed isolated bradyarrhythmias during sleep 188. In such patients, a sleep study is warranted before implanting a pacemaker because CPAP therapy may be sufficient to prevent arrhythmias associated with OSAS 188.

However, not all studies reported a significant association between heart rhythm disturbances requiring implantation of a pacemaker and SDB 189. In addition, few studies are available on bradyarrhythmias associated with SDB in the general population. In subjects from the Sleep Heart Health Study, no significant difference was shown for conduction delay arrhythmias between SDB-exposed and -nonexposed groups, but second-degree atrioventricular block type 1 and the percentage of pacemaker-implanted subjects tended to be higher in the group with an AHI ≥30 190.

Other studies have tested the hypothesis that OSAS might increase the prevalence of nocturnal ventricular arrhythmias through intermittent hypoxia during sleep. The only population study on this topic reported increased odds ratios for ventricular tachycardia and complex ventricular ectopy in patients with moderate-to-severe OSAS compared with non-OSAS subjects 190. In addition, ventricular premature beats decreased by 58% after 1 month of CPAP treatment in patients with OSAS and HF 191. These data are in agreement with the expected effects of hypoxaemia on the heart 179 and may help explain the pathogenesis of nocturnal sudden death in OSAS patients 58, 59.

The association of OSAS with atrial fibrillation, and the possibility that atrial pacing might improve SDB, are recent areas of clinical research. A high prevalence of OSAS has been found in patients with atrial fibrillation by some 192, 193 but not all studies 194. Case reports have shown the shift from sinus rhythm to atrial fibrillation during sleep in patients with untreated OSAS 195, and the disappearance of atrial fibrillation and heart block after initiation of CPAP treatment in severe OSAS 196. A pathogenetic role of OSAS is further suggested by the four-fold increase in the prevalence of atrial fibrillation in subjects with an AHI ≥30 reported by the Sleep Heart Health Study Investigators 190. Finally, the clinical relevance of the role of OSAS in atrial fibrillation is indicated by the high rate of recurrence of atrial fibrillation in inadequately treated OSAS patients 197. However, atrial fibrillation is also a risk factor for central sleep apnoea 163165, suggesting that its association with OSAS may reflect a general effect of sleep apnoeas rather than being specific for OSAS.

Atrial overdrive pacing was proposed as a possible therapeutic approach for sleep apnoea in an intriguing study by Garrigue et al. 198, who reported that AHI decreased when patients were subjected to pacing compared with baseline conditions. These results were not confirmed by subsequent studies 199203, even though the pathophysiology of cardiac pacing in SDB is still considered of interest, especially in HF patients 204, 205.


Besides the improved understanding expected from large clinical and epidemiological studies concerning the natural history of OSAS, and the largely unproven effects of treatment options other than CPAP, one of the most interesting and promising areas of current clinical and experimental research regards the inter-relationships between OSAS, energy metabolism 30, 31 and sleep deprivation 206. Complex hormonal interactions involve sleep and metabolism, and current lifestyle affects sleep–wake cycles, with the toll of progressively decreasing sleep times, at all ages, compared with previous decades. Several review papers have recently been published on the relationship between OSAS and the metabolic syndrome 30, 31, 206210, with special regard to the pathogenesis of increased cardiovascular risk. Increasing evidence indicates that OSAS and intermittent hypoxia may independently affect energy metabolism. The following subsections summarise data on: 1) insulin resistance in population and patient studies, and the effects of CPAP treatment; 2) the association of OSAS and diabetes; and 3) lipid metabolism and hepatic steatosis in OSAS.

