Multidimensional Sleep Health: Definitions and Implications for Cardiometabolic Health: A Scientific Statement From the American Heart Association

Abstract

Poor sleep health is associated with cardiometabolic disease and related risk factors, including heart disease, stroke, elevated blood pressure and lipid levels, inflammation, glucose intolerance, obesity, physical inactivity, poor diet, unhealthy substance use, poor mental health, and increased all-cause and cardiovascular mortality, and is associated with social determinants of cardiovascular health and health disparities. Therefore, sleep duration has been recognized by the American Heart Association as one of Life’s Essential 8. Although chronic sleep duration is the sole metric used in Life’s Essential 8, sleep health represents a multidimensional construct. This scientific statement outlines the concept of multidimensional sleep health (sleep duration, continuity, timing, regularity, sleep-related daytime functioning, architecture, and absence of sleep disorders) as it applies to cardiometabolic health. Considerations of how these dimensions are related to cardiometabolic health and patterned by sociodemographic status are explained, and knowledge gaps are highlighted. Additional data are needed to understand better how these various dimensions of sleep should be assessed and how interventions targeting sleep health in clinical and community settings can be leveraged to improve health.

In 2016, the American Heart Association published a scientific statement highlighting evidence and gaps in knowledge related to the associations between sleep duration and quality and cardiometabolic health (CMH).¹ Strong evidence from epidemiologic studies supported associations between short sleep duration and incidence of obesity, type 2 diabetes (T2D), hypertension, coronary heart disease (CHD), and stroke, and between long sleep duration and incidence of obesity, T2D, CHD, stroke, and total cardiovascular disease (CVD).¹ There was also some indication of causality from sleep restriction intervention studies, but data were limited and most related to the risk of obesity.¹

Since then, additional evidence associating and causally linking sleep to CMH has accumulated. As a result, in 2022, the American Heart Association added sleep duration as an eighth metric of CMH, elevating Life’s Simple 7² to Life’s Essential 8 (LE8).³ Although the writing group acknowledged that sleep health was multidimensional, with various facets relevant to CMH,³ only sleep duration was considered for inclusion in LE8 because of a scarcity of data and ambiguity around assessing other sleep metrics.

LE8 illustrates that CMH reflects a broader construct than merely the absence of disease. In that same vein, the purpose of this scientific statement is to introduce sleep health as a positive, multifaceted model that expands beyond the absence of known sleep disorders or short sleep duration. This is highly relevant because most US adults report at least some sleep difficulties.⁴ Our goals are to explain multidimensional sleep health, expose the determinants of multidimensional sleep health and disparities in sleep health in the general population, and evaluate the role of multidimensional sleep health in the context of CMH. We chose sleep metrics that are readily available in the literature, measurable in clinical settings, and relevant to sleep and CMH. We suggest methods to assess multidimensional sleep health in various settings and integrate those measures in CVD risk calculations. We propose testing interventions to promote multidimensional sleep health as a means of reducing the risk of CVD and improving healthy life span.

DEFINING MULTIDIMENSIONAL SLEEP HEALTH

Multidimensional sleep health represents a “multidimensional pattern of sleep–wakefulness, adapted to individual, social, and environmental demands, that promotes physical and mental well-being.”⁵ Conceptualizing sleep as multidimensional emphasizes that sleep is not a unitary physiologic experience, and multiple sleep dimensions affect functioning in various ways. Similar to other lifestyle behaviors, no single aspect of sleep health fully captures its physiologic and phenomenologic value. The dimensions of sleep duration, continuity, timing, satisfaction, regularity, and sleep-related daytime functioning have emerged as being relevant and can be measured using both objective and self-report measures.

Sleep quality is a term that can be (and has been) used to describe global sleep health and any one of multiple dimensions of sleep health. As a nonspecific term, we restrict its use in the current document to instances where that term is specifically used in a citation and cannot be better clarified otherwise.

Sleep duration refers to the amount of sleep obtained either per night or per 24-hour period captured from retrospective or prospective self-report (questionnaires and diaries, respectively) or objective estimation (eg, polysomnography, actigraphy). This dimension is included in the LE8.³ It is best represented in the literature regarding its association with outcomes of interest, can be represented by a single number, and can be compared with established guidelines.⁶

Sleep continuity reflects the ability to initiate and maintain sleep. This is typically summarized as sleep efficiency (proportion of time in bed spent sleeping), which is calculated on the basis of other sleep continuity measures, including the time to fall asleep (sleep latency), number of awakenings, time spent awake after sleep onset, and unplanned early awakening.

Sleep timing refers to the clock time during the 24-hour day when someone sleeps (bedtime to wake time). Studies evaluating sleep timing and CMH often attempt to establish “optimal” bedtimes (eg, 10 to 11 PM has been shown to discriminate CVD risk factors among adults⁷). The circadian clock helps regulate sleep so that sleep during the biologic night has greater continuity and results in improved satisfaction compared with sleep during the day.

Sleep satisfaction reflects a person’s reported perception of their sleep experience. It is typically a patient-reported outcome and may represent aspects of sleep health not otherwise assessed in other dimensions or aspects of sleep physiology.

