Stress and cardiovascular disease: an update

Abstract

Psychological stress is generally accepted to be associated with an increased risk of cardiovascular disease (CVD), but results have varied in terms of how stress is measured and the strength of the association. Additionally, the mechanisms and potential causal links have remained speculative despite decades of research. The physiological responses to stress are well characterized, but their contribution to the development and progression of CVD has received little attention in empirical studies. Evidence suggests that physiological responses to stress have a fundamental role in the risk of CVD and that haemodynamic, vascular and immune perturbations triggered by stress are especially implicated. Stress response physiology is regulated by the corticolimbic regions of the brain, which have outputs to the autonomic nervous system. Variation in these regulatory pathways might explain interindividual differences in vulnerability to stress. Dynamic perturbations in autonomic, immune and vascular functions are probably also implicated as CVD risk mechanisms of chronic, recurring and cumulative stressful exposures, but more data are needed from prospective studies and from assessments in real-life situations. Psychological assessment remains insufficiently recognized in clinical care and prevention. Although stress-reduction interventions might mitigate perceived stress levels and potentially reduce cardiovascular risk, more data from randomized trials are needed.

Introduction

Psychological stress is a prevalent aspect of human experience. Biological stress responses have a crucial role in human physiology and are vital for adapting to environmental challenges, but stress has also been associated with negative health consequences. Numerous investigations, using diverse definitions of stress, studying different populations and using different research designs, have described detrimental effects of stress on various facets of physical health and, in particular, cardiovascular disease (CVD). Stress is intricately linked to psychological well being, and stressful events often serve as precursors to several psychiatric disorders. Conditions such as depression and post-traumatic stress disorder (PTSD), which are also associated with an increased risk of CVD, often develop in the aftermath of stressful experiences.

Although there is broad consensus on the association between various stressors and CVD, results have varied in terms of effect sizes and the effect of adjusting for potential confounding factors. Additionally, the mechanisms — and therefore the potential causal links — remain speculative. Although the physiological responses to stress are well characterized¹, our understanding remains limited when it comes to elucidating how these responses extend to pathological changes that contribute to the development and progression of CVD. Epidemiological and clinical studies have predominantly relied on self-reported measures of stress, which often show weak correlations with physiological stress reactivity or with profiles of a dysregulated stress response².

In the past decade, renewed efforts have been directed towards investigating the effects of psychological stress in laboratory settings on various aspects of cardiovascular physiology, immune function, myocardial ischaemia, neurobiology and cardiovascular outcomes³. Emerging evidence supports a conceptual framework that chronic or cumulative stress causes dysregulation of adaptive stress response systems. These responses can, in turn, lead to physiological changes in immune, vascular and metabolic functions that have a role in the risk of CVD. In addition, stress can lead to alterations in vascular regenerative and repair physiology that also have implications for the risk of CVD. Altered responses to everyday stressors or resultant negative emotions can amplify these effects.

In this Review, we focus on issues with the measurement of psychological stress and on the underlying pathobiology connecting stress to the risk of CVD. Maladaptive biological stress responses are highlighted as key mechanisms mediating the connection between stress and CVD outcomes. Given that the damaging cardiovascular effects of psychological stress tend to be more pronounced in individuals with a high risk of CVD than in the general population, except for severe stress in childhood⁴, much of our discussion of mechanistic pathways is based on data from individuals with pre-existing CVD. Although mental health conditions such as depression and PTSD are linked to stress⁵, they are not a focus of this Review.

Measurement of stress

Stress is conceptualized as the perception of environmental demands surpassing the capacity of an individual to adapt to a given situation⁶. Stress is pervasive in our society but its effects on health and disease vary because they depend on factors such as timing, duration, severity, individual susceptibility and physiological responses². Consistent evidence has established a connection between exposure to stress and an elevated risk of CVD both in individuals with no history of CVD and in those with pre-existing cardiometabolic conditions^3,7–9. A similar association has been observed for mental health conditions linked to traumatic stress such as mood disorders and PTSD^10,11.

An important challenge of this research field lies in the myriad definitions of stress and the substantial inconsistencies in measurement techniques, timescale and dimensions between studies^2,4 (Box 1). Stress can be assessed across multiple dimensions, encompassing actual exposures in the social and physical environments, individual psychological perceptions, and the effect of exposures on cognitive, emotional, behavioural or biological domains². At times, the term ‘stress’ has referred to actual life experiences, ranging from minor daily annoyances to traumatic events, the latter defined as actual or threatened death, serious injury, or sexual violence¹². Alternatively, the term has been used to characterize ‘stress responses’, which in turn are multifaceted and encompass emotional, behavioural and biological reactions to stressful situations². Emotional responses can include a variety of states, such as anxiety, fear, sadness and anger, as well as perceptions of distress, vital exhaustion, helplessness and lack of control. Further complexity comes from the fact that stressful events can precede common psychiatric disorders, especially depression and PTSD, and symptoms of these mental conditions are sometimes used as proxies for stress^4,13.

Box 1 Variation in stress measurements.

Examples of variations in characteristics of common measurements of stress in epidemiology and clinical research. The categories are not mutually exclusive.