Insulin resistance

Ip et al. 211 reported that markers of OSAS severity (AHI and minimum oxygen saturation) were associated with insulin resistance, whilst Punjabi et al. 212 reported that, in mildly obese but otherwise healthy males from the general population, SDB was associated with insulin resistance. Cross-sectional data from the Sleep Heart Health Study and the Wisconsin Sleep Cohort found similar results in population cohorts 213, 214. However, the longitudinal data from the Wisconsin Cohort could not demonstrate a clear relationship between SDB and the subsequent development of diabetes 214. Other studies in snorers have confirmed the association of SDB and insulin resistance or diabetes in Korean nonobese 215 and Indian males 216. In all studies, the association was independent of obesity, even though obesity is known to increase the risk for impaired glucose tolerance. Similar results were found in subjects with suspected OSAS 217219 and by some 220 but not all 221, 222 studies in children with SDB. In normal subjects, hypoxia appears to be an important contributor to glucose intolerance 223. Treatment with CPAP rapidly restored insulin sensitivity, especially in patients who were not obese 224, but a paradoxical increase in glucose level was also documented during acute CPAP application in nondiabetic obese OSAS patients 225. In summary, the bulk of evidence supports the view that OSAS worsens glucose metabolism, but this effect was reversible with treatment.

Type-2 diabetes

The association of OSAS and type-2 diabetes was reported much earlier than the studies on insulin resistance, but the evidence was limited to small case series or epidemiological studies using snoring as a surrogate marker for OSAS 226. More recent epidemiological studies have convincingly shown that type-2 diabetes is often associated with OSAS and daytime sleepiness 210213, 227229. For example, in a large series of OSAS patients, type-2 diabetes and impaired glucose tolerance showed a 30 and 20% prevalence, respectively 218, and studies in snorers reached similar conclusions 230232. In prospective studies, the risk of developing diabetes was higher in snorers than nonsnorers 233, 234, especially in obese subjects 233. However, insulin sensitivity improved after CPAP treatment in diabetic OSAS patients 235238. Sleep apnoeas are also frequent in children with type-1 diabetes 239 and adult patients with diabetic neuropathy 240, 241, but this probably reflects the consequence of abnormalities of ventilatory control associated with diabetes rather than a direct effect of obstructive apnoeas 242244.

Lipid metabolism and hepatic steatosis

Altered lipid metabolism and hepatic steatosis in OSAS have been recently studied. Total cholesterol tended to decrease after CPAP treatment 245. It is also possible that HDL-C in OSAS patients is functionally less effective in preventing low-density lipoprotein oxidation in vivo 246. Mean HDL-C increased after CPAP treatment and this change correlated with the decrease in AHI 247, suggesting reversibility of OSAS-associated changes in plasma lipids. As for the prevalence of hepatic steatosis in OSAS patients, insulin resistance and nonalcoholic steatohepatitis (NASH) often occur in obese and diabetic subjects 248, and OSAS might contribute to NASH pathogenesis similar to its role in insulin resistance. Available data, although limited, indeed suggest this might be the case. According to Singh et al. 249, the prevalence of OSAS was ∼50% in patients with nonalcoholic fatty liver disease, while severe OSAS represented a risk factor for increased liver enzymes and steatohepatitis independent of body weight 250. One in three obese OSAS patients showed abnormal serum aminotransferase levels, reverting towards normal values during prolonged CPAP treatment 251. In addition to clinical data, experimental studies also suggest a role of intermittent hypoxia in the pathogenesis of hyperlipidaemia 252, 253, supporting an independent role of OSAS, in addition to obesity, in the pathogenesis of NASH 254, 255.


The mechanisms underlying cardiovascular disease in patients with OSAS are still poorly understood. The pathogenesis is likely to be a multifactorial process involving a diverse range of mechanisms including sympathetic nervous system overactivity, selective activation of inflammatory pathways, endothelial dysfunction and metabolic dysregulation, with the latter particularly involving insulin resistance and disordered lipid metabolism.

Sympathetic nervous system overactivity

The repetitive episodes of upper airway obstruction that are characteristic of OSAS result in intermittent hypoxia and large swings in intrathoracic pressure that trigger autonomic responses including sympathetic nervous system overactivity. Many reports have demonstrated sympathetic overactivity in patients with OSAS. Increased urinary catecholamine levels have been reported in patients with OSAS with levels falling after treatment by tracheostomy 256. A direct link between hypoxaemia and elevated sympathetic activity has also been proposed 257259 and elevated MSNA was attenuated during apnoea when hyperoxic conditions were maintained 258. Furthermore, Narkiewicz et al. 260 have demonstrated a selective potentiation of peripheral chemoreflex sensitivity in patients with OSAS compared with normal controls. Other reports have indicated a significant fall in both plasma and urinary catecholamines following nasal CPAP therapy 261263. MSNA has also been directly measured by insertion of a tungsten microelectrode into the peroneal nerve. Using this methodology, an increase in MSNA following an acute apnoea associated with hypoxia has been observed 264, together with positive correlations between MSNA and plasma noradrenaline levels 265, 266. Treatment with nasal CPAP significantly lowered MSNA 267.

