Background

Obstructive sleep apnea (OSA) is a prevalent disease with roughly 1 of every 5 adults having at least mild OSA (defined by apnea–hypopnea index (AHI) ⩾5) and 1 of every 15 having at least moderate OSA (AHI ⩾15).1 The high prevalence rates are disturbing considering that OSA patients have an increased risk of morbidity and mortality, particularly due to cardiovascular disease.2, 3, 4, 5, 6, 7 OSA patients are often excessively sleepy during the daytime causing traffic crashes and work injuries, and they commonly have an increased number and severity of cardiovascular risk factors, such as hypertension, obesity, atherosclerosis, insulin resistance, the presence of increased inflammatory markers and endothelial dysfunction.8 Although OSA patients commonly demonstrate one or more established cardiovascular risk factors, there is now a considerable body of evidence indicating that OSA itself is an independent cardiovascular risk factor.3, 9, 10, 11, 12, 13, 14, 15, 16

Arterial stiffness determines how quickly the pulse wave generated by the contracting heart travels to the periphery and is reflected back. The larger arteries of the cardiovascular system are critical in the conduction of blood flow to the periphery. The elastic properties of these vessels, which are dependent on the stiffness of the arterial walls, allow for the smoothing of oscillations in blood pressure, reduction of the pulse pressure and perfusion of the myocardium.17, 18 Central pressure and arterial stiffness offer an accurate estimation of the load imposed on the coronary arteries, cerebral arteries and aorta, and, in turn, of overall vascular damage and prognosis.19, 20, 21, 22, 23, 24

Arterial stiffness can be measured noninvasively by applanation tonometry, echotracking and Doppler ultrasound. These techniques have very good reproducibility and have been widely applied.25 The most simple and reproducible noninvasive technique to date is the measurement of arterial waveforms (obtained by applanation tonometry), and more specifically pulse wave velocity (PWV), as recommended by the European Network for Non-invasive Investigation of Large Arteries25 and the European Society of Hypertension European Society of Cardiology Guidelines for the Management of Arterial Hypertension.26 PWV is inversely related to arterial distensibility27 and expresses the speed of the pressure wave traveling through the arteries. The pulse travels at a higher velocity in a stiff artery, and vice versa. Carotid-femoral PWV (cfPWV) is considered the ‘gold-standard’ measurement of the stiffness of the aorta.25, 26 It has been used extensively and has the largest amount of epidemiological evidence to support its predictive value for cardiovascular events in the general and diseased populations.25, 28, 29, 30, 31, 32

Measurements of arterial stiffness are believed to reflect global arterial endothelial function.33 A strong correlation between arterial stiffness and the development of atherosclerosis at various sites in the arterial tree has been noted.34, 35, 36 In fact, a recent study showed that arterial stiffness is an early marker of atherosclerosis, as it is affected to a greater extent when compared with intima-media thickness in patients with vascular disease.37 This result is in line with another study, which showed that arterial stiffness is an independent predictor of vascular health and coronary artery disease above and beyond intima-media thickness.38 Arterial stiffness increases chronically with age, with the development of chronic conditions (for example, hypertension), and with the presence of vascular risk factors (for example, smoking).39, 40, 41

These observations have prompted researchers investigating OSA–cardiovascular interactions to study the role of OSA in increasing arterial stiffness. The purpose of this systematic review was to collect evidence from the available literature regarding the effect of OSA on arterial stiffness to determine if arterial stiffness is increased in OSA patients.

Methods

Data sources and study selection

Studies were identified through PubMed, Embase and the Cochrane Library using the keywords shown in Figure 1. Relevant articles were extracted using the search terms, reference lists and the ‘related articles’ links of articles selected for review. Included studies satisfied the following criteria: (1) controlled clinical observational studies and/or randomized controlled trials, and (2) the participants were OSA patients in which arterial stiffness was measured using a validated technique.

Figure 1
figure 1

Search procedure.

The search for articles was limited to those published between 1985 and November 2009, and written in English without any other limitations. Three independent researchers (RJD, PS and ML) conducted the search. All the authors participated in the final selection of the included studies.

Results

Arterial stiffness in OSA vs. control groups: observational studies and baseline measurements from interventional studies

A total of 24 relevant studies were identified, confirmed by all authors and included in this review. Figure 1 illustrates the search and study selection procedure. Tables 1, 2 and 3 summarize the basic study and patient characteristics and the main findings of the included studies.

