Sex Differences in Atherosclerotic Cardiovascular Disease Risk in Obstructive Sleep Apnea

Abstract

Purpose of Review:

This review outlines obstructive sleep apnea (OSA) associated atherosclerotic cardiovascular disease (ASCVD) risk and highlights emerging data suggestive of sex differences.

Recent Findings:

Females with OSA have greater hypertension risk, higher carotid intima-media thickness, elevated cardiac enzymes, and worse outcomes following ischemic cardiovascular events relative to males with OSA. Mechanistically, this parallels sex differences in nocturnal hypoxemia, immune cell activity, inflammation, and endothelial function which frequently coincide with low estrogen levels.

Summary:

OSA-associated ASCVD risk appears more pronounced in females than males. This could be attributable to sex differences in the etiology of OSA and resultant activation of pathophysiological mechanisms. However, more data are required to differentiate causality from epiphenomena and develop individualized therapies to mitigate ASCVD in patients with OSA.

Introduction

Atherosclerotic cardiovascular disease (ASCVD) is a leading cause of death worldwide. Traditionally, ASCVD is discussed as a pathological process that occurs within the coronary vasculature and leads to myocardial infarction. Established risk factors for ASCVD include cardiometabolic pathologies (hypertension, type 2 diabetes), lifestyle patterns (diet, exercise, tobacco use), in addition to genetic predisposition (e.g., PCSK9 mutations). The American Heart Association’s recent publication “Life’s Essential 8” listed sleep alongside these risk factors as a primary contributor to cardiovascular risk [1]. Indeed, prolonged habitual sleep duration [2] and irregular sleep/wake cycles [3] are associated with increased ASCVD risk. Outside of poor sleep habits, sleep pathologies are also associated with atherogenesis. Obstructive sleep apnea (OSA) is among the most common sleep disorders and markedly increases ASCVD risk [4]. While OSA is historically more prevalent in males [5], emerging evidence suggests that OSA may increase ASCVD risk to a greater extent in females compared to males. When considering that women are under referred to sleep clinics [6], and that ischemic stroke and heart disease are the leading causes of death in women globally [7], it is clear that more work is needed to understand sex differences in OSA-associated ASCVD risk. The present review will provide a brief overview of atherosclerotic lesions, the etiology of OSA, and OSA-associated ASCVD risk. We will then describe sex differences in the etiology and pathophysiology of OSA, review potential mechanisms, and conclude by offering opportunities for future research. Throughout the manuscript, terminology for both sex and gender will be used to describe differences in clinical or physiological observations. While we recognize these as independent sociocultural domains, we will reference data within the literature as described by the respective authors. To this point, an emphasis will be placed on data collected in humans with findings from animals being discussed in the absence of human studies. Furthermore, preclinical studies historically do not perform statistical comparisons on sex differences which limits their utility in the present review [8].

Formation of Atherosclerotic Lesions

It is important to differentiate atherosclerosis from arteriosclerosis. Although both contribute to cardiovascular events, they represent distinctly different biological processes. Arteriosclerosis reflects, broadly, the “stiffening” of arteries through a loss of elastin that often occurs concomitantly with collagen deposition within the arterial wall. Atherosclerosis refers to the intrusion of lipids through the vascular endothelium into the media causing smooth muscle cell proliferation, macrophage activation, and the formation of foam cells. Although there is some physiological overlap between arterio- and atherosclerosis, this review will focus on the latter.