Furthermore, evidence in favour of a significant contribution to the pathogenesis of OSA-related cardiovascular complications, by alterations in autonomic cardiovascular control, has been obtained by techniques exploring spontaneous sensitivity of baroreflex control of the heart 268, 269. OSAS patients are characterised by reduced baroreflex sensitivity during both wakefulness and sleep 270 and such impairment can be reversed by CPAP. This improvement is particularly evident with chronic treatment 89, although a small but significant improvement can also be detected even after short-term CPAP application 271.

Support for the role of sympathetic overactivity in the pathogenesis of hypertension in OSAS also comes from animal models. In both dog and rat models of OSAS, an increase in blood pressure was found, which declined once the airway occlusion was abolished 78, 80. These blood pressure changes were not observed with recurrent induced arousals without airway occlusion, indicating that it was the obstructive events rather than the associated arousals that were responsible for the observed effects 78. These changes in blood pressure were prevented by pharmacological and surgical blockade of the sympathetic nerve system in a rat model of chronic intermittent hypoxia 272, 273.


Systemic inflammation plays an important role in all stages of atherosclerosis. It occurs in the vasculature as a response to injury, lipid peroxidation, and perhaps infection. Two well-recognised markers of systemic inflammation have been evaluated in OSAS, namely C-reactive protein (CRP) and tumour necrosis factor (TNF)-α. CRP, an important serum marker of inflammation, is synthesised by the liver and regulated by cytokines, particularly interleukin (IL)-6 274. Prospective studies have shown that CRP is a strong predictor of future coronary events in apparently healthy males and females 275 in addition to peripheral arterial disease 276. OSAS was reported to be associated with higher CRP and IL-6 levels in otherwise healthy subjects, and these levels correlated with OSAS severity, supporting an important role for inflammation in the cardiovascular pathogenesis of the disease 277280. Furthermore, treatment with nasal CPAP has been reported to be associated with decreased levels of these markers 278. However, recent studies have failed to find an association between CRP and OSAS and the relationship is now less clear 281, 282. It is likely that studies involving large numbers of patients, adequately controlled for potential confounding factors, particularly obesity, will be required to finally resolve the question of a possible independent association.

TNF-α is another inflammatory cytokine, which is regulated by the transcription factor nuclear factor (NF)-κB, a master regulator of inflammatory gene expression 283, and contributes to atherogenesis 284. TNF-α levels have been reported to correlate with cardiovascular risk 285. TNF-α levels are elevated in OSAS and fall with CPAP therapy 286, 287 and both T-cells and monocytes have been reported as potential sources of this cytokine 288. Furthermore, a gene polymorphism associated with increased TNF-α production has recently been reported to be more common in OSAS 289.

Recent evidence from a cell culture model of intermittent hypoxia supports a selective activation of inflammatory pathways over adaptive pathways in response to intermittent hypoxia, which contrasts with sustained hypoxia where activation of adaptive and protective pathways predominate 287. This preferential activation of inflammatory pathways may be a consequence of the intermittent re-oxygenation that is characteristic of intermittent hypoxia and thus represents a variant of reperfusion injury 290. The activation of inflammatory transcription factors by intermittent hypoxia/re-oxygenation (IHR) has also been demonstrated in a rat model and the authors found a significant correlation with the degree of IHR and neurocognitive function, and also an improvement with reversal of IHR 291.

Levels of circulating soluble adhesion molecules, which mediate adhesion of leucocytes to the vascular endothelium, such as intracellular adhesion molecule-1, are elevated in patients with OSAS and improve with CPAP therapy 292. Furthermore, increased adhesion of lymphocytes to vascular endothelial cells has been demonstrated in OSA patients compared with controls 293. Inflammatory cytokines, such as TNF-α and IL-8 induce the expression of cellular adhesion molecule 294, 295, thus providing further evidence of an important role for inflammation in the cardiovascular morbidity of OSAS.