Sleep regularity is the consistency (ie, variability or invariability) of sleep timing or duration. This can be represented in several ways, such as weekday–weekend (work/study days–free days) discrepancy or other computed metrics describing variability in sleep duration or timing variables. Information can be obtained through questionnaires or objectively from wrist actigraphy for more complex assessments. Sleep regularity is different from timing. Timing refers to the occurrence of sleep in the 24-hour day relative to the optimal circadian phase. Regularity refers to the stability of that schedule across days, irrespective of its location in the 24-hour day.

Sleep-related daytime functioning is the degree to which aspects of functioning are preserved during the day, irrespective of nighttime measures, highlighting the importance of sleep health both at night and in the daytime. This is typically conceptualized as being alert, energetic, vigilant, and awake, and lacking daytime sleepiness (increased sleep propensity during the day) or fatigue (tiredness without sleep propensity). Sleep-related daytime functioning can be assessed through self-reports or objective laboratory tests and may also be represented by an absence of deficits in other facets of daytime functioning (eg, reaction time) that may be otherwise affected by nighttime sleep or daytime sleepiness and fatigue.

Sleep architecture refers to sleep stages (eg, slow-wave sleep, rapid eye movement sleep) and physiologic patterns measured by electroencephalographic wave-forms. Sleep stages represent various neurophysiologic functions, and the amount and timing of these are associated with physiologic and behavioral outcomes.

An often-used framework to conceptualize multidimensional sleep health is the RU_SATED model, defined on the basis of domains of regularity, satisfaction, alertness (sleep-related daytime functioning), timing, efficiency (sleep continuity), and duration. However, this model does not include the absence of sleep disorders or consider sleep architecture. Furthermore, no consensus exists on which sleep measure best represents each domain, nor on the thresholds considered optimal. Table 1 lists common sleep health dimensions and how they have been operationalized in research.

Table 1.

Common Sleep Dimensions Included as Components of Multidimensional Sleep Health

Dimension	Description	Common approaches to operationalization	Commonly used “optimal” cut points
Regularity/rhythmicity	How stable one’s sleep–wake patterns are over time	Within-person variability (eg, SD) of sleep timing (eg, wake time, midpoint) or duration measures (ie, total sleep time; A, D, Q) Rest–activity–rhythm variables (A) Social jet lag (difference in midpoint of sleep between free and nonfree days)	SD ≤60 min
Satisfaction/quality	One’s perception of how well they slept	Sleep quality (D, Q) Rested upon awakening (D, Q) Sleep depth (D, Q)	PSQI global score ≤5
Alertness/sleepiness	One’s ability to maintain an optimal level of vigilance throughout the waking day	ESS (Q) Napping behavior (A, D, Q) MSLT, MWT (P) PVT	ESS score ≤ 10
Timing	The time at which sleep is obtained within the 24-hour day	Bedtime or wake time (A, D, Q) Sleep midpoint (ie, middle of the sleep period; A, D, Q)	Midpoint 2:00–4:00 AM
Efficiency	How easily sleep is initiated and maintained	Sleep efficiency (A, D) Sleep onset latency (A, D, Q) Wake after sleep onset (A, D, Q)	Sleep efficiency ≥85% Sleep onset latency <30 min Wake after sleep onset <30 min
Duration	The total amount of sleep obtained, either at night or across 24 h	Sleep duration (A, D, Q)	7–9 h (self-report) or 6–8 h (A)
Disturbed sleep	The presence and magnitude of disturbed sleep	Sleep disorder diagnosis, such as SDB or CID (Q) Use of sleep medications or devices (D, Q)	No medication or device use No sleep disorder
Sleep architecture	The amount and temporal organization of different sleep stages across a sleep bout	Amount of time (min, or % of total sleep) spent in non-REM or REM sleep stages (P)	None previously used

A indicates actigraphy; AHI, apnea-hypopnea index; CID, chronic insomnia disorder; D, diary; ESS, Epworth Sleepiness Scale; MSLT, multiple sleep latency test; MWT, maintenance of wakefulness test; P, polysomnography; PSQI, Pittsburgh Sleep Quality Index; PVT, Psychomotor Vigilance Test; Q, questionnaire; REM, rapid eye movement; and SDB, sleep-disordered breathing.

Similar to LE8, multidimensional sleep health is typically measured as an aggregation of individual sleep metrics. Composite multidimensional sleep health scores have been created by summing the number of optimal sleep health dimensions one possesses on the basis of clinically relevant or empirically derived cut points. Multidimensional sleep health instruments, including the RU_SATED questionnaire^5,8 and the Sleep Health Index,⁹ capture this information. More sophisticated analytical techniques, such as cluster analysis,¹⁰ that may better capture variability in continuous measures and do not assume a priori cut points have also been used in research. These are more difficult to calculate, specific to the population being studied, and not readily adaptable for a clinical setting. Despite these limitations, incorporating multiple dimensions of sleep health, beyond duration or exclusion of disorders, is an important advancement in understanding the cardiovascular health (CVH) effects of sleep.

