Sex differences in sleep and sleep-disordered breathing

Abstract

Purpose of review

There is increasing evidence for relevant sex differences in pathophysiology, symptom presentation and outcomes in obstructive sleep apnoea (OSA). However, research on sex differences and sex-specific phenotypes in sleep-disordered breathing (SDB) is still in its infancy and data on sex differences in other SDB is still very scarce.

Recent findings

While OSA is more common in men than in premenopausal women, the prevalence of OSA doubles postmenopausally and becomes comparable to that of men. Women have a lower collapsibility of the upper airway and a lower arousal threshold. In addition, the rapid eye movement (REM)-apnoea–hypopnoea index (AHI) is typically higher in women than in men, but the non-REM-AHI and thus the total AHI is often lower. Women are often symptomatic at lower AHI and present more frequently with symptoms of sleep fragmentation and poor sleep quality. Both certain forms of OSA (e.g. REM-OSA) and certain phenotypes (e.g. COMISA) are more common in women. Men have a higher risk of high loop gain central sleep apnoea.

Summary

For a better understanding of sex-typical phenotypes with the aim of a more targeted treatment approach of SDB, adequately powered studies on sex differences in SDB should be conducted.

INTRODUCTION

Looking at the different groups of sleep disorders – particularly insomnia, sleep-related breathing disorders, and sleep-related movement disorders – there are consistent findings on sex differences for these sleep disorders in terms of prevalence, symptom presentation and, in some cases, treatment outcomes. Increasing attention is being paid to sex differences in sleep-disordered breathing (SDB), particularly obstructive sleep apnoea (OSA), central sleep apnoea (CSA) syndromes, and obesity-related hypoventilation. This difference is partly explained by differences in upper airway collapsibility, respiratory control and comorbidities, but for the most part, it is incompletely understood and often insufficiently systematically investigated, particularly in SDB other than OSA. In addition, there are sex differences in the consequences of sleep disorders, particularly the cardiovascular and metabolic consequences. However, women are generally underrepresented in most randomized controlled intervention studies on SDB, which makes it difficult to assess potential differences in treatment outcomes. This review briefly highlights known sex differences in sleep and control of ventilation, and then discusses the evidence base and recent findings on sex differences in SDB. 

Box 1.

no caption available

GENDER DIFFERENCES IN SLEEP AND EPIDEMIOLOGY OF SLEEP DISORDERS

There are differences between the sexes in sleep-regulating mechanisms and sleep-regulating hormones, and in women, the latter also fluctuate depending on the menstrual cycle and phase of life. Premenopausal women generally have a higher proportion of slow-wave sleep and spindle density than men, and progesterone promotes slow-wave sleep, but overall the differences in sleep stages between sexes are marginal [1,2]. However, perceived sleep quality is poorer in adult women than in men at all ages [3]. Melatonin secretion occurs slightly earlier in women and men are more likely to have a later chronotype [4]. There are consistent findings on sex-specific differences in the prevalence of sleep disorders, e.g. a higher prevalence of insomnia in women [5,6]. The two groups of sleep disorders that occur more frequently in women are insomnia and restless legs syndrome (RLS). The higher prevalence of RLS in women may be associated to fluctuations in dopamine levels. In contrast, OSA (at least up to the age of 50) and high loop gain CSA are more common in men.

ROLE OF MENOPAUSE IN SLEEP

In addition to the underrepresentation of women in studies of sleep disorders, epidemiologic studies of sex differences in SDB are complicated by the fact that menopause leads to an increase in both insomnia and SDB, which must be taken into account when assessing the prevalence of sleep disorders in women. During the menopause, sleep quality is particularly poor and some sleep disorders increase. A recent meta-analysis has shown a significant increase in OSA, insomnia and RLS after the menopause [7^▪]. Population-based studies show that sleep problems are linked to the menopause and changes in sex hormones and not simply to age. In addition to the increase in sleep disturbances around menopause, described as one of the symptoms of menopause and reported by 40–60% of perimenopausal women, the prevalence and severity of OSA increases around and after menopause [8]. The prevalence of OSA is about twice as high in men as in women up to the age of 50, but after the menopause, it is almost as high in women as in men [9,10]. In addition, menopause modulates OSA severity. The upper airway of women is less prone to collapse during sleep than that of men due to several mechanisms, including better muscle responsiveness due to female sex hormones. Oestradiol (and progesterone) increases the contractility of upper airway dilators and has a protective effect against upper airway collapse [11]. Progesterone also stimulates ventilation. Testosterone on the other hand is suggested to increase upper airway collapsibility and may increase arousal threshold.