Briefly, the formation of an atherosclerotic lesion is initiated by plasma lipoproteins passing through “leaky” endothelial cell junctions and entering the tunica intima. These lipids are aggregated and oxidized which causes an increased expression of adhesion markers on the surface of vascular endothelial cells. Circulating monocytes bind to these markers permitting their entrance into the tunica intima where they differentiate into macrophages. These macrophages interact with intimal lipids to form foam cells which create the lesion’s necrotic core. As the core develops, vascular smooth muscle cells (VSMCs) proliferate further impairing vasodilation. Larger plaques progressively invade the vascular lumen to reduce blood flow and can cause symptoms such as angina. Smaller plaques can become dislodged from the vessel wall, enter circulation, and occlude perfusion entirely as observed with ischemic strokes. Accordingly, elevations in circulating cholesterol levels, production of pro-inflammatory cytokines, immune disorders, and damage to the vascular endothelium all contribute to increasing ASCVD risk. We direct readers elsewhere for a comprehensive and contemporary review of atherosclerosis [9].

Overview of OSA

Patients with OSA have intermittent complete or partial interruptions in airway patency that reduce airflow while sleeping termed apneas and hypopneas, respectively. These events are summed then indexed per hour of sleep to generate the apnea-hypopnea index (AHI) which is used to identify the presence, and quantify the severity, of OSA. Paralleling these events are a resultant decline in arterial oxygen levels and elevation of carbon dioxide concentrations which often precipitate an arousal from sleep. Classically, OSA is treated by positive airway pressure (PAP) which, when used as prescribed, maintains nocturnal airway patency. Importantly, three randomized clinical trials from across the globe suggest that PAP does not mitigate ASCVD risk [10–12]. This has prompted the investigation of add-on therapies which complement PAP [13] as well as further exploration of pathophysiological mechanisms.

OSA-Associated ASCVD Risk

Patients with OSA have a marked increase in the likelihood of ASCVD evidenced by a more than two-fold risk of myocardial infarction, coronary revascularization, or death from cardiovascular disease [14]. Similar observations also apply to the risk of stroke [15] which have been replicated over the past five years [16]. As the effects of nocturnal blood gas fluctuations have systemic implications, there is no single mechanism or anatomically regionalized origin for OSA-associated ASCVD risk. The following subsections outline prominent contributors to atherosclerosis in OSA, and emerging evidence on mechanisms of action.

Adipose Tissue

Seminal data from Peppard et al. illustrate the intimate relationship between adiposity and OSA. Over four years, a 10% increase in body mass predicted a 32% increase in patients’ AHI. Conversely, reducing body mass by 10% was associated with a 26% reduction in their AHI [17]. The accumulation of adipose tissue, particularly in the abdominal cavity, is also associated with ASCVD risk [18]. Recent analyses of Multi-Ethnic Study of Atherosclerosis (MESA) data found that a 10cm² increase in visceral adiposity is associated with increased coronary artery calcification after adjusting for ASCVD risk factors such as body mass index, smoking status, and blood pressure [19]. This seemingly causal relationship between the accumulation of adipose tissue and OSA-associated ASCVD was an impetus for exploring tirzepatide, a glucagon-like peptide 1 analog, as a treatment for OSA. Recent findings from the SURMOUNT-OSA trial indicate that 52 weeks of tirzepatide reduced the AHI by 20.0–23.0 events/hr which paralleled reductions in ASCVD risk factors such as body mass, blood pressure, and inflammation [20]. In a smaller sample, O’Donnell et al. studied the effects of 24 weeks of liraglutide with and without concomitant PAP therapy on ASCVD risk factors in patients with OSA [21]. Interestingly, their findings indicate that liraglutide in isolation reduces visceral adiposity whereas liraglutide with PAP did not yield the same results. Epicardial adipose tissue (EAT), a subdomain of visceral adipose, accumulates alongside the heart and is associated with increased risk of myocardial infarction [22]. More recent studies have integrated EAT into ASCVD prediction models with some data supporting [23, 24], and refuting [25], its prognostic implications. This may indicate that the volume of adipose tissue may be less relevant to ASCVD risk than the adipocytes’ physiology. To this point, our in vitro experiments suggest that the intermittent hypoxia which characterizes OSA promotes senescence in adipocytes [26]. Here, senescent cells can secrete pro-inflammatory cytokines that impair adjacent cells and tissues and are implicated in a wide array of pathobiology [27]. Given the proximity of visceral adipose tissue (e.g., EAT) to key organs, it is likely that adipocyte-derived inflammation is one key mechanism of OSA-associated ASCVD risk.