ASSESSMENT OF SLEEP HEALTH

Unlike some other aspects of health, sleep cannot be measured directly; all measurements of sleep health are indirect because sleep–wake neurophysiology involves a wide range of structures in the brainstem, midbrain, and other regions. Measurement options include both self-report and objective methods. Although in-laboratory polysomnography represents the gold standard strategy for measuring sleep architecture and sleep apnea, it remains an imperfect tool and not always the optimal strategy, even when available. This is especially relevant for the dimensions of sleep health listed here that depend on comfortable and familiar sleeping conditions and environment. Therefore, sleep health is best assessed using multiple methods.

Polysomnography typically includes electroencephalography (measuring brain activity to determine sleep versus wake and sleep stages), electrooculography (for measuring eye movements, especially for rapid eye movement sleep), electromyography (for measuring muscle tone, especially for rapid eye movement sleep), and other signals, which typically include electrocardiography and sensors for respiratory physiology, and limb movements. This strategy provides high temporal resolution and physiologic accuracy, providing information on sleep architecture, arousals, limb movements during sleep, disordered breathing and its origin (central or obstructive), one’s ability to sleep under controlled conditions, and daytime sleep propensity (Multiple Sleep Latency Test for sleepiness), and is diagnostic for several sleep disorders. However, polysomnography has limitations, especially related to a single time point measure in an artificial situation that may distort findings. In addition, polysomnography cannot detect some aspects of sleep that are relevant to specific sleep disorders (eg, insomnia), psychologic wellbeing, or CVH, such as perceived effects on daytime functioning, experienced satisfaction, or difficulty falling or staying asleep (ie, subjective continuity). Some newer strategies include portable wearable devices that include electroencephalographic sensors; these may also have value for recording physiologic sleep outside of a laboratory but are not as accurate as in-laboratory measures because problems during the recording cannot be fixed synchronously and may lead to data loss.

Other objective measurements of nighttime sleep include sensors typically worn on the wrist, finger, or arm that rely on movement quantification (eg, accelerometry) and may also include other signals (eg, photoplethysmography, skin temperature). Many are well-validated for sleep–wake detection.¹¹ However, they tend to overestimate sleep compared with polysomnography and underestimate sleep compared with self-report. Adding photoplethysmography to a movement signal has some utility but is not equivalent to polysomnography. Accelerometry, preferably combined with sleep diaries (see the following), is ideal for assessing the sleep–wake cycle (ie, circadian timing and regularity) and its disturbances (ie, circadian rhythm sleep–wake disorders). Objective measures of alertness, beyond the Multiple Sleep Latency Test, include the psychomotor vigilance test, which measures vigilant attention and, thus, the cognitive aspects of daytime fatigue.

Sleep diaries are prospective, self-reported sleep assessments that include daily recording of nighttime sleep timing and continuity variables (eg, time to bed, sleep latency, wake after sleep onset, time awake) and daytime alertness (eg, frequency, timing, and duration of naps or dozing); they may also include self-reported duration and satisfaction. Standard approaches for collecting sleep diary data exist.¹² This strategy is the gold standard approach for assessing insomnia and sleep difficulties, although a lack of patient adherence is a limitation.¹³

Retrospective self-reported sleep assessments may also be useful. Validated questionnaires exist for many domains of sleep health, particularly multi-item self-reports of satisfaction with sleep, sleep continuity, sleep timing and regularity, insomnia, circadian timing, sleepiness, and fatigue. However, observational cohort studies often use single-item measures. Although suboptimal, single survey items have proven useful for demonstrating the importance of sleep health relative to outcomes of interest. Some of these single items may present with unique validity problems, however, clouding interpretation.

SOCIAL DETERMINANTS OF SLEEP HEALTH

Sleep health is not equitable. Health disparities, including avoidable, unjust, and unfair conditions, often adversely affect single or multiple attributes of sleep health.¹⁴ Individuals from historically underrepresented racial or ethnic groups are more likely to have short sleep duration, worse sleep continuity, less satisfaction with sleep, later and more irregular sleep, and high sleepiness in comparison with non-Hispanic White individuals,¹⁵ and people from underrepresented racial or ethnic groups may also be more likely to have undiagnosed sleep disorders. These disparities are observed across the life span and persist over time,^16,17 with Black adults demonstrating the greatest global sleep health disparity.¹⁸ The difference in sleep between people from or not from underrepresented racial or ethnic groups varies across studies, ranging between 1- and 2-hour differences in sleep duration between Black and non-Hispanic White individuals.^17,19,20 Research on sleep disparities is growing, but failure to include people from underrepresented racial or ethnic groups in research limits our understanding of the role of multidimensional sleep health on CMH, as well as ways to promote multidimensional sleep health and treat poor sleep among these populations.