DIFFERENCE IN CONTROL OF VENTILATION BETWEEN SEXES

Hypoxic ventilatory response is higher in men than in women while hypercapnic ventilatory response is comparable between the sexes [12]. Women have a lower hypoxic ventilatory response and start at a lower level during wakefulness, but show better maintenance of the hypoxic response during sleep than men [13]. Men have a greater sleep stage-related depression of hypoxic ventilatory response than women independent of the starting value. In men but not in women, the ventilatory response is related to the metabolic rate and CO₂ production. From these early physiological studies, it can be concluded that men have a higher risk of hypocapnic CSA due to increased hypoxic chemosensitivity. In addition, men have a stronger ventilatory response to arousal and are therefore more likely to exhibit postarousal central apnoeas, e.g. in OSA [14].

SEX DIFFERENCES IN OBSTRUCTIVE SLEEP APNOEA

Pathophysiology and endotypes

Women often have a lower AHI than men, a lower OSA dominance in the supine position and are symptomatic at a lower AHI [15,16]. The collapsibility of the upper airway is lower in women, which can lead to fewer apnoeas and more hypopnoeas or respiratory-related arousals [17]. In addition, women appear to have a lower arousal threshold in non-REM sleep, which also promotes respiratory-related arousals and symptom presentation with symptoms of disturbed sleep (Fig. 1) [18^▪,19^▪]. The proportion of REM-AHI is generally higher in women than in men. REM-OSA is a relevant form of OSA, as REM-related OSA is independently associated with the prevalence and incidence of hypertension [20]. The pathophysiology of OSA differs according to sex, and not only because of the more peripheral fat distribution and smaller neck circumference per body mass index (BMI) in women. A population study showed that women with OSA generally have a higher BMI than men with OSA with the same AHI [16]. The distance between the hyoid bone and the palate is generally larger in men, making upper airway collapse more likely [21]. However, the upper airway length increases with age particularly in women [22]. A comparison of the endotypic characteristics of OSA (airway collapsibility, upper airway muscle reactivity, stability of ventilatory control (loop gain) and arousal threshold) between the sexes showed that women in Non-REM sleep tend to have lower upper airway collapsibility, lower loop gain and possibly also a lower arousal threshold [18^▪,23]. Different prevalence of physiological endotypes may explain part of the sex difference in Non-REM-AHI [24^▪].

FIGURE 1.

Overview on endotypic traits and sleep study characteristics in premenopausal women and in men and changes in women postmenopause.

Symptom presentation and comorbidities

The different symptom presentation and the different spectrum of comorbidities in women with OSA compared to the classic male obese OSA patient with daytime sleepiness and arterial hypertension may lead to underdiagnosis of OSA in women [25,26]. Women are often diagnosed with OSA at an older age and with a higher BMI than men are. In addition to underdiagnosis, being female is associated with a higher risk of undertreatment [27]. Women with OSA are less likely to report snoring than men with OSA. Symptoms include more frequent insomnia, nocturia and fatigue and less frequent choking and daytime sleepiness [28]. Screening questionnaires and scores for assessing the pretest probability for OSA that are used in general practice (e.g. STOP-BANG, NoSAS) also lead to a falsely low assessment of the pretest probability for OSA in women, as they mainly award points for the characteristics of the classic obese male sleepy OSA phenotype. The pretest probability is also often estimated to be lower due to the less classic symptom presentation; the symptoms are incorrectly categorized and therefore less access to diagnostics is granted. In addition, polygraphy is less sensitive in certain types of OSA that are more common in women (e.g. REM-isolated OSA, upper airway resistance syndrome, nondesaturator) and correct diagnosis of these types of OSA requires polysomnography.

Presentation with insomnia symptoms in OSA and COMISA (combined OSA and insomnia) is described more frequently in women and initial data indicate a lower adherence to continuous positive airway pressure (CPAP) and a worse cardiovascular outcome with COMISA compared to OSA alone [29^▪▪,30^▪]. An observational study compared the endotypes between COMISA and OSA without insomnia and found an accumulation of low arousal threshold in the COMISA group – a finding that was not confirmed in another recent study [31^▪,32^▪]. In the case of the COMISA phenotype, multinight sleep studies could be particularly important for the diagnosis of OSA [33^▪].

As blood pressure is a key factor in cardiovascular health and is modulated by sleep, sleep disorders such as OSA, which activate the sympathetic nervous system, are associated with an increased risk of hypertension and a lack of nocturnal blood pressure dipping. Population-based studies have indicated that the risk of adverse cardiovascular effects of sleep fragmentation and sleep deprivation is higher in women [34,35^▪]. However, data on sex differences in relation to hypertension and cardiovascular disease are inconsistent, and this association has mainly been studied in a male population. Women with OSA are more likely to develop comorbidities such as depression and anxiety.