Systemic Inflammation

Early experiments demonstrated that intermittent hypoxia activates the nuclear factor-kappa B (NF-κB) system in HeLa cells which corroborates the increased tumor necrosis factor-alpha (TNFα) levels in patients with OSA [28]. Subsequent studies found that nocturnal hypoxemia and serum cholesterol were independent predictors of TNFα levels in patients with OSA [29]. Díaz-García et al. recently reported that oxidized cholesterol upregulates NF-κB signaling in monocytes from patients with OSA which, through NLRP3 activation, promotes systemic inflammation and the release of tissue factor [30]. Interestingly, tissue factor levels were positively associated with carotid intima-media thickness (CIMT), a marker of subclinical atherosclerosis [31], providing causal evidence for inflammation in OSA-associated ASCVD risk. Indeed, data from the Brazilian Longitudinal Study of Adult Health (ELSE-Brazil) study report that even mild OSA is associated with increased CIMT and the strength of this relationship is augmented by C-reactive protein and triglyceride levels [32]. These studies seemingly indicate a causal relationship of OSA with inflammation; however, prospective data suggest this association could be bidirectional. This evidence originated from data collected prospectively from the Nurses’ Heart Health (NHS and NHSII), Health Professionals Follow-up Study (HFPS), and the Multi-Ethnic Study of Atherosclerosis (MESA) projects. A pooled analysis revealed that higher C-reactive protein levels corresponded to a greater risk of developing OSA, although this risk was nearly abolished when adjusting for body mass index [33]. Regardless of causality, patients with OSA have significantly elevated levels of inflammation which increases their ASCVD risk. One mechanism for this increased risk is cytokine-derived damage to the vasculature.

Hypertension

The link between sleep-disordered breathing and vascular dysfunction dates back to the 1970s [34]. Following these early observations, increases in sympathetic-mediated vasoconstriction were observed in OSA [35] along with reduced vasodilatory responses to acetylcholine infusions into the brachial artery (e.g., microvascular endothelial function) [36]. Thereafter, a reduction in nitric oxide synthase phosphorylation, as well as higher levels of cyclooxygenase and nitrotyrosine, were reported in arterial endothelial cells of patients with OSA [37]. Counterintuitively, acetylcholine-mediated vasodilation in coronary arteries does not appear to be impaired in patients with OSA [38]. One explanation for this finding is a potential shift in the primary vasodilator from nitric oxide (mechanism for acetylcholine-induced vasodilation) to mitochondria-derived hydrogen peroxide in patients with ASCVD. Indeed, a recent study illustrated a key role for lipid phosphate phosphatase 3 via miR-92 in regulating this transition [39]. Data from Shang and colleagues recently found that the hypoxia-sensitive miR-210 may be instrumental in disrupting endothelial cell function by suppressing mitochondrial bioenergetics [40]. While these studies indicate a supporting, rather than starring, role for nitric oxide in OSA-associated ASCVD risk, increasing nitric oxide bioavailability may still confer cardioprotection in patients with OSA [41].

Taken together, the increased ASCVD risk in patients with OSA is well-established and multifactorial. Evidence supports a role for both the accumulation of adiposity as well as changes in adipocyte functionality, inflammation, vascular dysfunction, and hypertension. It should be noted that this list is not comprehensive and that these mechanisms are interdependent. Interestingly, it appears the end result of OSA on ASCVD risk appears to be more deleterious to females compared to males although the mechanism(s) of this sexual dimorphism are incompletely understood.