Timescale of stress exposure

• Acute stressors (for example, <1 week)

• Chronic stressors (for example, ≥6 months, continuous or recurring)

Perspective of stress measurement (subjective versus objective)

• Perceived, self-reported psychological distress

• Actual stressful or traumatic exposure (such as natural disaster or death of a loved one)

• Objectively measured stress exposure (such as stress biomarkers or wearable devices in daily life)

• Objectively measured stress response (such as biological or emotional changes with laboratory stress testing)

Severity of stress exposure

• Minor daily hassles or annoyances

• Social, interpersonal or emotional stressors captured on a continuous scale of severity

• Traumatic event (actual or threatened death, serious injury, or sexual violence)

Time window of stress measurement

• Current status (now)

• Retrospective (for example, past month, past year, ever in life)

• Prospective (for example, assessments across the life course or daily measurements through ecological momentary assessments)

Life period of stress measurement

• Childhood or early life

• Adult life

• Cumulative across lifespan

Content of stress exposure

• Global subjective stress (generic distress or multiple dimensions)

• Specific life domains (such as job stress, neighbourhood stress, financial stress or caregiving stress)

• Discrete stressful event (such as a natural disaster or acute bereavement)

Logistics of stress measurement

• Laboratory (for example, mental stress provocation)

• Natural environment (for example, recalled exposures or daily monitoring in real life)

Dimension of stress response

• Cognitive response (such as appraisals of threat, challenge, uncontrollability or social evaluation)

• Emotional response (such as negative affect, anger, anxiety or worry)

• Mental health response (such as symptoms of depression or post-traumatic stress disorder)

• Biological response (such as autonomic, haemodynamic or inflammatory changes during provoked stress)

• Behavioural response (for example, coping behaviours such as smoking, overeating, seeking or avoiding social interactions or situations)

Timescale and life course

The timing and duration of stress have been implicated in its consequences on health and disease. A distinction is often made between acute and chronic stressors.

Acute stressors.

Acute stressors are short-term stress-inducing events or situations, sometimes defined as lasting less than 1 week⁸. Given their brief duration, acute stressors are generally thought to have less of an effect on disease development than chronic exposures. Nevertheless, their high intensity can worsen pre-existing conditions and precipitate cardiovascular events in individuals with underlying coronary atherosclerosis^4,14.

Substantial evidence links discrete acute stressors with CVD events, including environmental disasters and situations that trigger acute emotional states such as bouts of anger, fear, bereavement or extreme excitement. Given the unpredictability of these exposures, studies have used ecological approaches or retrospective designs, such as the case–crossover method, to investigate them. Using ecological population studies, major natural disasters (such as earthquakes) or terrorist attacks have been associated with spikes in CVD events in the aftermath of the event, and when causes of death were examined, no increase in non-coronary fatalities was observed¹⁵. However, these population-level studies have limitations owing to a lack of information on the specific circumstances surrounding the cardiac events in the affected individuals. Indeed, other than acute emotional stress, cardiac events might be triggered by concurrent factors such as vigorous physical exertion (for example, running away).

More insight could be gathered from studies focusing on emotional triggers at the individual level, whereby people are asked about their experiences before the onset of symptoms, but individual self-reporting can be influenced by recall bias. To mitigate this bias, some studies have used the case–crossover design, which uses study participants as their own controls by comparing the frequency of a specific exposure in the hours immediately preceding symptom onset with its frequency during a control period such as a few days earlier. A systematic review incorporating nine independent case–crossover studies on anger outbursts reported a pooled risk estimate of 4.7 for acute cardiovascular events¹⁶. Other acute negative emotions, such as grief and sadness, have been reported as triggers of cardiovascular events^17,18. Although the absolute cardiovascular risk from these emotional triggers is small in the general population (contributing to between 3% and 4% of all acute cardiac events), the risk is higher among individuals with elevated cardiovascular risk status^17,19.

Chronic stressors.

Chronic stressors are stressful experiences that persist over an extended period, often for months or even years. Chronic stress has been defined as any experience that is demanding and distressing nearly every day for 6 months or more². Chronic stressors can be continuous or recurring, might not have a clear time onset or end, and might theoretically include personal challenges, enduring difficulties, conflicts and threats that people face in their daily lives but that remain difficult to measure. Because of their persistent nature, chronic stressors are thought to have a larger role than acute stressors in facilitating disease processes in the long term such as atherosclerosis progression and the development of cardiometabolic risk factors²⁰. Types of chronic stressors that have been consistently associated with cardiovascular risk include adverse neighbourhood environment, financial hardship, low socioeconomic status, work stress, caregiving stress and interpersonal stress (including discrimination and social isolation)^21–28. Definitions of psychological distress that have included composite dimensions of perceived stress and mental health disturbances have also been associated with increased risk of adverse cardiovascular outcomes^29,30.

Low socioeconomic status, sometimes characterized as low-level occupational position, is a recognized source of chronic stress because it predisposes individuals to a complex array of stressful circumstances, such as financial adversity, work-related stress, unhealthy neighbourhoods and job loss, coupled with limited resources to counteract adversity such as education and employment opportunities²⁶. Work-related stress, in itself, has been intensely investigated. Supported by established theoretical models, such as effort–reward imbalance and job strain (job demand–control), multiple domains of job stress have been associated with an elevated risk of CVD even after adjusting for education and common risk factors for CVD^27,28. People with pre-existing cardiometabolic conditions (CVD or diabetes mellitus) are especially vulnerable to an increased risk of death from work-related stress³¹.

However, the distinction between acute and chronic stressors is arbitrary⁸. Even if the stressor itself persisted for only a few minutes, its emotional effect can linger and become chronic, which is especially true for traumatic stress. Moreover, the recurrence of acute stressors over time can lead to cumulative effects on the health of an individual over the life course, and when a daily stressor is caused by the same ongoing situation, it can be conceptualized as a form of chronic stress².