Special phenotypes of obstructive sleep apnoea in women

Compared to men, women have more hypopnoeas than apnoeas and less oxygen desaturation during Non-REM sleep. In addition, women are more likely to have respiratory effort-related arousals and to present with upper airway resistance syndrome (UARS), which is characterized by a repetitive increase in respiratory effort that triggers arousals (respiratory effort-related arousals (RERAs): inspiratory flow limitations that do not fulfil the hypopnoea criterion and lead to an increase in respiratory effort that triggers an arousal), resulting in daytime sleepiness and other symptoms [36]. Women are more likely to suffer from REM-isolated OSA [15]. REM-OSA usually presents as mild or moderate OSA and low arousal threshold is commonly observed in REM-OSA [37]. As women have a lower arousal threshold, this may predispose them to REM-OSA (particularly premenopausal women) although men have a higher collapsibility of the upper airway which is highest during REM in both sexes [37]. Collapsibility may be less important during atonia and other endotypic traits may become more important REM-OSA is associated with an increased risk of hypertension, a nondipping blood pressure pattern and an increased cardiometabolic risk [20,38]. CPAP adherence in REM-OSA is often poor, but CPAP use should be long as REM mainly occurs in the morning. Data on Non-CPAP treatment of REM-OSA is lacking.

Role of sex for treatment outcomes

In a French study of almost half a million CPAP patients, female gender was associated with a slightly higher risk of CPAP discontinuation [39]. Overall, however, there is no evidence of sex-specific differences in CPAP adherence. OSA with insomnia-like symptoms is a more common presentation in women than in men, and even at lower AHI, this type of OSA is associated with a higher burden of cardiovascular comorbidities and lower CPAP adherence compared to sleepy OSA [40]. Men tend to require higher CPAP pressures, which could theoretically be associated with more adverse effects such as mask leakage [41]. Given the lower collapsibility of the upper airway and the lower arousal threshold in women, female-specific algorithms for auto-titrating CPAP to improve treatment effectiveness and tolerance are of interest. An algorithm with higher sensitivity to airflow limitations and slower pressure rise was developed for women and was found to be noninferior in controlling OSA at lower mean pressure [42]. One observational study found greater success of mandibular advancement devices in controlling OSA in women, but there are conflicting results from different studies [43].

With regard to cardiovascular outcomes, there are no data indicating a sex difference in blood pressure response to CPAP [44]. However, as women are generally underrepresented in RCTs on therapeutic outcomes and no RCT was designed to examine the role of sex for hard cardiovascular endpoints (19% women in the SAVE trial [45], 16% women in the ISAAC trial [46], 16% women in the RICCADSA trial [47], although the prevalence of OSA in men and women aged around 60 years and older as in these studies, may not differ that much), no reliable conclusions can be drawn about potential differences in vascular events and mortality. A meta-analysis of these three RCTs with individual patient data found no difference between the sexes in the effect of CPAP on major adverse cardiovascular events, but was also limited by the small number of women (18%) and the lack of specific matching between the sexes for relevant cardiovascular predictors [48^▪▪]. Real-world data have also not shown sex differences in cardiovascular outcomes between CPAP discontinuers and CPAP users [49^▪].

SEX DIFFERENCES IN CENTRAL SLEEP APNOEA

As men have a higher loop gain and thus more ventilatory control instability, they are more prone to high loop gain-CSA such as Cheyne Stoke's respiration in heart failure, treatment-emergent CSA and high-altitude periodic breathing [50,51]. Men are also more susceptible to hypocapnic central apnoeas and thus also to instable breathing due to a higher apnoeic threshold [52].

Observational studies on the prevalence and outcomes of Cheyne Stokes respiration in heart failure mainly included men [53]. A recent prospective observational study in patients with heart failure and sleep apnoea (OSA n = 145 or CSA n = 308), in which women were underrepresented (15.7%), found a higher likelihood of rehospitalization for heart failure or death in women with comparable severity of sleep apnoea between the sexes, but the difference between the sexes only occurred in OSA and not in CSA [54]. However, the available epidemiologic data do not allow a systematic comparison of presentation and outcomes in CSA between men and women. Identifying sex-typical phenotypes of CSA should be a research priority for the development of individualized treatment strategies, which are also desirable for CSA considering the phenotype-based differences in treatment adherence and outcomes observed in Cheyne Stokes respiration in heart failure [55^▪].