Sex Differences in Atherosclerosis

Before discussing how OSA impacts ASCVD risk differently between sexes, it is important to first provide an overview of sex differences in the pathophysiology and clinical impact of atherosclerosis. Broadly, men are considered to have greater ASCVD risk than women throughout most of the lifespan and tend to have higher blood cholesterol levels [42]. However, the associations between cardiovascular risk factors such as smoking, hypertension, and type 2 diabetes with ischemic cardiovascular events are stronger in women [43, 44]. This discrepancy may indicate a sexual dimorphism in plaque formation, composition, and/or stability. Indeed, data from the Providing Regional Observations to Study Predictors of Events in the Coronary Tree (PROSPECT) study indicate that, in patients with acute coronary syndrome, women have fewer non-culprit lesions, a smaller necrotic core, and were less likely to have plaque rupture compared to men [45]. Similar findings were observed by Kataoka et al. who noted that non-culprit atherosclerotic lesions in women were more likely to have signs of erosion, but less likely to indicate calcification and contain cholesterol crystals relative to men [46]. Although these data indicate that females have lower ASCVD risk than males due to biological differences in atherosclerotic plaques, sex-specific life events in females increase their ASCVD risk.

One example is pregnancy whereby females with an adverse pregnancy outcome, such as gestational diabetes, had greater three-year increases in ASCVD risk indexed via the Framingham Risk Score relative to females with “normal” pregnancies [47]. Data from the Women’s Health Initiative corroborates this notion whereby pregnancy-related hypertension increased the risk of ASCVD by 27% after adjusting for traditional risk factors (e.g., body mass) [48]. Preeclampsia specifically is associated with a 30% greater risk of coronary atherosclerosis and a two-fold increased risk for significant stenosis [49]. There is growing interest in studying the contribution of sleep-related breathing disruptions to cardiovascular health with pregnancy as clinical [50] and sub-clinical [51] nocturnal interruptions in airflow are indicative of hypertensive pregnancies. There are also stark differences in the etiology and pathophysiology of OSA that have implications for ASCVD risk.

Sex Differences in OSA

Etiology

Although it is widely accepted that OSA is more common in males than females, important sex differences in the etiology of OSA are worth noting. Principally, data collected during polysomnography indicate that OSA is more manifest during rapid eye-movement (REM) sleep in females when compared to males [52, 53]. As muscle atonia is a characteristic of REM sleep [54], this may indicate that females are more susceptible to airway collapse during REM. Indeed, tracheal diameter tends to be smaller in women compared to men and is associated with the AHI in women, but not men, with OSA [55]. However, a recent study by Hang et al. found that upper airway collapsibility was greater in men than women with OSA [56]. Interestingly, the authors noted that women with OSA had a greater rate of rise in loop gain with aging compared to men with OSA indicating a greater susceptibility to respiratory instability. As the ventilatory response to changes in arterial blood gas levels, specifically oxygen, is indicative of hypertension with OSA [35], this may indicate that OSA increases ASCVD risk to a greater extent in females than males. This notion is supported by data from Ko and colleagues who found that OSA increases the likelihood of ASCVD (e.g., myocardial infarction, stroke) by 72% in women but only 27% in men [57]. Females with OSA also have greater impairments in cognition and depressive symptoms after stroke compared to males [58] with similar results following myocardial infarction [59]. In summary, the intermittent changes in nocturnal airway patency which characterize OSA are REM-dominant in females, relative to males. This may have implications for ASCVD risk with recent studies purporting diverse mechanisms of action.

