Abstract
Sleep profoundly affects metabolic pathways. In healthy subjects, experimental sleep restriction caused insulin resistance (IR) and increased evening cortisol and sympathetic activation. Increased obesity in subjects reporting short sleep duration leads to speculation that, during recent decades, decreased sleeping time in the general population may have contributed to the increasing prevalence of obesity. Causal inference is difficult due to lack of control for confounders and inconsistent evidence of temporal sequence.
In the general population, obstructive sleep apnoea (OSA) is associated with glucose intolerance. OSA severity is also associated with the degree of IR. However, OSA at baseline does not seem to significantly predict the development of diabetes. Prevalence of the metabolic syndrome is higher in patients with OSA than in obese subjects without OSA. Treatment with continuous positive airway pressure seems to improve glucose metabolism both in diabetic and nondiabetic OSA but mainly in nonobese subjects.
The relative role of obesity and OSA in the pathogenesis of metabolic alterations is still unclear and is intensively studied in clinical and experimental models. In the intermittent hypoxia model in rodents, strong interactions are likely to occur between haemodynamic alterations, systemic inflammation and metabolic changes, modulated by genetic background. Molecular and cellular mechanisms are currently being investigated.
SERIES “THE GENETIC AND CARDIOVASCULAR ASPECTS OF OBSTRUCTIVE SLEEP APNOEA/HYPOPNOEA SYNDROME”
Edited by R.L. Riha and W.T. McNicholas
Number 6 in this Series
There is compelling evidence that sleep apnoea represents a major cardiovascular risk 1–8. Many studies have reported an independent association of obstructive sleep apnoea (OSA) with several components of the metabolic syndrome (MetS), particularly insulin resistance (IR) and abnormal lipid metabolism 9, 10. This association may further increase cardiovascular risk 11, since the MetS is recognised to be a risk factor for cardiovascular morbidity and mortality 12, 13.
Rapidly accumulating data from both epidemiological and clinical studies 14, 15 suggest that OSA is independently associated with alterations in glucose metabolism and places patients at an increased risk of the development of type 2 diabetes. Recent reports have indicated that many patients with type 2 diabetes have OSA 15. Even though there is emerging evidence that the relationship between type 2 diabetes and OSA is at least partially independent of adiposity 16, 17, there are several important limitations in the published literature that do not allow causality to be established, i.e. cross-sectional studies, use of snoring as a surrogate marker of OSA, and various techniques for the assessments of glucose metabolism and type 2 diabetes. Recent state-of-the art reviews have highlighted these limitations and emphasised the need for further clinical research in this direction 14, 15.
In this article, we will review the physiological effects of sleep on glucose metabolism and the possible role of sleep disruption on the pathogenesis of metabolic abnormalities, the clinical evidence linking sleep disordered breathing (SDB) and impairment of glucose metabolism, the current evidence regarding the impact of continuous positive airway pressure (CPAP) treatment on glucose and insulin control, and the major role of adipose tissue and visceral obesity. We will further discuss the possible mechanisms by which OSA may contribute to metabolic dysregulation in light of the published evidence in humans and animal models, i.e. increased sympathetic activity, sleep fragmentation and intermittent hypoxia. This article will also refer to a European Respiratory Society Research Seminar held in Dusseldorf (Germany) from November 30–December 1, 2007 in conjunction with two EU COST (Cooperation in the field of Scientific and Technical Research) Actions on “Cardiovascular risk in OSAS” (B26) and “Adipose tissue and the metabolic syndrome” (BM0602).
PHYSIOLOGICAL AND CLINICAL DATA
Sleep and metabolism
OSA may affect metabolism indirectly, by decreasing the amount and/or quality of sleep. Sleep loss profoundly affects metabolic pathways 18. In healthy subjects, experimental sleep restriction caused IR, together with increased evening cortisol and sympathetic activation 19. Sleep restriction was also shown to be associated with reduced leptin and increased ghrelin plasma concentrations and increased appetite 20. Modest acute sleep loss, such as selective slow-wave sleep deprivation, may alter glucose tolerance in normal subjects 21. In general population cohorts, short sleep duration was associated with altered plasma levels of leptin 22, 23 and ghrelin 22 and increased body mass index (BMI) 22, 23. In young adults, a prospective study found a significant risk of obesity in subjects reporting short sleep duration 24, leading to speculation that decreased sleeping time over the recent decades may have contributed to the increasing prevalence of obesity in the general population. In addition, in general population cohorts, difficulties falling asleep, difficulties in sleep maintenance and reduction in sleep duration have been found to be associated with an increased incidence of diabetes in males 25, 26.