Hypoxia also induces the activation of the adaptive pathway mediated by upregulation of the transcription factor hypoxia-inducible factor (HIF)-1. There is evidence of HIF-1-dependent gene activation in OSAS as indicated by increased levels of vascular endothelial growth factor 296, although another report indicates that OSAS associated with pure intermittent hypoxia (where interapnoea oxygen levels are normal) is not associated with elevated levels of another HIF-1-dependent gene, erythropoietin 287.

Oxidative stress

While there is evidence of increased release of reactive oxygen species in patients with OSAS 297, 298, as a likely consequence of intermittent re-oxygenation associated with recurring apnoea, the interaction with other molecular mechanisms such as inflammatory pathways has not been fully evaluated. Studies in rats have demonstrated that chronic intermittent hypoxia (CIH) results in oxidative stress that subsequently leads to left ventricular dysfunction 299. Similarly, CIH-associated oxidative stress has been shown to result in cortical neuronal cell apoptosis in mice 300. The oxidative stress-induced brain injury appears to be associated with hypersomnolence in a mouse model of CIH 301, 302. Oxidative stress may be responsible for reduced nitric oxide (NO) bioavailability, enhanced lipid peroxidation 303 and formation of isoprostanes 304, although there is recent evidence that CIH is associated with activation of inducible NO synthase in the brain 305. CIH has also been associated with reduced hypoglossal nerve output by oxidative stress 306. Furthermore, free radicals might upregulate transcription factors such as NF-κB and HIF 290.

Endothelial dysfunction

Vascular endothelium controls various vascular functions through regulation of vasoactive mediators in response to physical or biochemical stimuli and is the major regulator of vascular haemostasis. The endothelium maintains the balance between vasodilatation and vasoconstriction; if this balance is tilted towards vasoconstriction, endothelial dysfunction occurs, causing damage to the arterial wall. Endothelial dysfunction has been found to occur in response to cardiovascular risk factors and to precede or accelerate the development of atherosclerosis 284. Such dysfunction appears to have a predictive value for cardiovascular events in patients with chest pain and/or coronary artery disease 307. Endothelial dysfunction has also been shown to occur in OSAS patients with little evidence of cardiovascular disease in human studies that assessed intima-media thickness and carotid–femoral pulse-wave velocity 308. A role for this dysfunction in the pathogenesis of cardiovascular complications in OSAS has been supported by various studies demonstrating impairment in endothelium-dependent vasodilatation 309313. Furthermore, treatment with nasal CPAP has been reported to reverse endothelial dysfunction 314. There appears to be a sex difference in endothelial function in that flow-mediated vasodilation is more impaired in females with OSAS than males 315.

A major vasodilator substance released by the endothelium is NO 316, and decreased production or activity of NO may be an early sign of atherosclerosis. Decreased levels of NO have been found in OSAS patients and levels increase with CPAP therapy 317322. The endothelium also produces vasoconstrictor substances, such as endothelin and angiotensin II, and levels have been reported as increased in OSAS and to fall with effective CPAP therapy 323. However, other reports did not find an increase of endothelin in OSAS 324. The Sleep Heart Health Study has also reported evidence of vascular dysfunction among older participants, particularly arterial diameter 325, identifying OSAS as an independent risk factor for impaired flow-mediated vasodilation. However, endothelial dysfunction is often seen in patients with hypertension, hyperlipidaemia, diabetes or smoking and these comorbidities may limit the importance of OSAS as an independent risk factor for endothelial dysfunction.