Disparities in multidimensional sleep health also exist by socioeconomic position, including income, education, work status and position, and financial assets. These disparities are also observed to be associated with risk of chronic diseases. Although there have been conflicting findings regarding their associations with sleep health, there is growing consensus that higher socioeconomic status is related to better opportunities for optimal multidimensional sleep health.^21,22 A recent review of 336 studies²² revealed consistent associations between multiple characteristics of socioeconomic position and indicators of sleep continuity (eg, latency, awakenings, efficiency), symptoms of sleep disorders (eg, insomnia, sleep-disordered breathing), and sleep satisfaction. Associations with sleep duration were less consistent, possibly because of differences in work schedules and social, occupational, and familial demands. Excessive daytime sleepiness (EDS), a symptom of sleep deficiency and a contributor to poor daytime function, was also associated with social position.²³ Some of these factors likely interact. Race, ethnicity, sex, gender, and socioeconomic position represent intersecting identities that explain complex sleep disparities. A limited literature suggests that some measures of better socioeconomic position (eg, professional occupation, education) are associated with worse sleep health among Black and Hispanic or Latinx individuals, but better sleep health for White individuals.^20,24,25 These findings suggest the need to better understand the multilevel social and environmental factors (eg, racism, intersectional discrimination, home and neighborhood characteristics, such as light, air, and noise pollution, economic disadvantage, social cohesion, safety) that contribute to disparities in multidimensional sleep health,¹⁵ which may explain these apparently contradictory findings. However, despite many studies focused on the contributions of socioeconomic position to multidimensional sleep health, there remains considerable need to improve our understanding of the underlying mechanisms.

Social-Ecological Model of Sleep Health

The social-ecological model addresses multiple levels and factors that contribute to health outcomes and health disparities.²⁶ This framework posits that individuals are nested in broader environments (eg, household, community, neighborhood, societal) that shape sleep health behaviors and patterns.^27,28 Figure illustrates the role of individual-level, social-level, and societal-level factors based on those originally described in the social ecological model.²⁹ The individual level refers to characteristics that do not extend outside the individual (eg, that person’s demographic characteristics, physical characteristics, health behaviors, mental processes). The social level refers to the social constructs that exist outside of the individual but include them (eg, their occupation, family, culture, socioeconomic status, neighborhood). The societal level refers to the factors that exist at an additional layer of abstraction, representing issues that transcend individual social-level factors (eg, factors that affect all workplaces, policies, laws). This model describes these factors in multidimensional sleep health and downstream outcomes.²⁷ Some social and environmental factors (eg, neighborhood socioeconomic status) are partially attributable to structural racism and discrimination and contribute to multidimensional sleep health disparities³⁰ and CVD risk.³¹ Thus, it is imperative to address upstream or structural factors and the consequential social determinants on multidimensional sleep health.

Figure. Social ecological model in the context of multidimensional sleep health.

Sleep health encompasses several dimensions, including duration, timing, regularity, daytime functioning, satisfaction, and continuity, which represent separate but related aspects of sleep health. These dimensions are determined by upstream factors (individual, social, and societal) representing the social ecological model of sleep health. Individual-level factors proximally determine sleep, including demographic characteristics, thoughts, feelings, behaviors, and genetics. These are embedded within social-level factors that include home, family, work, neighborhood, social networks, and other factors that exist outside of the individual but include the individual. Social factors create the context within which individual-level factors operate and are influenced by societal-level factors, including issues such as globalization, policy, racism, economics, and geography. Many of these influence climate change, which affect individual and social factors. Downstream of sleep health are the conceptualized health domains affected by multidimensional sleep health. These include interrelated factors such as cardiovascular fitness, immunologic response, metabolic regulation, behavioral health, neuronal function (including cognition, sleep architecture, and sleep disturbances), and emotional wellbeing. Taken together, these factors influence long-term outcomes related to morbidity and mortality, reflected by a healthy life span.

According to this model, the proximal, individual-level contributors to sleep health exist solely within the individual and represent their characteristics (eg, sociodemographics, genetics) and internal processes (eg, health, choices, attitudes, priorities). These factors most directly influence an individual’s day-to-day sleep-related behaviors and physiology. However, these factors are embedded within a social level. Social factors exist outside of the individual but subsume them and provide context for biology, beliefs, and behaviors. These social-level factors are further embedded within societal factors, such as technology, globalization, 24/7 society, racism, and public policy, and other factors that represent more distal but still relevant contributors to behavior. For example, societal changes in technology use and increased globalization have led to changes in how families and occupations manage their communications and business online, which affects individuals struggling to put down screens at night. This can then lead to problematic technology use, which impairs sleep health.

MULTIDIMENSIONAL SLEEP HEALTH AND CMH

The evidence associating short (<7 hours) or long (>9 hours) sleep with obesity, hypertension, CVD, CHD, T2D, and stroke is strong,¹ and it has continued to accumulate in the past decade with the addition of novel outcomes. Evidence on the association of the most widely studied sleep health dimensions with CMH from recent systematic reviews and meta-analyses since the previous American Heart Association scientific statement is summarized in the following. Topics for which evidence was addressed by previous statements, such as obesity¹ and obstructive sleep apnea,³² or for which evidence is limited, such as sleep architecture, medications, or devices, were not included in the summary.