The causal relationship between sleep duration and obesity is far from being proven 27–30, as shown by recent critical reviews or meta-analyses of published data in this field 27, 31, 32. Although cross-sectional studies from around the world show a consistent increased risk of obesity among short sleepers in children and adults 32, prospective data seem to fail to show this 33. Causal inference is difficult due to lack of control for important confounders and inconsistent evidence of temporal sequence in prospective studies 27, 32. Moreover, effect size and importance of sleep duration in comparison to other risk factors for obesity have been recently challenged 27, 34, 35. However, causality is often difficult to establish in epidemiology owing to biological complexity and multiple interactions 36. Moreover, a modest effect size, such as the average decrease in BMI by 0.35 units associated with one extra hour of sleep in the general population 32, may be unimportant on an individual basis but of major significance in public health 36. From the available relative risk ratios and short sleep prevalence, Young 36 calculated that 5–13% of the total proportion of obesity in children and 3–5% in adults could be attributable to short sleep.
The mechanisms that are possibly involved are of interest. Sleep deprivation has been found to induce a pro-inflammatory state, with increased release of interleukin (IL)-6 37, 38 and production of IL-6 and tumour necrosis factor (TNF)-α by circulating monocytes 39. Nuclear factor (NF)-κB activation has been identified as a molecular pathway by which sleep restriction may influence leukocyte inflammatory gene expression and the risk of inflammation-related disease 40. The pro-inflammatory effects of sleep restriction may, at least partly, be mediated by stress activation, i.e. sympathetic and/or cortisol activation 41–43. In addition, the group of Knutson and Van Cauter 44 speculated that the adverse impact of sleep deprivation on appetite regulation is likely to be driven by increased activity in neuronal populations expressing the excitatory peptides orexins, which promote both waking and feeding 44–46.
In summary, sleep loss could affect metabolism via several mechanisms, but it is difficult to apply the currently available data to OSA. There are no studies specifically addressing the effects of sleep fragmentation (such as in OSA) on metabolism, as recently stated 47.
Association between sleep apnoea, glucose intolerance and diabetes
Early studies indicated a possible causal association between the presence of OSA and development of type 2 diabetes. However, most studies exhibited significant limitations including small sample size, highly selected populations, inadequate adjustment for confounders and use of surrogate markers of OSA 15. Methods have also been highly variable among studies. Table 1⇓ summarises current definitions used in clinical studies on glucose metabolism and the MetS 48–50.
Current definitions used in clinical studies
Table 2⇓ summarises the available epidemiological studies on the association of SDB with IR and diabetes 16, 51–62. In general population studies 25, 63, snoring was shown to be a risk factor for the development of diabetes over 10 yrs independent of confounding factors. Importantly, two population cross-sectional studies including only lean subjects (BMI <25 kg·m−2) found an independent association between frequent snoring and reduced glucose tolerance 56, 62. Several other sleep anomalies have also been related to type 2 diabetes 25, 26, 63. The relationship between self-reported sleep complaints and risk of diabetes may be less pronounced in females 63 than in males 25. The Sleep Heart Health Study 16 (performed in 2,656 individuals) showed that sleep-related hypoxaemia was associated with glucose intolerance independently of age, sex, BMI and waist circumference. OSA severity was also associated with the degree of IR after adjustment for obesity. More recently, these data have been confirmed in the same cohort, and the association between SDB and impaired glucose metabolism was found to be similar in normal-weight and overweight subjects 61. The Wisconsin Sleep Study (n = 1,387) demonstrated a significant cross-sectional association between OSA and type 2 diabetes for all degrees of OSA, which persisted for moderate-to-severe OSA after adjustment for obesity (odds ratio 2.3) 56. However, although the longitudinal data showed that OSA at baseline predicted the development of diabetes over 4 yrs, significance disappeared after adjusting for obesity 56. Finally, OSA was recently found to be independently associated with decreased insulin sensitivity in a population-based sample of females investigated with full-polysomnogram (PSG) and insulin sensitivity index (ISI) calculated from the results of an oral glucose tolerance test 62.