Theoretically, the effects of stress accumulate over time, with increased exposure leading to a higher likelihood of experiencing health issues². However, cumulative stress is burdensome to measure. Studies of chronic stress have mostly relied on recent perceptions of stress (for example, over the past month) and on retrospective checklists that often do not capture all the relevant stressors in a person’s life and can also be affected by recall bias. A life course approach, ideally with a prospective design, is needed to quantify the cumulative effect of stressful experiences on cardiovascular health and their interconnections over the life course. Job-related stress is one of few areas in which researchers have considered cumulative experiences across a long period of time. Higher cumulative exposures, as measured by repeated assessments, were associated with higher levels of allostatic load (the wear and tear on the body resulting from exposure to repeated or chronic stress) and cardiometabolic risk^32,33. However, these studies have focused only on work stress. Detrimental effects from a single psychological exposure can interact with other stressors and with a multitude of other risk factors present in the built, natural and social environments — these factors can collectively contribute to disease³⁴. Therefore, the study of chronic stress could benefit from an ‘exposome’ approach throughout the lifespan. This paradigm considers the intricate nature of these exposures and their cumulative effect on CVD³⁵.

A life course approach can also help to identify vulnerable periods. Adverse experiences in early life carry particular importance because they have the potential to induce long-term maladaptation through biological embedding that can accelerate the development or progression of CVD^36,37. In the general population, exposure to severe stress in childhood shows a larger effect size for incident CVD (pooled HR 2.1) than chronic stress in adulthood (such as job stress)⁴. Although unfavourable behavioural and cardiometabolic risk factors, such as smoking, substance abuse, obesity and diabetes, are more common among individuals who have faced higher levels of adversity in early life than in those who have not^38–41, the association between childhood adversity and CVD is not completely explained by lifestyle and traditional cardiovascular risk factors^39,42. Additionally, even though childhood stress forecasts stress in adulthood, these exposures predict CVD end points independently and show additive effects^40,42. Observations from population studies suggest that stress in early life is especially implicated as an aetiological factor in early-onset CVD, probably by accelerating the development of atherosclerosis among young or middle-aged populations^42,43, whereas stress in adulthood has a more pronounced role as a trigger of CVD events among individuals at high risk such as those with pre-existing clinical or subclinical CVD^4,14.

Subjective versus objective measures of stress

An important distinction in stress measurement is between perceived stress (the subjective appraisal of a stressor) and objectively measured stress (a stressor that is independently measured). Apart from ecological studies examining the effect of environmental disasters on death rates or other indices of CVD risk at the population level (described in the section “Acute stressors”), most epidemiological research has relied on subjective, self-reported measures of stress. Studies of objectively measured stressors and their effect on emotions and physiology have been mainly conducted in laboratory settings (which is addressed in a later section) because measuring stressful events objectively in real-life situations is challenging owing to their complexity and unpredictability.

Established and reliable objective ‘biomarkers’ of stress that can be measured unobtrusively in everyday life are lacking. With progress in novel, low-profile, multimodal monitoring technologies, interest is growing in the development of objective measures of stress with the use of digital biomarkers that measure the stress response from a physiological perspective⁴⁴. These measurements are conducted in conjunction with ecological momentary assessments of stress and mood, prospectively assessed in real time or at random time points in a person’s natural environment⁴⁵. Evolving wearable technologies offer enormous potential for the collection of ecologically valid data on dynamic, within-person processes in everyday life, which can inform the assessment of stressful events and the study of stress responses in real life⁴⁶. These techniques, sometimes known as ‘passive sensing’ digital phenotyping, are attractive because they have a low participant burden and use objective measures that do not rely on self-reporting and recall^46,47.

The vagus nerve has a crucial role in adaptation to the environment, and heart rate variability (HRV) has long been proposed as an index of vagal function and a possible link between everyday psychological moments and adverse health outcomes⁴⁸. HRV measures the variation between consecutive heartbeats and describes changes in autonomic balance in response to various psychological and physiological stimuli. HRV has shown a consistent association with CVD incidence and mortality^49–51 as well as with stress, depression and PTSD^52–54. However, HRV is a non-specific marker of autonomic function and can be influenced by many factors other than stress. Its value could be increased if measured in conjunction with ecological momentary assessments of stress in daily life⁴⁶. With advances in passive sensing technologies, novel indices of heart rate fluctuation patterns have been extracted from electrocardiographic data that might provide greater physiological specificity than HRV. For example, distribution entropy⁵⁵ and deceleration capacity^56,57 have shown prognostic value, and deceleration capacity carries specific physiological relevance as a marker of vagal function on the heart⁵⁸. Other measures of autonomic function, including pre-ejection period, galvanic skin response and T-wave amplitude, show promise as quantifiable, unbiased indicators of stress physiology that are measurable through ambulatory electrocardiography and other ambulatory sensors^59–61. Wearable device technologies might ultimately allow the collection and integration of a large amount of physiological data related to stressful exposures in a person’s natural environment.

Measuring exposure to stress objectively is important but equally important is the psychological response to stress. There are two examples in the cardiovascular literature. One is perceived financial strain, which is related to worse cardiovascular health⁶² and an increased incidence of CVD⁶³ in the general population as well as to worse health outcomes after myocardial infarction⁶⁴. These associations are independent of objective socioeconomic indicators such as education, employment status, income and access to care, suggesting that psychological distress, anxiety and worry might have an important role in addition to an individual’s financial resources. The other example is caregiving stress. Although the overall burden of caregiving has been associated with an increased risk of CVD²², the adverse effects of caregiving on health especially apply to caregivers who report emotional strain⁶⁵.

Provoked versus naturally occurring stress

Two research settings in which stress has been studied, specifically for the evaluation of stress responses, include the laboratory, where stress is experimentally provoked with a mental stress task, and real life, often through the use of wearable devices.

Stress testing in the laboratory.