Sleep Duration and CMH

Evidence of the contributions of short or long sleep duration to CMH is strong (Table 2).¹

Table 2.

Association of Common Sleep Dimensions With Cardiovascular Health Outcomes

Sleep dimensions, study types, authors (y)	Studies (participants, n)	Sleep dimension	Findings
Duration of sleep
Meta-analyses
Hua et al (2021)33	27 cross-sectional (340 492)	Short	OR, 1.06 (95% CI, 1.01–1.11), prevalent MetS
		Long	OR, 1.11 (95% CI, 1.04–1.17), prevalent MetS
	9 prospective (235 895)	Short	RR, 1.15 (95% CI, 1.05–1.25), incident MetS
		Long	RR, 1.02 (95% CI, 0.85–1.18), incident MetS
Vats et al (2024)34	8 cohort (2 063 045)	Short	HR, 1.18 (95% CI, 1.03–1.34), incident AFib
		Long	HR, 1.03 (95% CI, 0.92–1.14), incident AFib
Saz-Lara et al (2022)35	8 cross-sectional	Short	Not significantly associated with arterial stiffness
	6 longitudinal (97 837)	Long	ES, 0.21 (95% CI, 0.06–0.36), arterial stiffness
Kwok et al (2018)36	74 cohort (3 340 684)	Short	Not associated with risk of all-cause mortality
		9 h	RR, 1.14 (95% CI, 1.05–1.25), all-cause mortality
		10 h	RR, 1.30 (95% CI, 1.19–1.42), all-cause mortality
		11 h	RR, 1.47 (95% CI, 1.33–1.64), all-cause mortality
He et al (2017)37	16 prospective (528 653)	4 h	RR, 1.17 (95% CI, 0.99–1.38), incident stroke
		5 h	RR, 1.17 (95% CI, 1.00–1.37), incident stroke
		6 h	RR, 1.10 (95% CI, 1.00–1.21), incident stroke
		Short	RR, 1.35 (95% CI, 1.88–1.52), incident stroke (EU)
		Short	RR, 1.20 (95% CI, 0.99–1.46), incident stroke (US)
		8 h	RR, 1.17 (95% CI, 1.07–1.28), incident stroke
		9 h	RR, 1.45 (95% CI, 1.23–1.70), incident stroke
		10 h	RR, 1.64 (95% CI, 1.14–1.92), incident stroke
Huang et al (2022)38	71 cohort (3 800 000)	4.3–10.3 h	Nighttime sleep duration associated with the lowest risk of CVD at 7.5 h (nonlinear P<0.0001)
		5.09–9.7 h	Sleep length associated with a lower risk of stroke reverted at 7.5 h per night (nonlinear P =0.05)
Kruisbrink et al (2017)39	6 cohort (30 033)	Short	RR, 1.01 (95% CI, 0.93–1.10), incident dyslipidemia
		Long	RR, 0.98 (95% CI, 0.87–1.10), incident dyslipidemia
Systematic reviews
Forshaw et al (2022)40	17	Short	Moderately associated with nondipping BP
		Long	Not associated with nondipping BP
Utsumi et al (2022)41	13	Short; long	Inconsistent association with atherosclerosis and endothelial dysfunction related to sleep measurement
Sleep quality
Meta-analyses
Lo et al (2018)42	8 (36 971)	Poor	OR, 1.48 (95% CI, 1.13–1.95), prevalent HBP
Saz-Lara et al (2022)35	8 cross-sectional; 6 longitudinal (97 837)	Poor	ES, 0.13 (95% CI, 0.04–0.21), arterial stiffness
Kwok et al (2018)36	74 cohort (3 340 684)	Poor	RR, 1.44 (95% CI, 1.09–1.90), incident CHD; not associated with all-cause or CVD mortality
Chokesuwattanaskul et al (2018)43	5 cohort (14 179 909)	Insomnia	OR, 1.30 (95% CI, 1.26–1.35), prevalent AFib
		Awakenings	OR, 1.36 (95% CI, 1.13–1.63), prevalent AFib
Dean et al (2023)44	9 cohort (1 184 256)	Insomnia	RR, 1.69 (95% CI, 1.41–2.02), incident MI
Dai et al (2024)45	6 cross-sectional (5914); 2 longitudinal (n=1963)	Insomnia + short	OR, 2.67 (95% CI, 1.45–4.90), prevalent HBP; RR, 1.95 (95% CI, 1.19–3.20), incident HBP
Systematic reviews
Aziz et al (2017)46	13	Poor	Associated with endothelial dysfunction and arterial stiffness, inconsistently with atherosclerosis
Forshaw et al (2022)40	17	Poor	Moderately associated with nondipping BP
Sleep timing
Systematic review