Epidemiological studies on insulin resistance(IR) and metabolic syndrome (MetS) in sleep disordered breathing (SDB)
Similar data have been obtained in samples of OSA patients (tables 3⇓ and 4⇓) 64–80, with a large prevalence of positive 63, 64–79 rather than negative 78–80 studies. In clinical populations, OSA patients characterised by full-PSG were significantly more likely to have impaired glucose tolerance (IGT) and diabetes than subjects free of OSA syndrome (OSAS) 71. The relationship between SDB and impaired glucose–insulin metabolism was independent of obesity and age 71. A number of reports found increased IR and IGT in OSAS patients, independent of body weight 68–70, 75, and a worsening of IR with increasing apnoea/hypopnoea index (AHI) 54. However, other studies failed to demonstrate an independent effect of AHI owing to the major impact of obesity 64, 65, 80. Excessive daytime sleepiness (EDS) may also be of importance, as underlined by the recent findings that hyperglycaemia and IR only occurred in OSA patients presenting with EDS 77.
Clinical studies on insulin resistance(IR) in obstructive sleep apnoea (OSA): positive studies
Clinical studies on insulin resistance(IR) in obstructive sleep apnoea (OSA): negative studies
It may be concluded, in agreement with Tasali and Ip 9, that despite the abundance of cross-sectional evidence for the link between OSA and abnormal glucose control, further well-designed longitudinal and interventional studies are clearly needed to address the direction of causality.
OSA and the metabolic syndrome
According to clinical and epidemiological studies, the cluster of risk factors known as the MetS is associated with increased risk for diabetes, cardiovascular events and mortality in the general population 12. Although the definition of the MetS is still under debate 81–83, IR is considered as the major metabolic abnormality, and is usually associated with an increased amount of visceral (dysfunctional) fat 84. The World Health Organization definition of the MetS is based on the direct measurement of IR (table 1⇑) 50. Another definition (National Cholesterol Education Program–Adult Treatment Panel (ATP) III) is based on simple clinical findings (abdominal obesity, dyslipidaemia, hypertension and increased plasma glucose), and is easily applicable as it does not require tests to be performed in a specialised environment (table 1⇑) 49. Finally, the definition proposed by the International Diabetes Federation shares many features with the ATP III definition, but defines the cut-offs for waist circumference according to ethnicity 85, thus accounting for differences in body habitus between Caucasian and Asian populations. All these definitions should be considered as “in progress” and subject to change based on the evidence provided by current and future studies.
The MetS can be explained by viewing abdominal adipose tissue as an endocrine organ (see below), releasing into the circulation excess harmful free fatty acids (FFA), angiotensin II and adipokines. Increased blood FFA inhibits the uptake of glucose by muscle. Because excess FFA and angiotensin II damage the pancreas, insulin release is not sufficient to counteract hyperglyacemia, resulting in IR 86. The most prevalent form of this group of metabolic abnormalities linked to IR is found in patients with abdominal obesity, especially with an excess of intra-abdominal or visceral adipose tissue 87. It has been suggested that visceral obesity may represent a clinical intermediate phenotype, reflecting the relative inability of subcutaneous adipose tissue to clear and store extra energy resulting from dietary triglycerides, thus leading to fat deposition in visceral adipose depots, skeletal muscle, liver, heart, etc. Thus, visceral obesity may be both a marker of a dysmetabolic state and a cause of the MetS 87.
MetS is often found in OSAS patients (table 5⇓) 10, 58, 65, 74, 75, 88–92. However, the relative role played by OSA and obesity in the pathogenesis of MetS remains uncertain. Prevalence of the MetS is higher in patients with OSAS than in the European general population (15–20%) or in obese subjects without OSAS 81. In subjects with SDB, prevalence rates ranged from 19% in Korean snorers 57 to 87% in OSA patients from the UK 10. The risk of developing the MetS increased with severity of SDB in Western as well as Eastern populations 17, 58, 69, 73, 88–91, 93. The studies published to date agree on the estimate of a five-fold (or higher) risk of MetS in OSAS patients compared with controls.
Metabolic syndrome(MetS) in obstructive sleep apnoea syndrome (OSAS) patients
Most studies found a significant association between MetS and AHI, while the association with intermittent hypoxaemia was weak or absent. This result is at variance with the data obtained in animal models, which suggest a role of intermittent hypoxaemia in metabolic alterations (see below). In OSAS patients, apnoea or desaturation indexes showed stronger correlation with the amount of visceral fat than with global obesity indexes, such as BMI 69, leading some authors to propose that OSAS should be considered as a component of the MetS 94. However, the pathogenic mechanisms possibly leading from metabolic alterations to OSAS are still unclear. This is also illustrated by the fact that not all OSAS patients are obese, and not all obese subjects develop OSAS. Further studies with careful assessment of the amount and distribution of body fat are needed to better understand the pathophysiology of adipose tissue and its interaction with OSAS, taking into account the current progress in basic and clinical research on obesity.