The laboratory facilitates the measurement of acute stressors and provides a controlled environment for assessing an individual’s physiological responses to stress. In healthy people and those with CVD and other cardiometabolic conditions, studies of laboratory mental stress have reliably demonstrated the expected cascade of effects, including sudden increases in heart rate, blood pressure and inflammatory mediators together with an increase in biomarkers of neuroendocrine activation^2,66–68. Coronary and peripheral microvascular constriction and a transitory increase in arterial stiffness and endothelial dysfunction have also been demonstrated^69–72.

Although responses to acute laboratory stressors are transitory, they serve as indicators of an individual’s typical stress response pattern, carrying implications for long-term health. Furthermore, laboratory studies help to understand the potential mechanisms underlying the triggering of cardiovascular events by acute stress. However, even in the controlled setting of a laboratory, stressors vary in their characteristics, intensity and duration. In a meta-analysis of 208 laboratory studies, stressful tasks containing elements of lack of control and social–evaluative threat (such as performance evaluations that could be negatively judged by others) were associated with the largest acute changes in cortisol and adrenocorticotropin hormone and the longest times to recovery⁷³.

A limitation of laboratory studies of mental stress is that they can lack generalizability to stress in the real world. Stress occurring in a naturalistic setting can vary in typology, context, frequency and intensity compared with a laboratory stressor and can also be influenced by the variety of social interactions, activities and environments encountered in daily life. Furthermore, the laboratory paradigm inherently lacks the capacity to explore chronic stress, and ethical constraints prohibit the examination of intense stress in a controlled laboratory setting.

Stress in real life.

Studies have described associations between work-related stress and measures of autonomic physiology, such as blood pressure and HRV, sometimes using ambulatory devices^74,75. Exposures in these studies were usually assessed retrospectively. To date, little is known about individual stress responses when they dynamically and prospectively occur in the natural environments and in real time during daily life⁴⁶. Studies of situational stressors (such as giving a speech in real life or undergoing an oral examination) have generally shown increases in blood pressure and heart rate during these tasks⁷⁶ but, similarly to laboratory stress provocation, they lack ecological validity for everyday stressors. Some studies found that heart rate and blood pressure responses to giving a speech (or other high-demand tasks) in real life were similar to or higher than the responses to a laboratory stressor^76–78, but findings have not been consistent⁷⁷.

One method to circumvent these limitations is the use of wearable sensing devices in conjunction with ecological momentary assessments; the latter capture variations in experiences and behaviours in real time during daily life (such as by delivering brief surveys through a mobile device)⁷⁹. Studies using these techniques have generally confirmed that everyday stressors, psychological arousal, tension and anxiety are related to measures of cardiovascular reactivity⁸⁰. Most studies have focused on ambulatory blood pressure monitoring^81–83 and, more recently, ambulatory electrocardiographic monitoring^47,84,85. However, the recording time was often short and sample sizes were small. Most studies included young and healthy volunteers, and no studies examined populations at high risk such as those with CVD. Advances in mobile health technology will presumably allow more rigorous approaches for prospective daily sampling of self-reported stressful experiences in conjunction with passive ambulatory sensing of cardiovascular reactivity whilst considering physical movement and posture. A potential strength of this methodology is the ability to capture both between-person and within-person variations in exposures and responses in real life, which should improve the precision of measurements.

Pathophysiological mechanisms of stress and CVD

The systemic biological response to stress involves a well-characterized cascade of physiological changes in multiple interconnected systems, including neuroendocrine, autonomic, immune, haemodynamic, vascular and metabolic physiology⁶. Evidence indicates that the adverse effect of stress on the cardiovascular system is largely attributable to the consequences of maladaptive stress response physiology^3,7,80.

Stress response regulation: the autonomic nervous system

The autonomic nervous system has a crucial role in the regulation of systemic responses to stress. As demonstrated by studies of laboratory mental stress, sympathetic activation and parasympathetic withdrawal with acute stress cause a sudden increase in heart rate and blood pressure^69,70 as well as increased inflammation and hypercoagulability^67,86. These acute changes could heighten the risk of atherosclerotic plaque rupture and thereby trigger acute coronary events in vulnerable individuals^4,87. Chronic effects have also been implicated. Chronic or severe exposure to stress, especially if occurring early in the life course, has been hypothesized to cause dysregulation of autonomic nervous system responses⁸⁸, probably secondary to brain mechanisms discussed in the section “Neurobiology of stress and CVD”.

Responsiveness of autonomic function to everyday demands is vital for normal functioning and health, and dysregulation of autonomic responses to stress can manifest as either hyperreactivity or blunted reactivity^80,89. An area of emerging knowledge is the establishment of biomarkers of autonomic flexibility to help assess the physiological effect of stress and stratify individuals based on risk. Measures such as HRV and other ambulatory indices of autonomic function, such as deceleration capacity (a measure of vagal tone^56,57) and pre-ejection period (a measure of sympathetic activation^60,61), when assessed in conjunction with ecological momentary assessments of stress and mood in everyday life, can signal changes in autonomic function during stressful and emotional situations.

The cardiovascular reactivity hypothesis

The ultimate mechanisms by which maladaptive autonomic stress responses lead to pathological changes contributing to CVD have not received adequate attention because few studies have provided direct evidence for the underlying pathobiology⁸⁰.

In a widely accepted model (Fig. 1), the perception of stress initiates negative emotions, such as feelings of distress, fear and avoidance, and at the same time activates the neuroendocrine and autonomic response systems that are part of the acute stress response⁹⁰. This integrated cascade of effects ultimately alters vascular and haemodynamic physiology, together with immune function and inflammation³. According to the cardiovascular reactivity hypothesis, when stressors become recurrent or chronic, repeated activation of these response systems exerts cumulative, long-term alterations to vascular, immune, inflammatory and metabolic processes⁶. Exaggerated or protracted systemic responses to stress are theorized to cause shear or tensile mechanical stress on the walls of blood vessels — a ‘wear and tear’ phenomenon that might accelerate atherosclerosis, increase plaque vulnerability and trigger cardiovascular events⁶. Furthermore, chronic activation of stress response systems can cause metabolic and immune alterations, including dysregulated glucose and lipid homeostasis such as insulin resistance, obesity and hyperlipidaemia⁹¹.