Table 1 Subject characteristics of included studies
Table 2 Arterial stiffness in OSA compared with control groups (cross-sectional studies)
Table 3 Continuous positive airway pressure treatment and arterial stiffness (randomized and nonrandomized studies)

Several studies evaluated central arterial stiffness by measuring cfPWV. Tsioufis et al.42 found a higher cfPWV in OSA patients compared with controls. Drager et al.43 noted that although cfPWV was higher in OSA patients than in controls, it was even higher in OSA patients with hypertension. Another study by the same group compared the cfPWV of non-OSA controls with a mild to moderate OSA group, and a severe OSA group. Although they did not find the mild to moderate group to significantly differ from either the controls or the severe OSA group, they did note a significant difference between severe and non-OSA participants. A further study by Drager et al.44 examined the effect of the presence of OSA in patients with metabolic syndrome on arterial stiffness. They found cfPWV to be significantly higher in the presence as compared to the absence of OSA in these patients. Moreover, there was a significant positive correlation between cfPWV and AHI.45 A subsequent study by the same group examined arterial stiffness in OSA patients with and without masked hypertension.46 It was found that, whereas both types of OSA patients showed significantly higher arterial stiffness as compared to healthy controls, OSA patients with masked hypertension had the highest arterial stiffness. Chung et al.47 compared cfPWV in patients with mild to moderate OSA, severe OSA and control subjects. They found cfPWV to be significantly higher in patients with severe OSA as compared to both mild to moderate OSA and normal controls. Protogerou et al.48 determined cfPWV, augmentation index (AIx) and central blood pressure in mild, moderate and severe OSA patients. They found that these arterial stiffness parameters did not increase with the severity of OSA after adjustment for age, gender, body mass index, heart rate, mean blood pressure, smoking and diabetes mellitus. However, respiratory disturbance index (RDI) was found to be an independent predictor of cfPWV.

Shiina et al.49 measured brachial-ankle PWV (baPWV) in OSA patients and found that they had a higher baPWV compared with those without OSA. In addition, a significant independent correlation between OSA and baPWV was noted. Similarly, another study observed that baPWV was higher in OSA compared with OSA-free control patients. This trend was maintained even in individuals free of hypertension and cardiovascular risk factors, indicating that OSA itself may be involved in the causation of stiffening of the arteries.50 This conclusion is strengthened by the findings of Tomiyama et al.,51 who found AHI to be an independent predictor of baPWV. Furthermore, Kumagai et al.52 compared patients with mild and moderate to severe OSA using the cardio-ankle vascular index. They found cardio-ankle vascular index to be lower in patients with mild OSA as compared to patients with moderate to severe OSA.

Kohler et al.53 evaluated the stiffness of the arteries in minimally symptomatic OSA patients using AIx. They found increased AIx in OSA patients compared with controls. Likewise, Noda et al.54 noted an increased AIx in OSA patients compared with non-OSA controls.

Several studies used ultrasonographic methods to evaluate arterial stiffness in OSA patients. Tanriverdi et al.55 found that the stiffness index of OSA patients was higher as compared to control patients, and that aortic distensibility was lower than in control patients. They also noted a correlation between aortic stiffness and RDI as well as a negative correlation between RDI and distensibility (r=−0.41, P=0.001). Kasikcioglu et al.56 found decreased aortic distensibility and increased stiffness index in OSA patients compared with controls. These findings were furthered by another study that found decreased distensibility in OSA patients compared with controls and a further decrease in OSA hypertensive patients.57

Effects of continuous positive airway pressure treatment on arterial stiffness

Kohler et al.58 randomized 102 male OSA patients to 4 weeks of therapeutic or subtherapeutic effects of continuous positive airway pressure (CPAP). They noted a decrease in AIx in the therapeutic group compared with a slight, nonsignificant increase in the subtherapeutic group.58 They also tested the sensitivity of the baroreflex at these time points, and found significant improvements following only therapeutic CPAP. The authors hypothesized that CPAP may improve arterial stiffness by acting upon baroreflex sensitivity. Drager et al.59 performed a similar study with 4 months of CPAP, examining cfPWV, which is the ‘gold standard’ for arterial stiffness measurements. The CPAP group had a decrease in cfPWV, C-reactive protein and catecholamines, whereas changes in the non-CPAP control group were not significant.59