Evidence from Observational Studies

Some of the strongest evidence for a sexual dimorphism in OSA-associated ASCVD risk comes from a pooled analysis of data from the Sleep Heart Health Study (SHHS) and the Atherosclerosis Risk in Communities (ARIC) project. In this analysis, Roca et al. observed a greater impact of OSA on risk of a composite outcome inclusive of heart failure and cardiovascular mortality in women compared to men [60]. The authors also noted a positive association between left ventricular mass index and the severity of OSA in women, but not men, which paralleled high-sensitivity troponin-T levels. Subsequent data support a role for blood pressure dysregulation in these sex differences. That is, the respiratory event index (a proxy for AHI) is more strongly associated with hypertension in women than men after adjusting for age and body mass index [61]. Perhaps the most clinically insightful finding from this study was that this sex difference was apparent at a “subclinical” frequency of respiratory events (e.g., below 5.0 events/hr). Of direct relevance to ASCVD, retrospective data illustrate that the AHI during REM sleep is positively associated with CIMT [62]. It is important to recall that females with OSA have a REM-dominant pathology as compared to males. Indeed, when stratified by sex, the relationship between the AHI during REM sleep and CIMT was exclusively observed in women [62]. These conclusions were corroborated by data collected prospectively in 92 women [63]. The authors noted that the AHI during REM sleep, but not traditional ASCVD risk factors (e.g., blood pressure), at baseline predicted increases in CIMT over the 10-year follow up period. Collectively, these studies indicate that OSA increases ASCVD risk to a greater extent in females compared to males which translates to a sex difference in OSA-associated cardiovascular mortality. While our understanding of the mechanisms for sex differences in OSA-associated ASCVD risk continues to develop, several studies purport a protective role for estrogen.

Multifactorial Role for Estrogen

Simplistically, the marked increase in OSA prevalence which occurs after the menopausal transition insinuates that reduced estrogen production increases the risk of OSA in females [64]. This parallels an increase in many other ASCVD risk factors such as blood pressure, lipid levels, inflammation, and adiposity [65]. However, whether these observations are related, interdependent, epiphenomena, or attributable to aging is challenging to elucidate in humans. Evidence for a causal role of estrogen deficiency in OSA and ASCVD risk comes from a large-scale observational study of women undergoing premenopausal unilateral or bilateral oophorectomy. Data from Mielke et al. noted a two-fold increase in the risk of developing OSA during the 20 years post-oophorectomy [66]. Prior work from this group illustrates a 44% increase in cardiovascular risk following oophorectomy which appeared to be ameliorated in women receiving estrogen therapy [67]. The landmark Early Versus Late Intervention Trial with Estradiol (ELITE) study found that starting estrogen therapy within six years of initiating menopause delayed the age-related progression in CIMT by approximately 50% [68] supporting earlier work reporting a reduction in cardiovascular events and mortality [69]. However, this benefit was not observed in women initiating estrogen ≥10 years after the onset of menopausal symptoms. While these data strongly support ASCVD risk reduction via early adoption of estrogen therapy, recent evidence also suggests that estrogen therapy can reduce the frequency of respiratory events during REM sleep in women [70]. Data from the Kronos Early Estrogen Prevention Study (KEEPS) found that transdermal, compared to oral, estrogen therapy improves subjective sleep quality [71] which may lead to a commensurate reduction in ASCVD risk. However, the KEEPS study found that neither transdermal nor oral delivery of estrogen improved CIMT or coronary calcification [72]. In summary, OSA-associated ASCVD risk may be reduced by estrogen therapy if initiated shortly after the onset of menopause. Despite reductions in subclinical atherosclerosis being reported in some studies with estrogen therapy, the improvements in subjective sleep quality and nocturnal respiratory patterns make mechanistic interpretations challenging. To this point, estrogen receptors are expressed on many immune cell populations that are involved in the development of atherosclerotic lesions [73] and subsequent modulation of these cell may contribute to the sex differences in OSA-associated ASCVD risk.