Effects of OSA treatment
The effects of CPAP treatment on glucose metabolism have been evaluated in both nondiabetic and diabetic patients, as summarised in table 6⇓ 78, 93, 95–109, and may provide some clues as to the relative role of OSA and obesity in the pathogenesis of metabolic alterations. Until 2003, there were very few clear results owing to methodological issues and various confounders 92. In 2004, Harsch et al. 99 reported that CPAP treatment for 2 days rapidly improved the ISI in nondiabetic patients and that the positive effects of CPAP persisted after 3 months of treatment. Conversely, ISI improved only slightly and after prolonged treatment in obese patients (BMI >30 kg·m−2), suggesting that in the latter group insulin sensitivity is primarily determined by obesity and, to a lesser extent, by sleep apnoea 99. In nondiabetic patients, increased blood glucose was found after 1 night of CPAP treatment, with a tendency to higher fasting insulin and resistance to insulin (i.e. homeostasis model assessment (HOMA)-IR) after CPAP 101. Such an increase in blood glucose might be related to CPAP-associated increase in growth hormone 95, 97, leading to an increase in plasmatic FFA owing to growth hormone lipolytic effects and thus to reduced glucose utilisation by skeletal muscles.
Clinical studies on glycaemic control and insulin resistance(IR) in obstructive sleep apnoea syndrome (OSAS) patients before and after continuous positive airway pressure (CPAP) treatment
CPAP treatment does not greatly affect the metabolic status of obese OSA patients. A randomised, placebo-controlled, blinded crossover trial comparing cardiovascular and metabolic outcomes after 6 weeks of therapeutic or sham CPAP reported no change in glucose, lipids, IR or the proportion of patients with MetS in obese males, while positive effects of treatment on blood pressure and EDS were clearly present in the therapeutic CPAP group 105. Whether EDS is also a critical determinant of the response to CPAP treatment, as recently reported 77, needs further evaluation in randomised control trials of large samples.
The limited duration of randomised controlled studies in OSA patients could, at least partly, account for the nonsignificant effects of CPAP treatment on glucose metabolism found in most studies. An observational study in a highly selected sample of OSAS patients found improved insulin sensitivity in patients with good compliance to CPAP after 2.9 yrs of treatment 108. Similarly, the Swedish Obesity Study reported a three-fold incidence of diabetes and hypertriglyceridaemia in patients with witnessed apnoeas compared to subjects with no OSA at 2 yrs follow-up 110. Two studies reported that visceral fat decreased after CPAP treatment 103, 111 while a more recent study 93 found no change in IR or visceral fat after CPAP for 3 months. Therefore, large longitudinal studies focusing on different aspects of this complex topic are needed to assess the potential long-term effect of OSA treatment.
In type 2 diabetic patients with OSAS, several studies have assessed the impact of CPAP treatment on glycaemic control. Recent observational studies using continuous glucose monitoring techniques have reported positive effects of CPAP on glycaemic control, already present during the first night of treatment, as variability of glycaemic values decreased compared with baseline conditions 107. Dawson 109 found decreased glucose levels and variability without significant changes in haemoglobin (Hb)A1c levels. Conversely, Babu et al. 102 reported the results of 72-h continuous monitoring of interstitial glucose and measurements of HbA1c levels in 25 patients before and after 3 months of CPAP. Post-prandial glucose values were significantly reduced 1 h after treatment, and HbA1c level decreased in patients with abnormally high baseline HbA1c (>7%). Furthermore, in subjects who used CPAP for >4 h·day−1, the reduction in HbA1c level was significantly correlated with CPAP use 102. A retrospective study also confirmed a slight reduction in HbA1c in diabetic patients with OSA treated with CPAP 100. However, obesity was likely to be a major confounding factor, since a randomised controlled trial comparing therapeutic (n = 20) or placebo CPAP (n = 22) for 3 months found no difference in terms of glycaemic control or IR in these patients 105. In summary, a huge impact of obesity is also present in type 2 diabetic patients, which may offset the impact of CPAP 112, 113.