Chaput et al (2020)47	41 (92 340)	Later onset	Associated with higher adiposity (low quality of evidence) and biomarkers of cardiovascular risk (moderate quality of evidence)
Original studies
Tse et al (2021)48	1 (136 652)	Bedtime ≥ midnight	OR, 1.2 (95% CI, 1.12–1.29), general obesity; OR, 1.2 (95% CI, 1.12–1.28), abdominal obesity
Knutson et al (2017)49	1 (13 429)	Later wake time	Associated with fasting glucose ≈3% increase per h later; P<0.05
		Later midpoint	Associated with insulin resistance of ≈1.5% increase per h later; P<0.05
Abbott et al (2019)50	1 (2156)	Delayed midpoint by 1 h	Associated with systolic BP +0.73 mm Hg (95% CI, 0.30–1.16), P<0.01; associated with diastolic BP +0.53 mm Hg (95% CI, 0.17–0.90), P<0.01
Fan et al (2021)51	1 (4576)	Weekday bedtime > midnight	HR, 1.63 (95% CI, 1.09–2.43), incident myocardial infarction
Sleep regularity
Meta-analysis
Arab et al (2024)52	43 (231 648)	Social jet lag	OR, 1.2 (95% CI, 1.02–1.14), overweight or obesity
Systematic review: panel consensus statement
Sletten et al (2023)53	63	Variability in timing	Associated with increased risk of cardiovascular disease, hypertension, inflammation, and obesity
		Catch-up (extending sleep by 1–2 h on nonwork days)	Associated with positive cardiometabolic health outcomes
Systematic reviews
Forshaw et al (2022)40	17	Variability in timing	Associated with nondipping BP in both shift workers and nonshift workers
Zhu et al (2022)54	63	Variability in timing and duration	Associated with weight gain, obesity, CVD, and metabolic syndrome; possibly associated with HbA1c in people with underlying type 1 diabetes but not in healthy populations; inconsistent association with risk of hypertension and diabetes; no association with insulin resistance
Original studies
Windred et al (2024)55	1 (60 977)	Consistent sleep–wake timing	22%–57% lower risk of cardiovascular mortality (P<0.001 to P=0.048)
Cribb et al (2023)56	1 (88 975)	Variability in timing	HR, 1.88 (95% CI, 1.61–2.21), CVD mortality
Full et al (2023)57	1 (2032)	Irregular duration (SD >120 min)	PR, 1.33 (95% CI, 1.03–1.71), high coronary artery calcium burden; PR, 1.75 (95% CI, 1.03–2.95), abnormal ankle‐brachial index
Sleepiness
Meta-analysis
Wang et al (2020)58	17 cohort (153 909)	EDS	RR, 1.28 (95% CI, 1.09–1.50), incident CVD; RR, 1.28 (95% CI, 1.12–1.46), incident CHD; RR, 1.52 (95% CI, 1.10–2.12), incident stroke; RR, 1.23 (95% CI, 1.13–1.33), all-cause mortality; RR, 1.47 (95% CI, 1.09–1.98), CVD mortality
Original studies
Bixler et al (2005)59	1 (1741)	EDS	OR, 3.3 (95% CI, 2.9–3.8), prevalent depression; OR, 1.45 (ES, 4.3), prevalent obesity; OR, 1.9 (95% CI, 1.6–2.4), prevalent T2D; OR, 1.53 (ES, 1.9), prevalent smoking
Fernandez-Mendoza et al (2015)60	1 (1741)	Incident EDS	OR, 2.0 (95% CI, 1.2–3.3), T2D; OR, 1.8 (95% CI, 1.2–2.8), obesity
		Persistent EDS	OR, 2.8 (95% CI, 1.8–4.3), weight gain
		Remitted EDS	OR, 2.7 (95% CI, 1.4–5.2), weight gain; OR, 2.5 (95% CI, 1.2–5.1), weight loss
Wang et al (201962	1 (452 071)	EDS	rg, 0.1694 (SE, 0.0358), obesity; rg, 0.1527 (SE, 0.0335), CHD
Meta-analysis
Sun et al (2022)61	14 (720 523)	Daytime napping	RR, 1.29 (95% CI, 1.20–1.40), incident obesity; RR, 1.20 (95% CI, 1.05–1.37), incident T2D; RR, 1.14 (95% CI, 1.05–1.23), incident dyslipidemia; RR, 1.17 (95% CI, 1.07–1.28), incident MetS; RR, 1.14 (95% CI, 1.08–1.21), incident CVD; RR, 1.18 (95% CI, 1.07–1.31), all-cause mortality; RR, 1.24 (95% CI, 1.04–1.47), CVD mortality