Furthermore, we identified five nonrandomized clinical observational studies. Keles et al.60 measured aortic elastic parameters, aortic strain and aortic distensibility using echocardiography in 24 patients with newly diagnosed, previously untreated, moderate or severe OSA and 17 healthy controls. Following 6 months of CPAP treatment, measurements were repeated in the OSA group, and a significant improvement in aortic elastic parameters was found. Similarly, Phillips et al.61 found that 2 months of CPAP treatment significantly decreased AIx, but not peripheral blood pressure. In the same study, withdrawing CPAP for a week in patients who had undergone CPAP treatment for at least 1 year did not significantly affect AIx, but led to an increase in peripheral blood pressure in the morning as compared to the late evening.61 A study by Kitahara et al.62 found that 4 months of CPAP treatment significantly decreased baPWV. Furthermore, they found a significant correlation between the severity of sleep apnea, as measured by arousal index, and baPWV. Similarly, Shiina et al.63 found 3 months of CPAP to lead to a significant decrease in baPWV but not peripheral blood pressure. Saito et al.,64 however, found 6 months of CPAP treatment to decrease both baPWV and peripheral blood pressure in both hypertensive and normotensive patients. Interestingly, following the significant decrease observed after the first 6 months of treatment, baPWV gradually increased in both groups over the following 18 months under study, whereas peripheral blood pressure did not.

Discussion

This systematic review has presented data from 24 studies evaluating the effect of OSA on arterial stiffness. The results indicate that arterial stiffness is increased in OSA patients, and the majority of the studies point to an association between the severity of OSA and the magnitude of the effect on arterial stiffness. The consistency of the literature is of note. All studies presented in this review found an association of OSA and arterial stiffness. As to the association between the severity of OSA and arterial stiffness, the literature is somewhat less consistent. Out of the seven studies investigating the association between severity of OSA and arterial stiffness,45, 47, 48, 51, 52, 60, 62 six provided full support for the association45, 47, 51, 52, 60, 62 and one provided mixed results.48 More specifically, the study by Protogerou et al.48 did not find a direct association between OSA severity categories (moderate, severe and very severe) and arterial stiffness. However, they did find RDI to predict cfPWV significantly, the gold standard of arterial stiffness measurements. This is in line with findings by Keles et al.,60 who found a significant correlation between AHI and aortic elastic parameters. Clearly, more research is needed to further our understanding of this association.

The results presented in this review are in keeping with previous studies evaluating endothelial function of OSA patients using flow-mediated dilation. Numerous studies noted that flow-mediated dilation, measured locally in the brachial artery, was decreased in OSA patients compared with non-OSA controls,2, 53, 65, 66, 67 indicating that endothelial dysfunction occurs in OSA. However, here we discuss the available literature that furthers our understanding of the effect of OSA on arterial function, showing that not only local, but also global arterial function is impaired in OSA.

One of the strongest lines of evidence supporting an independent effect of OSA on arterial stiffness is the finding of the two randomized, controlled studies performed to date, which found that CPAP treatment of OSA significantly improved arterial stiffness.58, 59 Similar findings have also been reported from other studies evaluating OSA treatment effects on arterial stiffness which did not meet our inclusion criteria.68 Also of note, Phillips et al.61 found that 8 weeks of CPAP treatment lead to reductions in AIx and central systolic blood pressure without a concomitant decrease in peripheral blood pressure. Thus, although the numbers are relatively small in the randomized studies to date, and further larger studies are required with longer follow-up of both arterial stiffness and cardiovascular event outcomes, the current literature points to an independent association between OSA and arterial stiffness. However, as only two randomized trials have been performed to date, more randomized trials are needed to further explore this relationship.

Importantly, it should be noted that arterial stiffness was found to be elevated throughout the day, in the awake state. This implies a carryover effect of OSA on arterial stiffness. More specifically, arterial stiffness is not only elevated acutely at night, when the respiratory disturbance occurs, as shown by Jelic et al.,69 but also during the awake state.

Mechanisms

A number of mechanisms could explain the role of OSA in increasing arterial stiffness. Inflammation, oxidative stress and sympathetic activity affect endothelial function and are all altered in OSA.9, 10 Intermittent hypoxia is associated with increased production of reactive oxygen species, and therefore, increases oxidative stress.70, 71 Prolonged oxidative stress disturbs cellular function, promotes endothelial dysfunction and increases inflammation, which can further increase the metabolic and cardiovascular complications of OSA such as atherosclerosis.70 Oxidative stress also reduces the activity of nitric oxide synthase and thus leads to decreased production of nitric oxide, increasing arterial stiffness.72 Various studies have also found that C-reactive protein, tumor necrosis factor-α, interleukin-6 (IL-6), IL-8 and cell adhesion molecules such as intercellular adhesion molecule-1 are increased in the circulation in OSA patients independent of obesity.10, 15, 73, 74 Increased inflammation may have a role in the pathogenesis of OSA, increasing arterial stiffness and contributing to the early atherosclerosis found in OSA patients.3 Furthermore, it also appears that endothelial dysfunction in OSA is partially mediated through apoptotic dysregulation; circulating apoptotic endothelial cells were found to be associated with AHI (r=0.56, P=0.004). This was partially reversed after 8 weeks of CPAP treatment.2