SEX DIFFERENCES IN OBESITY HYPOVENTILATION SYNDROME

The differences in the prevalence of obesity hypoventilation syndrome (OHS) between the sexes and the change in association with menopause have not been systematically investigated. However, as with OSA, OHS is underdiagnosed and diagnosis is often delayed in women [56]. Most patients with OHS have moderate to severe OSA. However, OHS with lone hypoventilation (without relevant OSA) is more common in women, which is also reflected in the patient populations of the Pickwick studies, where about 80% of OHS patients without severe OSA were women. The prevalence of OHS increases after menopause. This could be explained both by the increase in upper airway collapsibility and OSA severity and by the decrease in progesterone as a stimulator of ventilation. In clinical practice, higher respiratory drive and lower CO₂ partial pressure during sleep are often observed in younger obese premenopausal women compared to younger men in the same BMI range. In addition, OSA may contribute less to hypoventilation during sleep in women due to lower collapsibility of the upper airway. Furthermore, a more central fat distribution, as commonly observed in obese men, might have more negative effects on dynamic lung volumes and respiratory mechanics at the same BMI compared to a more peripheral distribution in women. However, there is little epidemiologic data on sex differences in prevalence, symptom presentation, poly(somno)graphic features, and treatment outcome in OHS, and there is a need for future research. The comparison of comorbidities and obesity-associated diseases as well as pulmonary hypertension between the sexes would also be of interest. A better understanding of female and male phenotypes of OHS could contribute to a more individualized assessment and treatment strategy.

Table 1 provides an overview of sex differences in SDB.

Table 1.

Typical characteristics of sleep-disordered breathing in women and in men

	Women	Men
Prevalence of OSA	Lower in premenopausal women than in men and in postmenopausal women.	Up to the age of 50 years more than twice as high than in women
Endotypic traits of OSA	Lower arousal threshold compared to men More hypopnoeas than apnoeas More likely to have respiratory-related arousals	Higher collapsibility of the upper airway than women
Specific forms and symptoms of OSA	More likely to have upper airway resistance syndrome than men More likely to have nondesaturator type OSA (hypopnoeas defined with arousal but not with a ≥3% oxygen desaturation) More likely to have REM-isolated or REM-predominant OSA More likely to present with sleep-fragmentation and insomnia than men	More likely to have supine-predominant OSA More likely to present with a history of snoring and witnessed apnoeas
Pathophysiology of CSA	Lower awake respiratory drive in response to hypoxaemia	Higher hypoxic ventilatory drive Higher loop gain.
Prevalence of CSA	Less likely to develop high loop gain CSA compared to men	Higher prevalence of high loop gain CSA (Cheyne Stoke's respiration in heart failure, high altitude periodic breathing, treatment emergent CSA)
Prevalence of OHS	Prevalence increases postmenopause Lone hypoventilation without severe OSA more likely than in men	Likely to develop OHS at a lower body-mass-index than women

CSA, central sleep apnoea; OHS, obesity hypoventilation syndrome; OSA, obstructive sleep apnoea.

IMPLICATIONS FOR CLINICAL PRACTICE AND RESEARCH

Sex-specific aspects and menopausal status should be taken into account when assessing the symptom presentation of SDB and selecting the type of sleep study, as well as when recommending treatment. Currently used screening questionnaires for OSA such as STOP-BANG or NoSAS are less sensitive in women than in men, especially as they rely on symptoms and characteristics of the classic male obese OSA patient [57,58]. As hypopnoea associated with significant desaturations may underestimate the AHI in women, polysomnography should be performed more frequently in women instead of polygraphy. In addition, the REM-AHI should be considered specifically in women, as the total AHI based on the non-REM and REM-AHI could mask the prevalence of OSA of clinical significance in women [18^▪]. A low AHI in connection with OSA symptoms should not prevent a CPAP trial or other OSA therapy, especially in women. However, there are still no specific recommendations or practice guidelines for women or for sex-specific phenotypes of SDB.

Simple investigations into differences between men and women are likely to be as simplistic as asking about differences between AHI-based OSA severity categories and are unlikely to lead to consistent conclusions. Patients need to be more accurately characterized and differences between phenotypes based on multiple markers, including endotypes, need to be sought to support targeted treatment - and some phenotypes are likely to be characterized by female gender [59^▪].

CONCLUSION

The prevalence, underlying pathophysiology and symptoms of SDB, particularly OSA, differ between genders and change after menopause. While men with OSA are more likely to report snoring and daytime sleepiness, women may present with a history of sleep disturbances and insomnia. Due to the different symptom presentation, the lower sensitivity of OSA screening questionnaires and some specific types of OSA more common in women (e.g., REM-OSA, nondesaturator or upper airway resistance syndrome), women with clinically relevant OSA are underdiagnosed and undertreated. The prevalence and severity of OSA increase after the menopause. Men have a higher risk of high-loop CSA such as Cheyne Stokes respiration and treatment-emergent CSA. A better understanding of sex-typical phenotypes with the aim of a more targeted treatment approach to SDB should be a research priority.

Acknowledgements

None.

Financial support and sponsorship

No specific funding.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest

• ▪▪ of outstanding interest