AFib indicates atrial fibrillation; BP, blood pressure; CHD, coronary heart disease; CVD, cardiovascular disease; EDS, excessive daytime sleepiness; ES, effect size; HBP, high blood pressure; HR, hazard ratio; MetS, metabolic syndrome; MI, myocardial infarction; OR, odds ratio; PR, prevalence ratio; rg, genetic correlation from genome-wide association study; RR, relative risk; SD, standard deviation; SE, standard error; and T2D, type 2 diabetes.

Six meta-analyses have reported associations between short sleep and metabolic syndrome,³³ atrial fibrillation,³⁴ stroke,⁶³ and nondipping blood pressure,⁴⁰ and between long sleep and metabolic syndrome,³³ arterial stiffness,³⁵ stroke,⁶³ and all-cause, stroke, or CVD mortality.^36,37 However, inconsistent associations have been found between short sleep and arterial stiffness,³⁵ dyslipidemia,³⁹ and all-cause, stroke, or CVD mortality,³⁶ and between long sleep and metabolic syndrome,³³ atrial fibrillation,³⁴ and dyslipidemia.³⁹ These contrasting findings may be explained by the definitions used for either short or long sleep and the measurement method.^38,41 The largest meta-analysis to date has shown that the risk of CVD reaches a nadir at 7.5 hours,³⁸ with wide confidence intervals. A systematic review of 13 studies showed an inconsistent association between short and long sleep with atherosclerosis and endothelial dysfunction and indicated that the source of variability was self-report versus objective sleep measures.⁴¹ Thus, both category thresholds and measurement methods are important when predicting CMH outcomes from sleep duration as a standalone dimension of sleep.

Sleep Continuity Disturbance and Satisfaction and CMH

Some recent meta-analyses and systematic reviews (Table 2) demonstrated that other elements of poor sleep health (eg, reduced satisfaction with sleep, impaired sleep continuity) are associated with hypertension,⁴² arterial stiffness,^35,46 CHD,³⁶ nondipping blood pressure,⁴⁰ and endothelial dysfunction,⁴⁶ but not with atherosclerosis,⁴⁶ dyslipidemia,³⁹ or all-cause or CVD mortality.^36,37 Associations between sleep continuity complaints (reported as insomnia symptoms) with CVH outcomes have also been observed, specifically with atrial fibrillation⁴³ and myocardial infarction.⁴⁴ Consistent with the concept of multidimensional sleep health, a recent meta-analysis supported a synergistic association between self-reported sleep continuity complaints (eg, insomnia) and objectively measured short sleep, extending beyond each individual sleep dimension, with hypertension.⁴⁵ Disruption of sleep continuity influences distribution of sleep across various stages throughout the night. One meta-analysis noted that disrupting slow-wave sleep led to greater insulin resistance compared with undisturbed sleep.⁶⁴

Sleep Timing and CMH

Inappropriate timing of sleep (Table 1) is likely to be associated with CVD risk, but the evidence is limited. A systematic review (Table 2) showed that later sleep timing was associated with higher adiposity,⁴⁷ although the quality of evidence was low. Later sleep timing was associated with greater levels of CVD risk overall, and the quality of evidence was rated as moderate.⁴⁷ Most studies (63%) used self-reported measures of sleep timing, creating high heterogeneity across studies, and reported linear associations only, making it difficult to quantify what constitutes “late timing.” However, having a bedtime of midnight or later, compared with earlier, has been associated with higher odds of general and abdominal obesity,⁴⁸ and with incident myocardial infarction.⁵¹ In cohort studies, dose–response relationships were found between later wake time and fasting glucose levels, and later midpoint of sleep and insulin resistance⁴⁹ and systolic and diastolic blood pressure.⁵⁰ A study using data from the Korean National Health and Nutrition Surveys summarized better CMH with bedtime cut points of 9 PM to 12:30 AM for men and 10 to 11 PM for women.⁷ Most of these studies used sleep timing metrics as surrogate markers of circadian alignment, assuming that an individual’s sleep timing is well aligned with their endogenous circadian phase (circadian clock) when socioecologic factors allow.

Sleep Regularity and CMH

Irregular sleep duration and timing have been associated with increased CVD risk, independent of sleep quantity and quality. For example, social jet lag (difference in midpoint of sleep between weekdays and weekend days) has been associated with 1.20-fold greater odds of overweight or obesity (Table 2),⁵² and day-to-day variability in sleep timing has been associated with greater risk of CVD, hypertension, inflammation, obesity,⁵³ and nondipping blood pressure,⁴⁰ but not with insulin resistance.⁵⁴ The majority of studies on sleep variability have been observational, with variability in measures used across studies.⁵³ In large-scale observational studies using objective measures of sleep regularity, greater consistency of sleep–wake timing was associated with 22% to 57% lower risk of cardiovascular mortality,⁵⁵ whereas higher sleep irregularity was associated with nearly 2-fold increased risk of CVD mortality.⁵⁶ Sleep irregularity was also associated with subclinical atherosclerosis in the Multi-Ethnic Study of Atherosclerosis Sleep Ancillary Study, with participants with greater sleep duration irregularity (SD >120 versus ≤60 minutes) being more likely to have high coronary artery calcium levels and abnormal ankle‐brachial index.⁵⁷ Data from the UK Biobank study revealed that irregular sleep confers increased risk of T2D despite meeting recommendations for adequate sleep duration, and individuals who fail to meet sleep duration recommendations and have highly irregular sleep have the greatest risk.⁶⁵