The effects of CPAP treatment on the MetS are controversial, as recently summarised in two reviews 9, 114. A recent observational study in patients with severe OSAS reported decreased blood pressure and plasma cholesterol and improved HOMA index after 8 weeks of CPAP treatment in patients with good compliance to therapy; the estimated effect of CPAP over 10 yrs was a decrease in cardiovascular risk from 18.8% to 13.9% 106. However, a randomised controlled study in patients with moderate-to-severe OSAS showed that IR or other MetS variables were unaffected after 6 weeks of effective CPAP 104. Moreover, some studies suggested positive effects on plasma lipids after CPAP in patients showing a good compliance to treatment 106, 115, 116, while other studies found no effect of OSA treatment on plasma cholesterol or triglycerides 93, 104.
Other markers of glucose metabolism have been assessed in OSA patients, such as insulin growth factor (IGF)-1 and adiponectin. A high IGF-1 concentration is predictive of decreased risk of type 2 diabetes and impaired glucose tolerance 117, 118, whereas low IGF-1 concentrations were found to be associated with increased risk of cardiovascular disease 119. The complex interactions between IGF-1, its binding proteins and insulin sensitivity promote IGF-1 as an important regulator of glucose homeostasis. While fasting insulin and blood glucose are subject to short-term changes, IGF-1 is a more stable variable subject to long-term regulation. In a population-based cohort, IGF-1 significantly increased after CPAP treatment 117. However, in OSA patients’ improvement in IGF-1 after CPAP was reported to occur only in patients presenting with EDS 76. Adiponectin is known to counteract the effects of IR 120, and the effects of OSA treatment on adiponectin have been assessed with controversial results. Some studies found increased adiponectin after 1 night 121 or 2 weeks 122 of CPAP treatment, while other studies found no change in adiponectin levels after OSA treatment for 1 night 123 or 1–3 months 93, 124, 125.
It is possible that OSA treatment may positively affect only some MetS components, rather than affecting all of them 110, 115. The available results need further confirmation. There is a strong need for controlled prospective studies, to evaluate whether some patient subgroups might especially benefit from the effects of CPAP treatment on metabolic variables 14.
OSA and glucose metabolism in paediatrics
Children and adolescents represent a very important clinical population since epidemic obesity in paediatrics is a major health concern 125, 126 and is associated with high MetS prevalence 127, although the real relevance of the MetS in adolescents is currently under discussion 128. In addition, children classically represent a good clinical model for examining the relationship between SDB and glucose metabolism with limited coexistent comorbidity, even though differences between adult and paediatric OSA may have become smaller due to the current high prevalence of obesity at a young age [129]. The causal role of SDB in paediatric metabolic abnormalities is currently unclear, as indicated in a recent review 130.
In the general population, SDB in children seems to be associated with MetS 131. In the Cleveland Cohort 131, after adjusting for age, race, sex and preterm status, children with SDB had a 6.49 increased odds of MetS compared with children without SDB 131. Approximately 25% of the sample was overweight and 19% had MetS. In clinical samples of obese children, SDB was found to correlate with fasting insulin levels independent of BMI 132, 133. This has been challenged among children with suspected SDB, in whom IR and dyslipidaemia seem to be determined primarily by the degree of body adiposity rather than by the severity of SDB 134, 135. In nonobese children, severity of SDB was not a significant predictor of fasting insulin or HOMA index values 136.
As for adipokine levels in children with SDB, obesity appeared as the primary determinant although SDB and associated hypoxaemia may contribute to elevated leptin levels 137. In a recent study conducted in obese and nonobese children, Gozal et al. 138 showed that SDB was associated with altered lipid homeostasis and systemic inflammation. In the presence of obesity, SDB also affected glucose metabolism through reduction in insulin sensitivity, independent of obesity 138. Therefore, in obese children there could be an interaction between increased adiposity and SDB to promote and amplify IR.
Few data have been obtained in children on the effects of treatment for OSA on metabolic abnormalities. A small study reported a slight improvement in plasma high-density lipoprotein cholesterol after adenotonsillectomy, but no major changes in insulin level 135. Leptin and sympathetic markers were found to be increased at baseline in children with SDB compared to simple snorers, and decreased after CPAP treatment for 3 months 139. However, IR was unaffected by treatment 139. Also, no change was shown in insulin level or HOMA index compared to baseline measurements in a sample of Greek children after adenotonsillectomy 140.