Therapeutic and clinical implications

It has been well established that OSA is an independent risk factor for hypertension, myocardial infarction and stroke.2, 3, 4, 5, 6, 7 Arterial stiffness is independently associated with cardiovascular risk19, 20, 21, 22, 23, 24, 26, 28, 29, 30, 31, 75, 76, 77, 78 and may therefore provide a potential explanation for the increased risk of cardiovascular events in OSA patients. Furthermore, a strong correlation between arterial stiffness and the development of atherosclerosis at various sites in the arteries has been noted.35, 36 Therefore, measuring arterial stiffness can be an important clinical tool for the monitoring of disease progression and treatment efficacy, as recommended by the European Network for Non-invasive Investigation of Large Arteries25 and the 2007 European guidelines for the management of arterial hypertension.26 The use of arterial stiffness measurements may be particularly useful in assessing the endothelial damage induced by certain cardiovascular risk factors, such as OSA. However, not all parameters of arterial stiffness are equal; cfPWV is the best measure of arterial stiffness. It is considered the ‘gold standard’ for the assessment of aortic stiffness,25, 26 and cfPWV and AIx have the greatest amount of evidence to support their predictive value for cardiovascular events in healthy and diseased populations.26, 28, 29, 30, 75, 79, 80 Therefore, other parameters of arterial stiffness, such as local distensibility or compliance, may not be as valuable to assess the vascular effects of OSA and are not as useful clinical parameters.

Limitations

This review has a number of limitations. First, the studies use slightly different protocols and methods; the time of day and the conditions under which the studies took place (for example, fasting) were not standardized between studies or not reported. Furthermore, the age groups of subjects, severity of OSA, body mass index, renal function, lipid profile, menopausal status, medication use, including statin use, were not uniform between studies or subject groups. Furthermore, studies were typically cross-sectional, with small sample sizes, and links to long-term outcomes (for example, cardiovascular events or progression of sleep apnea) have yet to be investigated. Moreover, studies examining the effect of CPAP on arterial stiffness have had short follow-ups, thereby not providing information about long-term outcomes.

These factors were also often not reported or corrected for; clearly, future studies should recruit more homogeneous populations. The methods of assessing arterial stiffness also differed between studies; however, it is encouraging that eight studies used cfPWV, AIx or both as parameters of arterial stiffness.42, 43, 44, 45, 47, 48, 53, 54 Furthermore, limiting our search to English language publications may have led to an omission of potentially relevant studies. Owing to methodological constraints, a meta-analysis of studies aiming to also quantify the overall magnitude of the effect of OSA on arterial stiffness could not be conducted.

Many studies did not report the time of the day at which arterial stiffness measurements were performed. A study by Phillips et al.81 demonstrated circadian variation of arterial stiffness in OSA patients. More specifically, arterial stiffness was found to be higher in the morning as opposed to the evening. Although Phillips et al. did not recruit healthy controls to compare this effect in both healthy and OSA patients, other studies have shown this circadian effect also in healthy populations.82 Furthermore, a study in healthy individuals by Papaioannou et al.82 suggested that the aforementioned circadian effect is at least in part due to increased blood pressure in the morning. As to OSA patients, it may represent a carryover effect from increased stiffness as well as blood pressure observed during respiratory disturbances at night.69 Nevertheless, it should be noted that this effect still bears importance, as most cardiovascular events take place in the morning.83, 84

Conclusion

To the best of our knowledge this is the first systematic review on the effect of OSA on arterial stiffness. We have reviewed results from 24 studies evaluating the effect of OSA on arterial stiffness. Arterial stiffness is increased in OSA patients compared with non-OSA control patients. The magnitude of the increase in arterial stiffness appears to be greater in more severely affected OSA patients and in those with other cardiovascular risk factors, such as hypertension. Therefore, arterial stiffness may have a role in the increased risk of cardiovascular complications in OSA. Interestingly, the use of CPAP was found to improve arterial stiffness in patients with OSA. Assessment of arterial stiffness in OSA patients by means of currently available, convenient and noninvasive methods may be an effective method to also monitor disease progression and treatment efficacy.