Immune Dysregulation

Seminal data report that experimental sleep restriction increases the number of circulating immune cells that have pro-atherogenic properties such as natural killer (NK) cells and monocytes [74]. Succeeding experiments show that the immunoinflammatory responses to experimental sleep restriction diverge between sexes with females showing a greater response than males [75]. We recently translated these findings to patients with OSA by observing that females with OSA have a greater proportion of cytotoxic NK cells relative to females with OSA with no such finding in males [76]. Importantly, NK cells participate in the hyperinflammatory state of numerous clinical conditions [77]. Data from a murine model of hyperlipidemia purport a strong pro-atherogenic role for NK cells [78] whereas others report conflicting evidence [79]. These discrepancies have led to more thorough analyses of NK cell subpopulations and are the topic of emergent research in ASCVD [80], although the translatability to humans remains unclear [81]. A secondary finding in our study was that regulatory T cells were more common in females with OSA relative to females without OSA [76]. As regulatory T cells suppress the production of cytokines by other cells and reduce the progression of atherosclerotic lesions [82], this further supports the hypothesis that dysregulation of the immune system may contribute to sex differences in OSA-associated ASCVD risk potentially through immune-derived inflammation.

Vascular Dysfunction

Vascular endothelial cell function and integrity are key factors in the formation of atherosclerotic lesions. Seminal studies show that patients with OSA have a blunted hemodynamic response to intra-brachial acetylcholine infusion (i.e., worse endothelial function) relative to age- and weight-matched controls [36]. This finding was corroborated by data collected from 30,000 subjects, indicating that peripheral artery function assessed via the reactive hyperemia index was attenuated in patients with OSA [83]. The authors noted that after adjusting for ASCVD risk factors, OSA was indicative of poor endothelial function exclusively in women with OSA. When endothelial function was assessed using the gold-standard flow mediated dilation (FMD) technique, AHI was inversely associated with FMD in women with OSA whereas no relationship was observed in men with OSA [84]. Of note, endothelial function assessed via FMD reflects coronary vascular health [85] and a 1% reduction in FMD is associated with a ~10% increase in the risk of ASCVD events such as myocardial infarction [86]. One potential explanation for these sex differences is the concentration of OSA during REM sleep [52, 53] whereby adverse changes in REM sleep are associated with reductions in FMD [87]. Mechanistically, sympathetic nerve activity (SNA) is greatest during REM sleep [88] and the REM-dominant OSA observed in women may indicate greater autonomic dysfunction [89]. Indeed, increasing SNA acutely impairs FMD independent of change in arterial blood pressure [90] and autonomic blockage with intravenous trimethaphan improves endothelial function in obese patients with hypertension [91]. Collectively, OSA impairs endothelial function to a greater extent in females, compared to males, which may be due to higher levels of SNA.

Future Directions

The notion that OSA increases ASCVD risk to a greater extent in females than males is gaining attention. While this is likely to prompt more investigations into the underlying mechanisms, an important first step would be to increase the number of females being screened for sleep-disordered breathing. This would be especially pertinent for interventional cardiologists who have the ability to directly assess coronary vascular health. The resultant data could be pooled, merged with clinical metadata, then used to develop focused observational studies. Another opportunity is to increase the number of studies specifically investigating sex differences in patients with OSA. These projects would ideally integrate ASCVD risk factors (e.g., lipid levels, inflammation) with physiological measurements (e.g., muscle SNA, FMD) to guide more mechanistic research projects. Although intermittent hypoxia fails to replicate all components of OSA [92], findings from in vitro studies can reflect data collected in humans. However, to truly understand sex differences in OSA-associated ASCVD risk, these experiments may need to incorporate hypercapnia and a hyperadrenergic exposure. Finally, there is consistent under-recruitment of women in randomized controlled trials. Increased enrollment of women in these studies would be critical to more firmly establish any greater risk of OSA in women, and potentially their greater benefit from therapy.

Conclusions

ASCVD is a leading cause of death worldwide and is more common in males than females. Patients with OSA have increased risk of developing atherosclerotic lesions and this appears to be more impactful in females than males. There are a host of mechanisms that may contribute to sex differences in OSA-associated ASCVD risk that likely originate with etiological differences such as the nocturnal hypoxic burden. However, much more work is needed to comprehensively understanding the mechanisms of this sexual dimorphism.