In summary, the field of SDB and its metabolic consequences together with its interaction with obesity is rapidly developing, but uncertainties remain significant as highlighted by recent reviews and editorials 141–143.
MECHANISMS AND EXPERIMENTAL DATA
Role of adipose tissue and visceral obesity
White adipose tissue is considered to be a key endocrine and secretory organ that releases a large number of adipokines with a major link to inflammation and immunity. The paradigm shift in adipose tissue biology was initiated in 1994 by the discovery of leptin 144. Subsequently, a growing number of proteins, peptides and other factors released from white adipocytes, collectively termed adipocytokines, have been described 145. Most of these adipocytokines are linked to inflammation and their production is increased in obesity. To date, only adiponectin is known to exert anti-inflammatory and anti-diabetic activity, and is reduced in obesity and type2 diabetes 146, 147.
Human obesity is characterised by increased rather than low leptin production. In OSAS patients, several studies reported increased leptin levels compared to weight-matched controls 69, 148–152, which correlated with OSA severity 148, 150, 151 and decreased after CPAP 111, 148, 153, 154. Similar results were recently reported in paediatric OSA 139. Therefore, OSA may exert an independent effect on leptin levels, causing leptin resistance, possibly through hypoxia, which acts by increasing leptin gene transcription 155.
Adipose tissue inflammation is thought to play a key role in the development of MetS, type 2 diabetes and cardiovascular disease 156. In 2004, Trayhurn and Wood 157 suggested that adipose tissue inflammation may represent a specific response to relative hypoxia in clusters of adipocytes that become distant from the vasculature as cell size increases. It has since been demonstrated that hypoxia occurs in adipose tissue of obese mouse models and triggers expression of inflammatory adipokines 158.
Hypoxia-induced factor (HIF)-1 plays a key role in the response to hypoxia in most tissues. Transcription factors such as NF-κB and CREB are downstream targets of HIF-1. The number of hypoxia-sensitive genes is continuously growing, and to date >70 genes have been described as targets of HIF-1. These genes include proteins involved in angiogenesis, cell proliferation, apoptosis and energy metabolism 159. Hypoxia may increase expression and secretion of a variety of inflammation-related adipocytokines such as IL-6, macrophage migration inhibitory factor and vascular endothelial growth factor. Therefore, hypoxia is likely to affect adipocyte function and promote adipose tissue inflammation. This may play a critical role in obesity-related disorders and may trigger the development of peripheral resistance to insulin and thus promote the development of type 2 diabetes and the MetS. The relationship between HIF-1 and inflammation has been discussed in detail in another article of this series 160.
Analysis of secretory products from primary human adipocytes revealed that these cells release classical adipocytokines such as TNF-α, IL-6, leptin, and adiponectin, as well as newly discovered adipocytokines i.e. tissue inhibitor of metalloproteinases-1 and monocyte chemotactic protein (MCP)-1 161. MCP-1 was first described as a secretory product of monocytes and endothelial cells with a prominent role in arteriosclerosis but it is also associated with the obese state. MCP-1 exhibits IR-inducing capability in adipocytes and myocytes 162.
Increased expression and secretion of adipokines in obesity may be a marker of low-grade chronic inflammation in adipose tissue. Protein kinase C and IκB kinase (IKK) are two kinases known to be involved in the inflammatory processes underlying IR. IKK influences insulin sensitivity, especially in skeletal muscle, by inhibiting insulin signalling by insulin receptor substrate-1 phopshorylation on serine residues and by activating NF-κB. In turn, NF-κB regulates production of pro-inflammatory cytokines such as TNF-α and IL-6 163, and generates both hepatic and systemic inflammation as well as IR 164.
In summary, the adipose tissue in obesity shows abnormal function and evidence of hypoxia and inflammation. This might be worsened by the occurrence of apnoeas during sleep, with further hypoxia and inflammation. The relative role of obesity and OSA in the pathogenesis of metabolic alterations is still unclear and is being intensively studied both in clinical and experimental models.
OSA, oxidative stress, inflammation and adipose tissue
OSA cardiovascular and metabolic consequences are now viewed as a component of a systemic disease resulting from oxidative stress 165, and systemic and vascular inflammation 166–172. Inflammation appears to be mostly confined to the vascular compartment, while systemic inflammation is often absent or mild. This may account for the variable level of C-reactive protein, which is often found not to be elevated in OSA patients without comorbidities 173, 174.