Although seemingly contradictory to findings showing associations between greater sleep regularity and greater CVH, “catch up” sleep has also been shown to be associated with favorable CVD outcomes in the few studies available to date.⁵³ Although it would be better to achieve regular bedtimes and wake times 7 days a week, as a countermeasure that may be beneficial for CVH outcomes, the National Sleep Foundation recommends extending sleep by up to 1 to 2 hours on nonwork days in individuals who cannot meet sleep duration recommendations on work days.⁵³ Some individuals may choose to compensate for insufficient sleep at night with daytime napping. This may not be the best solution for CVH; a meta-analysis showed that daytime napping is associated with obesity, T2D, dyslipidemia, metabolic syndrome, CVD, and all-cause or CVD mortality.⁶¹ However, daytime naps of <30 minutes are not significantly associated with most CVH outcomes among adults younger than 60 years.⁶¹

Daytime Sleepiness and CMH

EDS has been significantly associated with CVD, CHD, stroke, and all-cause or CVD mortality (Table 2).⁵⁸ There seems to be bidirectionality to the associations between EDS and CVD risk⁶²; for example, EDS has been associated with obesity, T2D, depression, and smoking,⁵⁹ and weight gain, T2D, and depression have been associated with the incidence of EDS.⁶⁰ Weight loss has been associated with remission of EDS,⁶⁰ highlighting benefits of lifestyle modifications on multidimensional sleep health.

CONCLUSION, KNOWLEDGE GAPS, AND FUTURE DIRECTIONS

Research conducted over the past several decades led to the recognition of sleep duration as an important contributor to CMH. We now recognize sleep as a multidimensional health factor and have noted here various metrics beyond sleep duration that contribute to health. Variability in findings of associations between these sleep metrics and CMH, due to heterogeneity among studies (eg, type of measurements, definitions of sleep metrics or cut points), hinders progress in this field. Research thus far has contributed to the understanding that poor sleep health (eg, short sleep duration, irregular sleep schedules, otherwise unadvisable sleep behaviors) contribute to adverse cardiovascular outcomes. However, guidance on multidimensional sleep health cannot be made without evidence that improving sleep health leads to better CMH. Some studies have focused on single aspects of sleep health, mainly lengthening sleep duration, but none has been comprehensive. Nonetheless, increasing sleep duration generally has a beneficial effect on CMH factors.⁶⁶

To start, actionable recommendations on multidimensional sleep health cannot be disclosed without firm and reproducible findings. This will require a concerted effort from various disciplines, including but not restricted to sleep, cardiology, endocrinology, gastroenterology, nephrology, and hepatology, within the scientific community, and support from funding agencies for research that will provide such information. We therefore call for population studies to evaluate associations of multidimensional sleep health, using tools and definitions provided herein, with CMH, and to evaluate empirically developed cut points for multidimensional sleep health, each of its dimensions, and each method of measurement in predicting CMH. The concept of CMH, as an outcome in relation to sleep, would be enhanced by evaluations of immunologic responses, behavioral and cognitive health, emotional well-being, and cardiovascular fitness that we were not able to address fully herein. In addition, studies should examine joint or additive effects across sleep dimensions to identify which specific combinations confer greatest risk (eg, insomnia complaints with short sleep duration, delayed timing with sleep irregularity). Physiologic studies should also provide more robust evidence on the role of specific sleep architecture metrics on elevating cardiovascular risk beyond other sleep dimensions. We also call for clinical studies to test causation through interventions to improve sleep health as a means of enhancing CMH as well as for basic research to evaluate mechanisms through appropriate animal models. In the meantime, clinicians are encouraged to discuss sleep using the multidimensional sleep health framework presented herein with their patients to gather a more comprehensive assessment of sleep health, beyond traditional clinical history and sleep disorders screening, that can be leveraged to improve their patients’ quality of life. Open-ended questions such as “How is your sleep?” would provide opportunities for patients to disclose information on dimensions of sleep health described herein. These data should be included the patient’s chart so that all health care professionals involved in the patient’s care are aware of their sleep health, and can evaluate the contributions of sleep health to the patient’s condition as well as the effects of the patient’s health condition on their sleep health. This can then lead to a concerted effort to holistic treatment and, if needed, escalation to different levels of care and inclusion of appropriate specialists.

This scientific statement is aimed to bring attention to the multifaceted nature of sleep and raise awareness of the many sleep dimensions relevant to CMH in adults. However, these same sleep dimensions are highly relevant to youth,^67,68 and should be considered in younger age groups as they form the foundation for healthy life span. We urge the medical and scientific communities to consider evaluating multiple facets of sleep health (in addition to sleep disorders), such as sleep satisfaction, continuity, timing, regularity, and architecture, as well as daytime alertness, sleepiness, and fatigue, across the life span when evaluating CMH.⁶⁹ The American Heart Association guidelines in LE8 focus on sleep duration because it has been studied extensively. This is due, in part, to the ease with which sleep duration can be assessed, using both self-report and objective measures. Other metrics will need the same concerted effort to reach the level needed for inclusion into health guidelines. Therefore, future research gathering information from a wide range of sources, including devices and questionnaires, suggested in this report and as appropriate for the research question and setting, on multiple dimensions of sleep will provide the most comprehensive information that will serve to augment our knowledge and guide public health recommendations. In the meantime, simple tools that could be readily implemented in clinical settings, such as RU_SATED, could further guide personalized health care recommendations.