Sleep Apnea and Cardiovascular Disease

Abstract

Purpose of review:

Obstructive sleep apnea (OSA) is associated with incident coronary and cerebral vascular disease. The mechanisms underlying this association are thought to include increased sympathetic nervous system activity, oxidative stress, and systemic inflammation, with these effects mediated in part by elevated blood pressure and impaired glucose metabolism. In observational studies, OSA treatment with positive airway pressure (PAP) is associated with a reduction in cardiovascular disease risk. The aim of this review is to evaluate evidence from recent clinical trials that tested the impact of OSA treatment on major cardiovascular disease outcomes.

Recent findings:

Multicenter randomized trials have demonstrated a significant, albeit modest, reduction in blood pressure with OSA treatment. Treatment of OSA has generally not demonstrated improvement in type 2 diabetes mellitus, although limited evidence suggests that treatment may be effective in the prediabetic period. However, recent randomized trials of PAP treatment for OSA failed to demonstrate a reduction in incident or recurrent cardiovascular disease events. This may reflect the enrollment of a mostly non-sleepy study sample, as recent evidence suggests that sleepiness is a predictor of adverse cardiovascular outcomes from OSA.

Summary:

PAP treatment of OSA lowers blood pressure and may improve glucose metabolism; however, randomized clinical trials do not indicate a reduction in cardiovascular risk with treatment of minimally symptomatic OSA patients.

Introduction

Obstructive sleep apnea (OSA) is characterized by repeated episodes of partial or complete airway collapse during sleep, resulting in reduced airflow with consequent intermittent hypoxemia and hypercapnia, large swings in intrathoracic pressure, and transient arousal at the termination of the respiratory event. Early case series indicated a high prevalence of OSA in patients with established cardiovascular disease, although a causal association was unclear, as OSA and cardiovascular disease share multiple common risk factors, including obesity, male sex, and older age. Clinic-based case-control [1–4] and longitudinal [4–11] studies have demonstrated that this association persists after adjustment for multiple potential confounders, including body mass index, with a two- to five-fold increase in cardiovascular disease risk reported in patients with OSA. These observational study designs are limited by the potential for unmeasured confounding, particularly as body mass index is only modestly correlated with visceral adiposity, which appears to be more strongly related to OSA severity than is total body adiposity [12], and is also a stronger correlate of cardiovascular and metabolic disease risk. While sleep clinic-based observational studies demonstrate that OSA patients adherent to positive airway pressure (PAP) therapy have a lower risk of major adverse cardiovascular and cerebrovascular events (MACCE) than patients who decline or are non-adherent to therapy [4,6,7,13,14], these studies have a high risk of bias due to a “healthy adherer” effect, which has been found in multiple prior cardiovascular clinical trials to be larger than the effect of assignment to the active treatment group [15,16]. A causal relation of OSA to vascular disease has strong biological plausibility, however, with both animal studies and human physiological studies suggesting that increased sympathetic nervous system activity, oxidative stress, and systemic inflammation play important subclinical pathogenic roles [17]. These effects are thought to be mediated in part by clinical intermediates of increased blood pressure, including nocturnal blood pressure non-dipping, and impaired glucose metabolism. Small randomized clinical trials have shed light on the impact of OSA treatment on these clinical intermediates, and several large randomized clinical trials have recently addressed the question whether PAP treatment of OSA reduces the risk of incident or recurrent MACCE. The purpose of this review is to critically evaluate the evidence from these trials, which generally found no evidence of reduced cardiovascular risk with PAP therapy.

Impact of OSA treatment on blood pressure

Elevated sympathetic nervous system activity during both sleep and wakefulness has long been recognized in patients with OSA. As both muscle sympathetic nerve activity and urinary excretion of catecholamines decrease with effective treatment of OSA, the elevated sympathetic nervous system activity is due to OSA, and not to confounding by factors such as adiposity [18,19]. This increased sympathetic activity is a result of OSA-related intermittent hypoxemia, hypercapnia, and sleep fragmentation. While increased sympathetic activity may contribute to cardiovascular disease via multiple pathways, the most well-established mechanism is through blood pressure elevation. Animal models have demonstrated that sympathetic activity is the major cause of both the acute increase in blood pressure that occurs with obstructive apneas and the sustained hypertension that develops with intermittent hypoxemia in rodents or experimental obstructive sleep apnea in dogs [20–23].