Obesity associated with OSA appears to be the strongest determinant of systemic inflammation 175. Adipose tissue inflammation may play a critical role in OSA-associated morbidity 176, with peri-vascular adipose tissue especially contributing to the release of cytokines, TNF-α, pro-atherogenic chemokines, and pro-angiogenic peptides 177, 178 (fig. 1⇓). Whether these factors contribute directly to alterations in the function and structure of the vascular wall and the development of atherosclerosis and cardiovascular complications in OSA remains to be studied.
Effects of hypoxia on adipokines and their interactions with insulin metabolism and endothelial function. The main factors involved are leptin, angiotensinogen (Ang), resistin, C-reactive protein (CRP), tumour necrosis factor (TNF)-α and plasminogen activator inhibitor (PAI)-1. Leptin promotes (red arrows) insulin resistance and endothelial dysfunction, whereas adiponectin is protective (blue arrows). Obesity, a state of leptin resistance and endothelial dysfunction, also exhibits hypoxia, which is known to activate (red arrow) promoting adipokines and inhibit (blue arrows) adiponectin production. In OSA, obesity and night-time hypoxia might act synergistically in producing inflammation at the systemic and vascular level, and in promoting metabolic and cardiovascular dysfunction. HDL: high-density lipoprotein; OxLDL: oxidised low-density lipoprotein; CD40L: CD40 ligand; VCAM-1: vascular cell adhesion molecule-1; ICAM-1: intercellular adhesion molecule-1; +: activation/promotion; −: inhibition/protection. Modified from 178.
Intermittent hypoxia
Intermittent hypoxia (IH) is considered the peculiar pathophysiological aspect of OSA and has been extensively studied in lean and obese rodent models. Recent reviews have summarised experimental and clinical data linking IH to cardiovascular and metabolic alterations 178. Metabolic and atherosclerotic changes have been shown in mice exposed to chronic IH 179–183.
In the chronic IH model (35 days) in mice, both systemic and localised inflammation of small and large arteries occurred, with evidence of peri-adventitial localisation of T-cells infiltration highly suggestive of a critical role of the peri-adventitial fat in the IH-related vascular inflammation (C. Arnaud, University of Grenoble, Grenoble, France; personal communication). Indeed, this does not rule out haemodynamic factors as, in another study in mice 180, platelet endothelial cell adhesion molecule-1, a marker of the endothelial cell, was decreased at both the heart and aorta level with a specific gradient, without loss of endothelial cells, possibly indicating a role for shear forces applied to the heart and aorta. Thus, vascular remodelling may result from either haemodynamic or inflammatory changes, or both. From these studies and others already published 166, 179, 183–185, strong interactions are likely to occur in response to chronic IH between haemodynamic alterations, systemic inflammation and metabolic changes, and modulated by genetic background 178, 182. Inflammation may largely contribute to glucose homeostasis dysregulation.
Indeed, from a metabolic perspective, several pieces of evidence support a role for IH in the metabolic alterations seen in OSA. Exposure of lean mice (C57BL/6J) to IH for 5 days increased serum cholesterol and phospholipids levels, up-regulated triglycerides and phospholipid biosynthesis, and inhibited cholesterol uptake in the liver 179. These effects may be mediated through HIF-1 activation for triglycerides and the post-transcriptional regulation of lipid biosynthesis (sterol regulatory element binding protein-1) but not for serum cholesterol levels 186.
IH may result in acute IR in otherwise lean, healthy animals, and the response is associated with decreased glucose utilisation of oxidative muscle fibres, independent of autonomic nervous system activation 185. The magnitude of metabolic alterations may also depend on the severity of IH 184. However, in contrast to the persistent effects of chronic IH on sympathetic activity and blood pressure, the effects of IH on glucose homeostasis appear to be limited to the periods of hypoxic exposure 187. Moreover, combining IH exposure and glucose infusion amplified the alteration of blood glucose diurnal rhythm and led to high rates of apoptosis in β-cells 187. The overall effects of IH on glucose homeostasis are summarised in table 7⇓ 179, 185, 188–192, and appear highly complex, in part, because IH causes loss of weight in lean animals, which counteracts IR. However, these data support the findings in humans suggesting a synergistic effect of increased adiposity and SDB in promoting metabolic dysfunction. Finally, it should be reminded that sleep fragmentation and intermittent hypoxia may also interact in modulating glucose homeostasis in animal models as well as in OSA, but the effects of sleep fragmentation are extremely difficult to study, even in animal models 182.