Fig. 1 |. Mechanistic model linking psychological stress and CVD.

Chronic stressors cause dysregulation of adaptive stress response systems by acting on specific brain areas that are part of the corticolimbic system, which are also implicated in the development of stress-related mental health conditions. These brain regions, in turn, regulate neuroendocrine and autonomic response systems. The latter influence peripheral vascular, immune and haemodynamic physiology. These effects can be chronic but can be exacerbated by dynamic perturbations during everyday acute stressors as well as by mental health comorbidities such as depression and post-traumatic stress disorder (PTSD). Ultimately, these pathways can cause atherosclerotic plaque disruption, myocardial ischaemia, atherothrombosis and cardiac arrhythmias, especially among individuals with a high risk of cardiovascular disease (CVD), leading to acute coronary events. However, the effects of stress vary between individuals and population groups. These differences might reside in variations in stress response physiology and in the molecular pathways connecting peripheral stress responses to the risk of CVD.

Although exaggerated reactivity has received the most attention in the cardiovascular literature, blunted reactivity, observed especially in individuals with comorbidities or pre-existing CVD, is also maladaptive and can affect everyday function and future health^80,92,93. Although the reasons for a blunted response to stress are unclear, it could be caused by a protracted adrenergic overdrive (for example, from chronic stress or comorbid conditions), with subsequent loss of normal fluctuations in cardiac β-adrenergic receptor inotropic responsiveness. This situation could result in a decrease in normal heart rate and blood pressure variability and a blunted increase in cardiac output in response to stress^60,80,92. Both excessive and blunted cardiovascular reactivity to acute mental stress have been linked to cardiometabolic risk factors^66,68,89. In a 2023 study, blunted haemodynamic reactivity was associated with adverse outcomes among individuals with CVD⁹³.

Despite this accumulating knowledge, empirical evidence for the cardiovascular reactivity hypothesis remains limited. Previous research on stress reactivity has focused on healthy populations and blood pressure as an outcome, with limited information on other cardiovascular end points^66,80,94. Nevertheless, evidence suggests that physiological responses to stress, especially those involving haemodynamic, vascular and immune function, have a fundamental role in cardiovascular risk^{3,7,80,93,95–98}. Much of this research has involved mental stress testing in the laboratory. Future progress in digital phenotyping using monitoring devices and novel biomarkers of stress and stress reactivity will hopefully extend this knowledge to real-time stressful exposures in natural settings.

Haemodynamic changes and vasomotion

During mental stress, there is an elevation in both heart rate and blood pressure, akin to, albeit to a lesser extent than, the response observed during aerobic exercise. These changes lead to an increase in cardiac output and myocardial oxygen demand, estimated by the rate–pressure product. At the same time, during mental stress, there is an increase in systemic vascular resistance and cardiac afterload from sympathetically mediated peripheral vasoconstriction. This situation differs from that during exercise, when there is a decrease in systemic vascular resistance as a result of β₂-adrenergic receptor activation and the release of local vasodilators in skeletal muscle⁹⁹. Haemodynamic changes (as manifested by an increase in heart rate and blood pressure) and vasoconstrictive responses during mental stress have been demonstrated in a number of studies^69,100. Men tend to have a more pronounced peripheral vasoconstriction in response to mental stress than women^101,102; however, women tend to have a more pronounced microvascular response to stress¹⁰³.

At the coronary artery level, in the absence of coronary artery disease, mental stress induces dilatation in both epicardial arteries and microvessels whereas, in the presence of coronary atherosclerosis, a paradoxical constriction of coronary arteries and resistance vessels occurs^104,105. A notable correlation exists between coronary microvascular changes with mental stress and endothelium-dependent (but not with endothelium-independent) coronary microvascular function⁷⁰. Additionally, a robust correlation is observed between coronary and peripheral vasomotor responses to stress, such that individuals with reduced coronary microvascular flow during mental stress also have reduced peripheral microvascular flow in response to the stressor⁷⁰. These findings suggest that vasoreactivity to mental stress is a generalized phenomenon and that microvascular endothelial dysfunction has a role⁷⁰.

Immunity, inflammation and vascular regenerative pathways

Inflammatory and immune mechanisms are widely recognized for their involvement in atherosclerosis progression, plaque vulnerability and CVD risk. Acute and chronic psychological stress as well as stress-related mental disorders, such as depression and PTSD, have been linked to elevated levels of circulating inflammatory markers and indices of immune dysregulation^{7,67,106–108}. Children who endured adversity have elevated systemic inflammation 20 years later, a phenomenon that persists even after accounting for other childhood exposures and health behaviours¹⁰⁹. Therefore, the immune system might be a crucial intersection between psychological stress and cardiovascular risk.

Extensive intercommunication exists between the sympathetic nervous system and the immune system¹¹⁰, given that stress-induced inflammatory responses might evolutionarily have had short-term adaptive value¹¹¹. Noradrenaline release during stress activates the transcription factor nuclear factor-κB in circulating monocytes, initiating the inflammatory cascade¹¹². The hypothalamic–pituitary–adrenal axis is also involved in the regulation of immune responses through the glucocorticoid receptor¹¹³. The bone marrow, spleen and other lymphoid organs have extensive sympathetic innervation, and immune cells and their progenitors express adrenergic receptors¹¹⁰. Therefore, through β-adrenergic receptor signalling, chronic and acute stress can trigger the immune activation of haematopoietic organs, increasing the production of progenitor cells and inducing leukopoietic proliferation and the release of pro-inflammatory cytokines^7,114,115.