Hypertension is highly prevalent in OSA patients, and numerous small, single-center studies have reported improvement in blood pressure with PAP treatment. In the past decade, several larger, multicenter trials have confirmed a benefit of PAP therapy for reducing blood pressure, although the overall effect was modest. One study screened patients with newly diagnosed hypertension for the presence of OSA and randomized 340 patients with moderate to severe OSA to either active or sham PAP therapy. This study found that over 12 weeks of treatment, the 24-hour mean arterial pressure was reduced by 1.5 (95%CI 0.4 to 2.7) mmHg [24]. Another study randomized 359 non-sleepy patients with moderate to severe OSA and hypertension to either PAP or conservative management, and found that the impact of PAP on blood pressure increased throughout the 12-month follow-up period, with an approximately 2.7 mmHg lower mean arterial pressure than conservative management at the end of 12 months [25]. We randomized 318 patients with multiple cardiovascular disease risk factors who screened positive for moderate to severe OSA to either PAP, nocturnal supplemental oxygen, or a sleep education control for a period of 12 weeks. Despite blood pressure that was generally well controlled at baseline (mean 24-hour blood pressure of 124/71 [SD 15/8]), PAP was associated with a 2.8 mmHg greater reduction in mean arterial pressure compared to control, while supplemental oxygen alone did not reduce blood pressure [26]. The effect of PAP on blood pressure is somewhat higher in patients with OSA and resistant hypertension. In the HIPARCO study, 194 patients with resistant hypertension and moderate to severe OSA were randomized to PAP or a usual care control for 12 weeks. PAP was associated with a 3.9 (95%CI 1.3 to 6.6) mmHg reduction in 24-hour mean arterial pressure, with even larger effects reported in several smaller trials of PAP therapy in patients with resistant hypertension [27,28]. These mean pressure measures may understate the potential benefits of OSA treatment, however, as the three studies that measured 24-hour ambulatory blood pressure each found a greater reduction in nocturnal than diurnal blood pressure, with a significant reduction in the proportion of patients with nocturnal non-dipping blood pressure, which may carry a particularly increased risk of cardiovascular disease [24,26,27]. In a study of OSA patients with good long-term adherence to PAP, 153 patients were randomized to either subtherapeutic PAP or continued therapeutic PAP for a period of 2 weeks. Compared to patients continued on therapeutic PAP, those randomized to subtherapeutic PAP had an increase in morning systolic blood pressure of 9.0 (95%CI 5.7-12.3) mmHg and diastolic blood pressure of 7.8 (95%CI 5.6 to 10.0) mmHg, suggesting that over time and with good PAP adherence a considerably larger impact of PAP therapy on blood pressure may be attainable [29].

Impact of OSA treatment on glucose metabolism

While obesity is an important risk factor for both OSA and type 2 diabetes mellitus, a recent meta-analysis of prospective observational studies found that OSA is associated with a significantly increased risk of incident type 2 diabetes independent of obesity, with an adjusted relative risk of 1.35 (95%CI 1.24 to 1.47) [30]. Animal models of OSA have shown that intermittent hypoxia induces insulin resistance, particularly in the setting of obesity, an effect that may be mediated by a combination of sympathetic nervous system activation, production of reactive oxygen species, and systemic inflammation [31]. Short-term human studies using the frequently-sampled intravenous glucose tolerance test have demonstrated that as little as two nights of sleep fragmentation or a single 5-hour period of exposure to intermittent hypoxemia will induce insulin resistance [32,33]. Despite the association of OSA with type 2 diabetes in observational studies and the mechanistic evidence from animal studies, clinical trials of OSA treatment in patients with established type 2 diabetes have generally failed to demonstrate improvement in diabetes control [30]. In the largest such trial, 298 patients with type 2 diabetes who screened positive for moderate to severe OSA were randomized to usual care with or without PAP therapy. Although the PAP-treated group had a substantially larger reduction in sleepiness, there was no significant difference between groups in measures of glycosylated hemoglobin or fasting blood glucose [34].

The failure of PAP therapy to improve measures of glycemic control in established type 2 diabetes may reflect irreversible damage from OSA, as animal models suggest only partial reversibility of the effects of intermittent hypoxia, with continued worsening of inflammation in pancreatic islets when the hypoxia exposure is discontinued [35]. Several small human studies have demonstrated improvement in insulin sensitivity following PAP therapy in patients with OSA and impaired glucose tolerance or prediabetes [36–38]. To date, no randomized clinical trials have examined the impact of OSA treatment to prevent progression from prediabetes to overt type 2 diabetes.