Summary data on metabolic variables in studies on the effects of intermittent hypoxia(IH) in lean and obese mice
A recent area of investigation is the potential role of IH as a “second hit” stimulus for the transition from hepatic steatosis to nonalcoholic steatohepatitis (NASH) 190–192. Recent studies in mice exposed to chronic IH would support this possibility, since animals fed a regular diet developed mild liver injury, while animals fed a high-fat diet showed evidence of inflammation and fibrosis of the liver 191. In both groups, there was evidence of hepatic oxidative stress. NASH is likely to be associated with hepatic IR and this may further contribute to glucose homeostasis dysregulation. This topic is still largely unexplored in the clinical context, since few studies to date have examined hepatic function in OSA patients. The available data suggest that at least some patients, both adults and children, may show evidence of hepatic dysfunction correlated with the severity of nocturnal IH 193, 194.
CONCLUSIONS
Alterations in sleep quantity or quality may affect glucose metabolism. However, although cross-sectional studies from around the world show a consistent increased risk of obesity among short sleepers in children and adults, large prospective studies are needed. In addition, in SDB, despite the abundance of cross-sectional evidence for the link between OSA and abnormal glucose control, further well-designed longitudinal and interventional studies are clearly needed to address the direction of causality. The available evidence also suggests that CPAP has little or no effect on the metabolic status of obese subjects, presumably owing to the major impact of visceral obesity. However, recent data obtained in diabetic OSA patients by using the technique of continuous monitoring suggest that CPAP treatment may improve glycaemic control. Thus, the synergistic negative effects of obesity and SDB represent a major research challenge, as shown by the complex picture emerging from studies in animal models. The interaction between hypoxia and metabolism possibly involves stress activation, oxygen radical production, and multiple cellular pathways (NF-κB, HIF and apoptosis) and cell types (inflammatory cells, vascular endothelium, adipocytes). There is a potential role for adipose tissue inflammation both regarding vascular remodelling and metabolic dysfunction. Clinical and translational research is urgently needed in this field.
Statement of interest
None declared.
Footnotes
-
Previous articles in this series: No. 1: MacLeod AK, Liewald DCM, McGilchrist MM, et al. Some principles and practices of genetic biobanking studies. Eur Respir J 2009; 33: 419–425. No. 2: Riha RL, Gislasson T, Diefenbach K. The phenotype and genotype of adult obstructive sleep apnoea/hypopnoea syndrome. Eur Respir J 2009; 33: 646–655. No. 3: Jennum P, Riha RL. Epidemiology of sleep apnoea/hypopnoea syndrome and sleep-disordered breathing. Eur Respir J 2009; 33: 907–914. No. 4: Garvey JF, Taylor CT, McNicholas WT. Cardiovascular disease in obstructive sleep apnoea syndrome: the role of intermittent hypoxia and inflammation. Eur Respir J 2009; 33: 1195–1205. No. 5: Lavie L, Lavie P. Molecular mechanisms of cardiovascular disease in OSAHS: the oxidative stress link. Eur Respir J 2009; 33: 1467–1484.
-
Earn CME accreditation by answering questions about this article. You will find these at the back of the printed copy of this issue or online at www.erj.ersjournals.com/current.dtl
- Received November 5, 2008.
- Accepted February 25, 2009.
- © ERS Journals Ltd
References
- 1.↵
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.
- 29.
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.
- 43.↵
- 44.↵
- 45.
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 123.
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.↵
- 137.↵
- 138.↵
- 139.↵
- 140.↵
- 141.↵
- 142.
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵
- 148.↵
- 149.
- 150.↵
- 151.↵
- 152.↵
- 153.↵
- 154.↵
- 155.↵
- 156.↵
- 157.↵
- 158.↵
- 159.↵
- 160.↵
- 161.↵
- 162.↵
- 163.↵
- 164.↵
- 165.↵
- 166.↵
- 167.
- 168.
- 169.
- 170.
- 171.
- 172.↵
- 173.↵
- 174.↵
- 175.↵
- 176.↵
- 177.↵
- 178.↵
- 179.↵
- 180.↵
- 181.
- 182.↵
- 183.↵
- 184.↵
- 185.↵
- 186.↵
- 187.↵
- 188.↵
- 189.
- 190.↵
- 191.↵
- 192.↵
- 193.↵
- 194.↵