CD34⁺ progenitor cells in the bone marrow can differentiate into haematopoietic and endothelial cells, and CD34⁺ progenitor cell counts in peripheral blood are independent predictors of adverse cardiovascular outcomes^116,117. In addition to being a marker of immune cell activation, progenitor cell counts in the blood are indicative of endogenous regenerative and repair capacity because their production is stimulated by endothelial injury and cardiovascular risk factors¹¹⁸. These cells are also mobilized after acute mental stress⁹⁷, a phenomenon that is especially pronounced in patients with CVD who develop ischaemia with mental stress¹¹⁹.

Important sex-specific differences exist in stress and inflammation. Young and middle-aged women with CVD have an elevated inflammatory response to stress (particularly circulating levels of IL-6) compared with male counterparts¹²⁰. Furthermore, among women, a direct association was observed between the stress-induced increase in plasma IL-6 levels and adverse cardiovascular events, whereas no such association was found among men⁹⁸. A parallel sex-specific relationship was identified for the chemokine CCL2. Although the mechanisms underlying these sex-specific differences are not fully understood, variations between women and men in neuronal circuitry and brain–immune interactions might have a role¹²¹.

From stress response perturbations to CVD

The substantial body of literature connecting stressful stimuli to haemodynamic, vascular and immune disturbances is supported by brain imaging data. These responses correspond to distinct activation patterns in brain regions associated with emotion regulation, including the insula, parietal cortex, medial prefrontal cortex, amygdala and other limbic system areas^7,97,122,123. Nevertheless, the prognostic implications of these alterations have only recently been explored.

Over the past 5 years, cohort studies of individuals with CVD have revealed that those with greater digital vasoconstriction during mental stress (measured using digital tonometry) had a heightened risk of adverse CVD events during follow-up⁹⁶. Furthermore, a transient decline in endothelial function with mental stress, measured through flow-mediated vasodilatation of the brachial artery, was also an independent predictor of adverse CVD events⁹⁵. A blunted haemodynamic response to mental stress, expressed as a lower increase in the rate–pressure product during the stressor, was also significantly related to adverse cardiovascular outcomes⁹³. These stress-induced vascular and haemodynamic changes significantly improved risk prediction in patients with CVD, above and beyond the use of traditional risk indicators, severity of disease, and resting vascular and haemodynamic status^95,96. Even the magnitude of progenitor cell mobilization from the bone marrow with mental stress emerged as an independent predictor of adverse cardiovascular events⁹⁷. Of note, sex-specific differences were observed, indicating that peripheral microvascular and inflammatory responses to mental stress are more robust predictors of cardiovascular outcomes in women than in men^98,103.

Collectively, these observations suggest that the physiological response to stress has a fundamental role in cardiovascular risk, and that haemodynamic, vascular and immune perturbations triggered by stress are especially implicated. These data align with a central role of the acute stress response and the cardiovascular reactivity hypothesis (Fig. 1). However, how these laboratory observations extend to the pathology affecting long-term risk remains elusive. Dynamic perturbations in autonomic, immune and vascular response systems are likely also to be risk mechanisms for chronic, recurring or cumulative stressors in daily life, but more data are needed from prospective studies and from assessments in real-life situations.

Mental stress-induced myocardial ischaemia

In about 15–20% of patients with clinically stable coronary heart disease, acute mental stress provoked in laboratory testing induces myocardial ischaemia, a condition that is detectable through myocardial perfusion imaging^124,125. Mental stress-induced myocardial ischaemia (MSIMI) is analogous to the ischaemia provoked by exercise (for example, treadmill testing), except that the stimulus is psychological rather than physical^99,126.

The prognostic relevance of MSIMI had long been suspected on the basis of findings from small studies mostly conducted decades ago¹²⁶. A large study has now unequivocally demonstrated that MSIMI, measured with single-photon emission computed tomography myocardial perfusion imaging, is a powerful predictor of adverse cardiovascular events among individuals with CVD¹²⁷ (Fig. 2). After multivariable adjustment, MSIMI was associated with a more than twofold increase in the risk of cardiovascular events, which was more pronounced than the risk associated with myocardial ischaemia induced by a conventional stress test (such as exercise or pharmacological stress). Of note, cardiovascular risk factors, other clinical characteristics, psychological disturbances and even a propensity to develop ischaemia with a conventional stress test did not account for the excess risk associated with MSIMI. This finding suggests that this phenomenon holds clinical importance beyond that of established risk indicators.

Fig. 2 |. Cumulative incidence of cardiovascular disease end points for separate myocardial ischaemia phenotypes.

The graph shows the incidence of cardiovascular death or non-fatal myocardial infarction (MI) during a median follow-up of 5 years, classified according to the type of stress test used to induced myocardial ischaemia, in 899 individuals with coronary heart disease. Myocardial ischaemia induced by mental stress, conventional stress or both was assessed at baseline with single-photon emission computed tomography myocardial perfusion imaging. Mental stress was induced in the laboratory using a speech task that included elements of social–evaluative threat. Conventional stress testing was performed using either an exercise stress test or a pharmacological stress test on a separate day within 1 week from the mental stress. Reprinted with permission from ref. ¹²⁷, JAMA.

MSIMI differs from ischaemia induced by exercise stress in a number of ways. MSIMI is typically silent, occurs at a lower haemodynamic workload, has not consistently shown an association with coronary atherosclerosis, and can manifest in the absence of ischaemia with a conventional stress test or after successful revascularization^125,128. Psychological factors, such as depression^129,130, PTSD¹³¹ and a composite measure of chronic distress¹³, have been associated with MSIMI but not with conventional stress ischaemia. These characteristics suggest that MSIMI and ischaemia induced by a conventional stress test are distinct phenomena with different mechanisms.