Blood coagulation abnormalities

Increased cardiovascular risk in OSAS patients may also be linked to abnormalities of coagulation and excessive platelet activation and this topic has been recently reviewed 326. Increased circulating levels of activated coagulation factors have been reported by Robinson et al. 245 in untreated OSAS patients, but CPAP treatment appeared not to modify them. Interestingly, two groups have recently reported an increased d-dimer level in untreated OSAS and its correlation with the severity of nocturnal hypoxaemia, suggesting that a hypercoagulable state is potentially involved in cardiovascular risk in OSAS patients 327, 328. Other investigators have found the following: 1) increased blood viscosity in untreated adult OSAS 329; 2) increased fibrinogen level in both adults 329 and children with SDB 330; and 3) evidence of platelet activation, which decreased after CPAP treatment 331, 332. However, some uncertainties still remain on the independent role of OSAS on increased blood coagulability, due to the common coexistence of other cardiovascular risk factors 331, the lack of correlation between markers such as fibrinogen and severity of SDB in children 330 and the incomplete normalisation of coagulation after CPAP treatment 245, 331.

Metabolic dysregulation

The detailed clinical and epidemiological evidence concerning metabolic dysregulation in OSAS has been discussed above. OSAS-related factors that may contribute to metabolic dysregulation include increased sympathetic activity, sleep fragmentation and intermittent hypoxia.

Glucose intolerance

The mechanisms of impaired glucose tolerance in OSAS particularly involve insulin resistance. A number of reports have found increased insulin resistance and impaired glucose tolerance in OSAS patients, independent of body weight 211, 333, 334, and a worsening of insulin resistance with increasing AHI 212. In nondiabetic OSAS patients, circulating advanced glycation end-products have been reported to correlate with the severity of intermittent hypoxia 335 and a cause–effect relationship between hypoxia and glucose intolerance has been shown by studies in healthy humans 223. Animal studies also support an important role for intermittent hypoxia in the development of insulin resistance, which appears to be dependent on the disruption of leptin pathways 336.

The leptin pathway

Leptin is an adipocyte-derived hormone that regulates body weight through the control of appetite and energy expenditure. Leptin may predispose to platelet aggregation and has been implicated as an independent cardiovascular risk factor 337. Leptin has been extensively studied in recent years due to its role in appetite regulation 338, but its functions are probably more complex than initially believed. Human obesity is associated with increased leptin levels and a state of leptin resistance, while lack of leptin causes obesity in animal models 339. Leptin likely exerts pleiotropic functions in OSAS, not only by its effects on metabolism and obesity, but also by affecting ventilatory control 339. Expression of the human leptin gene is regulated by hypoxia 340. Hypercapnic OSAS patients also show a higher degree of leptin resistance compared with nonhypercapnic subjects 341, and similar data have been reported in obese non-OSAS subjects 342.

Several studies have reported that OSAS is associated with hyperleptinaemia, although some were not adjusted for obesity and visceral fat distribution. One study reported that elevated leptin levels in OSAS were only found in obese subjects 343, whereas other reports found that sleep hypoxaemia was the principal determinant 344. A report from the Cleveland family study also demonstrated BMI to be an important confounding factor in the relationship between OSAS and leptin levels 345. Effective treatment with CPAP has been reported to be associated with a decrease in leptin levels 346, 347, although in one report the fall in leptin levels was only observed in nonobese patients 347. The study by Shimizu et al. 348 was particularly interesting as changes in plasma leptin levels were related to cardiac sympathetic function and, thus, sought to link different pathogenetic mechanisms of cardiovascular complications in OSAS. While the results were not conclusive, the findings indicated a significant fall in leptin levels with CPAP therapy. However, another report found that leptin levels, when adjusted for body fat distribution, were not related to indices of OSAS 349 and a further study indicated that leptin levels were more closely associated with indices of obesity and lipid dysfunction than with indices of OSAS 350. Thus, the possibility of an independent relationship of leptin and other adipocytokines (such as adiponectin and ghrelin) to OSAS requires further investigation.


General priorities

While substantial progress has been made, particularly in recent years, in the identification of possible mechanisms to explain the association between OSAS and cardiovascular disease, substantial deficits remain. Many of the studies to date, particularly those related to identifying the basic mechanisms involved in these associations, have suffered from small sample size, inadequate control populations and failure to adequately control for potential confounding variables such as body mass index (BMI). Thus, it is not surprising that conflicting findings have been reported in many of these studies. There have been few translational studies that have explored basic mechanisms of cardiovascular disease in OSAS and applied the findings to the clinical setting.