Impact of OSA treatment on incident cardiovascular disease: results of recent clinical trials

While non-randomized clinic-based studies suggest that PAP therapy of OSA reduces risk of MACCE to that of patients without OSA, recent randomized clinical trials saw no reduction in cardiovascular risk with PAP therapy as either primary or secondary prevention [39–42]. The one study of primary prevention of cardiovascular disease randomized 725 patients with moderate-to-severe OSA to either usual care or PAP therapy. Patients who reported excessive sleepiness or had an elevated score on the widely used Epworth Sleepiness Scale questionnaire were excluded. The primary outcome was a composite of incident hypertension and incident cardiovascular disease. During median follow-up of 4 years, there were 59 incident cardiovascular events, with no significant difference between the PAP therapy and control groups [39]. The Randomized Intervention with Continuous Positive Airway Pressure in CAD and OSA (RICCADSA) study randomized 244 non-sleepy patients who screened positive for moderate-to-severe OSA following coronary revascularization procedures, with half receiving PAP therapy and half usual care. Median follow-up was 57 months. The primary endpoint was a composite of myocardial infarction, stroke, need for repeat coronary revascularization, or cardiovascular death. The primary endpoint was reached in 18.1% of the PAP group and 22.1% of the usual care group (hazard ratio 0.80, 95% CI 0.46-1.41). The effect was not significant for the composite endpoint or for any of its individual components [40]. Both of these studies were underpowered to detect potentially important risk reduction with PAP therapy; however, two larger randomized trials reached a similar conclusion. The Sleep Apnea Cardiovascular Endpoints (SAVE) study screened patients with known cardiovascular or cerebrovascular disease for the presence of OSA, and randomized 2717 patients with moderate-to-severe OSA to either PAP or usual care. The threshold for exclusion of patients based on excessive sleepiness was less stringent than in the RICCADSA study, but the enrolled patients were mostly non-sleepy. Although well powered to detect a 25% reduction in risk of MACCE (myocardial infarction, acute coronary syndrome, stroke, transient ischemic attack, cardiovascular death, or heart failure hospitalization), no significant impact was noted on the primary composite end point (hazard ratio 1.10, 95% CI 0.91-1.32) or any of its components [41]. The Impact of Sleep Apnea Syndrome in the Evolution of Acute Coronary Syndrome (ISAACC) screened patients with acute coronary syndrome for the presence of OSA, and randomized 1264 patients with moderate-to-severe OSA to either PAP or usual care. This study was also powered to detect a 25% reduction in MACCE. Excessively sleepy patients were excluded. Again, there was no significant difference between PAP and usual care groups in the rate of the primary composite outcome or in any of its components [42].

These studies have been criticized for generally low adherence to PAP therapy, which is not surprising given the lack of excessive sleepiness in most patients. When the patients randomized to PAP therapy were stratified by adherence to therapy, two of the studies reported significantly reduced risk of incident or recurrent cardiovascular events in patients using PAP for 4 or more hours per night [39,40], while the other two studies found that adherent patients had a non-significant but potentially clinically important 20% risk reduction [41,42]. While this suggests that meaningful reductions in cardiovascular risk may occur in OSA patients who are adherent to PAP therapy, these analyses are susceptible to the same healthy adherer bias as are observational cohort studies.

Another possible explanation for the discrepancy between clinic-based observational studies and these randomized clinical trials is the exclusion of sleepy patients from the randomized trials. Because PAP has been clearly demonstrated to improve OSA-related sleepiness, this exclusion reflects a concern that it would be unethical to randomize sleepy patients to no PAP therapy for an extended period of time. There is growing evidence, however, that sleepiness is an important marker of OSA-related cardiovascular risk. In the Sleep Heart Health Study, for instance, excess cardiovascular risk was observed only in those participants whose OSA was associated with excessive sleepiness [43]. This finding has recently been replicated in an independent cohort of 780 patients with moderate-to-severe OSA, in which patients with an excessively sleepy phenotype had more than a 3-fold increased risk of cardiovascular mortality compared to OSA patients without excessive sleepiness [44]. A similar association of sleepiness with increased risk of recurrent MACCE has been reported for post-myocardial infarction patients with moderate to severe OSA [45]. Thus, exclusion of sleepy patients from randomized clinical trials may have limited the ability of recent randomized trials to detect a benefit of PAP therapy.

Conclusions

OSA is strongly associated with incident and recurrent coronary and cerebral vascular disease in both clinic patients and population-based cohorts. This association appears to be independent of shared risk factors including adiposity, and physiologic mechanistic studies in both human and animal models provide ample evidence for a causal association of OSA with vascular disease. Clinic-based observational studies indicate lower rates of MACCE in OSA patients treated with PAP therapy, and randomized clinical trials have demonstrated modest reductions in blood pressure with treatment. In contrast, several recent randomized clinical trials in non-sleepy patients with OSA have not found a reduction in cardiovascular risk with PAP therapy. However, these randomized trials generally excluded patients who were excessively sleepy, one of the cardinal features of clinical OSA syndrome. As emerging data suggest that OSA is associated with excess cardiovascular risk only in OSA patients with a sleepy phenotype, the exclusion of sleepy patients from randomized clinical trials may explain these null findings. Nonetheless, while treatment of OSA for control of symptoms is clearly indicated, treatment of OSA in asymptomatic or minimally symptomatic patients solely for the purpose of reducing cardiovascular risk is not currently justified by high-grade evidence.