Patients who develop MSIMI have a more pronounced peripheral vasoconstriction with mental stress than patients who do not have MSIMI^69,132. These patients also have a pronounced haemodynamic response during mental stress as measured by the rate–pressure product⁶⁹. These findings suggest that a combination of generalized vasoconstriction leading to reduced coronary perfusion, coupled with increased myocardial oxygen demand, is an important substrate for MSIMI^70,99. This situation differs from conventional stress ischaemia, which is primarily provoked by flow-limiting coronary stenoses. Platelet hyperactivity¹³³ and subclinical myocardial dysfunction¹³⁴ have also been associated with MSIMI.

There are important sex-specific differences in MSIMI. First, this condition occurs more often in women, particularly younger women, than in men of a similar age^124,135,136. Second, the underlying mechanisms can differ between women and men. In keeping with observations that microvascular responses to stress are more pronounced and more prognostic in women than in men¹⁰³, microvascular reactivity with stress is a more important substrate for MSIMI among women^101,124,137 whereas obstructive coronary artery disease and haemodynamic response to stress have a greater role in men^101,128. Furthermore, women with MSIMI, unlike men, more often report angina in everyday life, even though MSIMI is typically asymptomatic in a laboratory setting¹³⁸. These characteristics share similarities with the syndrome of ischaemia with no obstructive coronary arteries, which is also more common in women than in men and is similarly linked to psychological factors, daily angina and microvascular disease^138,139. MSIMI and ischaemia with no obstructive coronary arteries might share a similar pathophysiology, particularly in women.

Neurobiology of stress and CVD

For an event to be registered by the body as stressful, it must be perceived by the senses. A stressful event, such as a tiger jumping out at you whilst you are walking down a trail, is initially perceived through the senses (vision and hearing). These sensory perceptions activate specific areas of the sensory cortex of the brain, and this information is then processed in the corticolimbic system, including the medial prefrontal cortex, the hippocampus and the amygdala. The hippocampus creates a memory that it compares to previous memories and sends a message to the amygdala, the centrepiece of the fear response¹⁴⁰. The amygdala then initiates the peripheral fight-or-flight response, with activation of the sympathetic nervous system and release of cortisol from the adrenal gland. Studies using ¹⁸F-fluorodeoxyglucose PET have linked amygdalar resting metabolic activity with the risk of subsequent CVD events^7,115. Consistent with the notion that the sympathetic nervous system and the hypothalamic–pituitary–adrenal axis regulate immune function, amygdalar activity was also associated with bone marrow activation and arterial inflammation^7,115. Brain areas involved in the stress response are also implicated in stress-related psychiatric disorders, including depression and PTSD, conditions that are both associated with CVD^10,11. The hippocampus, in addition to mediating memory, is highly sensitive to stress, and PTSD and depression are associated with a smaller hippocampal volume and deficits in hippocampal-based declarative memory¹⁴⁰.

The medial prefrontal cortex, which includes the anterior cingulate, the orbitofrontal cortex and adjacent brain regions, sends inhibitory inputs to the amygdala and thereby has an important role in the extinction, or suppression, of learned fear responses¹⁴¹. Investigations using laboratory-based mental stress tests in conjunction with imaging of the heart (using single-photon emission computed tomography) and the brain (using PET) have mapped mechanisms by which stress can confer an increased risk of adverse CVD outcomes. Patients with CVD who developed MSIMI had increased function in the anterior cingulate gyrus (medial prefrontal cortex) with mental stress compared with patients with CVD who did not have MSIMI¹⁴², and brain activation in this region correlated with increased stress-induced peripheral vasoconstriction¹²². Of note, activity in this brain region with mental stress, and specifically in the rostromedial prefrontal cortex, was associated with adverse cardiovascular outcomes in individuals with CVD¹²³. This area has an important role in the regulation of emotions, autonomic function and immune responses to stressful stimuli, and the association of activity in this brain area with CVD outcomes was partially mediated through a stress-induced decrease in HRV and increased inflammation (circulating IL-6 levels)¹²³. This brain region is also implicated in self-perception and social hierarchy¹⁴³; therefore, its role is consistent with observations that stressful tasks involving social rejection and lack of control have large effects on physiological responses and long-term health^2,73.

Collectively, these findings provide direct neurobiological evidence linking psychological stress to adverse cardiovascular outcomes through the regulation of autonomic and immune responses to stress, with the medial prefrontal cortex having a modulating role (Fig. 3). Of note, individual susceptibility to these effects varies. The patterns of brain activation during mental stress were found to differ based on sex¹⁴⁴ and the presence of obesity¹⁴⁵, a history of depression¹⁴⁶, early-life trauma¹⁴⁷ or angina¹⁴⁸. Furthermore, individual neurobiological resilience might have a role. Although the research on neurobiological resilience is still emerging, it has been conceptualized as the ability to maintain a stable psychological and physical equilibrium in the face of severe loss or trauma¹⁴⁹. Brain and other biological markers associated with the stress response have been implicated in resilience^150,151. A neurobiological marker of resilience, defined as lower stress-related amygdalar metabolic activity despite chronic socioeconomic or environmental stress, was associated with a decreased risk of CVD¹⁵².

Fig. 3 |. The central role of stress response physiology in cardiovascular risk secondary to psychological stress.