In general terms, there is a clear need for large-scale collaborative studies of carefully defined patient populations with OSAS that are adequately controlled for potential confounders. Such large-scale studies carry the prospect of evaluating potential interactions between different basic mechanisms operating in OSAS and cardiovascular disease, and including interactions with other related disorders such as diabetes and dyslipidaemia. Studies that explore the interaction of several factors that could lead to the development of cardiovascular disease in OSAS are singularly lacking. For example, insulin affects both NO release and the sympathetic nervous system, and sleep deprivation contributes to the development of insulin resistance. Possible strategies could be as follows: 1) inclusion of large samples with various degrees of confounding factors including obesity, central obesity and other elements of the metabolic syndrome that may not be solely related to OSA; or 2) studying highly selected sub-groups, such as lean subjects with severe OSA and limited risks factors otherwise.

Furthermore, prospective studies of carefully defined patient cohorts would likely provide important information on the evolution of cardiovascular disorders in OSAS patients, but the recognised efficacy of CPAP in treating the disorder makes long-term randomised studies difficult to design because of the potential dangers of withholding effective therapy from severely affected patients, particularly accident risk. Special attention should be paid to respiratory disorders during sleep in specific patient categories.


In children, the prevalence of obesity, metabolic syndrome and type-2 diabetes are rising 351, 352. As a consequence, the prevalence of OSAS can be expected to shift towards younger ages if the current trend towards sedentary lifestyle and excess caloric intake is not reversed with effective interventions aimed at the young population. Conversely, in past decades, OSAS in children was considered as a mostly mechanical problem of upper airways secondary to adenotonsillar hypertrophy, easily solved by surgery and useful to study the clinical effects of upper airway obstruction free from the many confounders typically associated with OSAS in adults.

The elderly

In elderly subjects, obstructive apnoeas during sleep are quite frequent but their clinical and prognostic impact appears lower than in young to middle-aged subjects 32, 353. This finding, especially puzzling given that common comorbidities in the elderly would be expected to further worsen disease outcome, has led some authors to hypothesise that nocturnal intermittent hypoxia might activate chronic protective mechanisms against cardiovascular damage similar to those described for classic ischaemia–reperfusion damage 354. Survival bias and lack of clinical information on sleep apnoea in subjects who die because of acute cardiovascular events are alternative possible explanations for the apparent decrease in the consequences of OSAS in the elderly.


Studies on OSAS and cardiovascular risk have predominantly assessed male, middle-aged, obese patients with severe OSAS, leaving many questions still open on the pathophysiology of OSAS in females 355, 356. In pre-menopausal females, hormonal influences may directly or indirectly (by influencing body fat distribution) protect the female sex from increased collapsibility of the upper airway and account for the low prevalence and severity of OSAS in females compared with BMI-matched males 357. Nevertheless, SDB often occurs in late pregnancy 358 and may contribute to the pathophysiology of pre-eclamptic states 359 by potentiating cardiovascular responses to respiratory events during sleep or endothelial dysfunction 360, 361. Markers of endothelial dysfunction appear more strongly associated with SDB severity in females compared with males 315, but the contribution of obstructive sleep apnoea to hypertension may be lower in females compared with males 64. A role of sex hormones is suggested by the increased risk to develop sleep apnoea in the post-menopausal period apparently prevented by use of hormonal replacement therapy 33, 362, 363. Little is known about possible sex-related differences in the prognosis of OSAS, but a preliminary report from the Wisconsin Sleep Study suggested a much higher mortality rate in females compared with males with OSAS of comparable severity 364.

The foregoing account indicates that the basic mechanisms involved in the pathogenesis of cardiovascular disease in OSAS are complex and some aspects, particularly those related to the molecular mechanisms of intermittent hypoxia, may be more suited to cell culture and/or animal models. In addition, current knowledge on the interactions between genetic background and environmental factors in the pathogenesis of OSAS is still limited. Future studies will help clarify the interactions between putative OSAS genes and disruption of cardiovascular regulation or body weight control in individual patients 365. A schematic outline of future research priorities is proposed in figure 1.

Fig. 1—

Future research priorities for obstructive sleep apnoea (OSA) syndrome.

Specific priorities

Inflammatory molecular pathways

While there is emerging evidence that inflammatory pathways play an important role in the evolution of atherogenesis and subsequent vascular disease, the precise mechanisms remain to be elucidated. In particular, the balance between different molecular mechanisms of inflammation such as TNF-α and CRP, remains unclear, as does the interaction of inflammatory and adaptive molecular pathways, such as those involving NF-κB and HIF-1, respectively. There is emerging evidence that intermittent hypoxia plays a central role in these mechanisms via the effects of intermittent re-oxygenation, which is an integral feature of intermittent hypoxia, and possibly involving the generation of ROS. Additional studies involving cell culture models and translational studies involving bench, animal and human models are required to adequately explore the mechanisms involved.

Sympathetic excitation

Evidence has accumulated over the past decade of an important role for sympathetic nervous system overactivity in the pathophysiology of vascular dysfunction in OSAS, which appears to be relieved by CPAP therapy. However, the underlying mechanisms remain unclear, particularly the role of intermittent hypoxia and sleep fragmentation and the effects on different types and sizes of blood vessels. Animal models may be most suited to this area of investigation. Furthermore, the importance of sympathetic nervous system over activity in hypertension in general is still under debate; current evidence suggests a role in the later rather than earlier stages of disease development. Further controlled studies, involving larger groups of subjects and allowing for potential confounding factors are necessary to more clearly define the role of sympathetic excitation in the pathophysiology of cardiovascular complications in OSAS.


While platelet dysfunction and/or hypercoagulability play an important role in the pathogenesis of vascular disease, there are limited studies on the potential role of hypercoagulability in the development of vascular disease in OSAS. However, published reports demonstrating a potential link between hypercoagulability and OSAS indicate this to be a potentially important area for further study.

Endothelial dysfunction

Endothelial dysfunction is an important precursor of atherosclerosis and subsequent vascular disease and there is emerging evidence of such dysfunction in OSAS. To date, however, studies have been limited and there is little information on potential interaction of vasoactive substances such as NO and other molecular mechanisms involving inflammation.

Metabolic dysregulation

While there is now good evidence that OSAS is associated with insulin resistance, independent of potential confounders, the detailed mechanisms of this relationship have not been identified. There is some evidence that intermittent hypoxaemia may be more important than other indices of OSAS. Furthermore, while there is evidence from several reports of abnormal lipid metabolism in OSAS, particularly relating to leptin, the detailed mechanisms remain to be identified.


Obstructive sleep apnoea syndrome and cardiovascular disease are highly prevalent in the European population and there is now convincing evidence that obstructive sleep apnoea syndrome is an independent risk factor for hypertension, ischaemic heart disease, and probably stroke. However, there is a need to collect more epidemiological data in Europe taking into account both sex- and age-related issues. The ongoing Action for the Cooperation in the field of Scientific and Technical Research (COST) is a step in this direction 39. The basic mechanisms involved in the increased cardiovascular risk of obstructive sleep apnoea syndrome remain unclear. In particular, there is a substantial deficit in translational studies that have evaluated specific aspects of this relationship in cell culture or animal models and applied the findings to human populations with obstructive sleep apnoea syndrome. Research in these relationships carries the prospect of identifying novel therapies that should improve the overall health of the population.


The Management Committee of COST Action B26 - Obstructive Sleep Apnoea is as follows: W.T. McNicholas (Chair), M.R. Bonsignore (Past-Chair), P. Levy (Vice-Chair), A. Alonderis, F. Barbe Illa, P. Calverley, W. De Backer, K. Diefenbach, V. Donic, I. Fietze, K. Franklin, L. Grote, J. Hedner, P.J. Jennum, L. Lavie, P. Lavie, J. Montserrat, G. Parati, T. Penzel, O. Polo, R. Riha, D. Rodenstein, R. Schulz, E. Sforza, P. Sliwinski, Z. Tomori, G. Varoneckas and J. Zielinski.

  • Received February 23, 2006.
  • Accepted August 28, 2006.


View Abstract