The medial prefrontal cortex and other regions of the corticolimbic systems have a pivotal role in the regulation of both emotional and physiological responses to stress. Physiological responses to stress that have a demonstrated connection with cardiovascular disease (CVD) events in prospective studies include haemodynamic, ischaemic, vagal, vascular and inflammatory responses to stress. Most of these physiological responses have also been linked to specific brain activation patterns in brain imaging studies of acute mental stress.

Individual variability and special populations

The consequences of stress on health vary from person to person. Factors such as the genetic predisposition of an individual, coping strategies, social support, health behaviours and contextual factors (cultural and economic resources and physical environment), can influence how acute and chronic stressors are perceived, how people react to stress, and the potential effect of stress on health.

Certain demographic groups show distinct vulnerabilities towards stress, including women, younger individuals and those from ethnic minority groups. Physical and mental comorbidities also affect the risk. The connection between stressful exposures, whether acute or chronic, and the risk of subsequent CVD events and death is exacerbated in individuals at high risk with pre-existing subclinical or clinical CVD and diabetes compared with those who are at low risk^4,31. Indeed, individuals with CVD and those with diabetes have dysregulation of stress response systems that might increase their risk of CVD^68,93,123. Furthermore, certain mental health conditions, such as depression and PTSD, can develop in some individuals after severe stress and further contribute to cardiovascular risk^10,11. These differences between individuals and groups might reside in variations in stress response physiology and in the molecular pathways involved in cardiovascular risk (Fig. 1).

Women

Women have a distinctive burden of psychological stressors, especially at young ages and in midlife due to, for example, financial adversity, sexual abuse, domestic violence and caregiving stress^153,154. Women also have a higher prevalence of stress-related mood and anxiety disorders than men¹⁵⁴. Some epidemiology studies have reported that early-life stress and depression are more powerful predictors of CVD incidence and mortality in young and midlife women than in men of a similar age^155–157. Psychosocial stress and depression were also found to be important correlates of future hospitalizations, delayed recovery and mortality in young women with early-onset coronary heart disease^158,159. In general, studies in animal models have confirmed the differential vulnerability of female and male individuals to the development of cardiometabolic disorders as a consequence of stress¹⁶⁰.

Psychological reactions to mental stress are not consistently different between women and men^124,161; however, biological responses differ. As described in the section “Pathophysiological mechanisms of stress and CVD”, women with coronary artery disease, particularly those in younger age groups, have a propensity to MSIMI and show a more pronounced inflammatory and peripheral microvascular response to stress. Coronary microvascular dysfunction in women probably contributes to MSIMI^137,162 and is also a feature of stress-induced cardiomyopathy, which is typically triggered by strong emotions and manifests predominantly among women¹⁶³. As noted above, in women with CVD but not in men with CVD, inflammatory and microvascular changes with mental stress have both been linked to adverse cardiac events^98,103.

Ethnic minority groups

Black American individuals have a higher exposure to social adversity than other racial or ethnic populations in the USA^164–166. This exposure includes more adverse life events, more discrimination, lower economic resources, lower access to care and more chronic stressors in general. Black American individuals might also have heightened vulnerability to the negative cardiovascular effects of psychological factors. In a study of patients with CVD in the USA, Black individuals were more likely than their non-Black counterparts to develop endothelial dysfunction with mental stress, and this response explained a substantial portion of their excess risk of adverse cardiovascular events¹⁶⁷. Among survivors of a myocardial infarction in the USA, self-reported exposure to discrimination was related to the development of MSIMI, and this association was especially strong among Black women¹⁶⁸. Nonetheless, the extent to which psychosocial stressors can help to explain the well-documented disparities in cardiovascular risk according to ethnicity needs more evaluation^166,169.

People with mental health conditions

Individuals with depression and PTSD show evidence of inflexibility in autonomic function compared with those without these conditions as suggested by lower HRV and abnormal baroreflex function^88,170–172. Extensive evidence also shows increased systemic inflammation and immune dysregulation in individuals with depression or PTSD^107,108. Individuals with major depression have larger increases in inflammatory markers than control individuals wihtout depression in response to laboratory stressors¹⁷³. Patients who survived a myocardial infarction and met the clinical criteria for PTSD also had an exaggerated inflammatory response during acute stress^174,175.

Exaggerated sympathetic reactivity to psychological stress and autonomic dysregulation have long been proposed as mechanisms linking PTSD to CVD¹¹, and data suggest a key role of both coronary and peripheral microvascular responses to stress. In PTSD, an abnormal autonomic response occurs during trauma recall tasks, suggesting a combination of increased sympathetic nervous system activity and impaired cardiac contractility^61,170,171. Even among individuals from the general population, PTSD is associated with reduced coronary microcirculatory function, which also undergoes a faster deterioration over time¹⁷⁶, and with exaggerated peripheral and systemic vasoconstrictor responses to traumatic stress reminders¹⁷⁷. Increased peripheral vasoconstriction during trauma recall in PTSD can contribute to arterial stiffness and endothelial dysfunction and can also increase cardiac afterload¹⁷⁷. At the coronary artery level, microvascular dysfunction reduces myocardial blood flow and can cause myocardial ischaemia through a microcirculatory mechanism¹⁷⁸. Indeed, MSIMI is more common in individuals with PTSD than in those without¹³¹.

Management considerations

Despite the increasing evidence linking psychological stress to CVD, stress is not normally addressed in routine clinical care. In the USA, professional societies have been hesitant to endorse the assessment and management of stress in advisory statements and guideline recommendations, although psychological influences are acknowledged and evaluation is encouraged as part of the social determinants of health^179–181. However, the European guidelines on prevention of CVD recognize psychosocial stress as an effect modifier and provide more specific guidance¹⁸². This reluctance might stem, in part, from viewing the psychological aspects as foundational to other risk factors¹⁸¹. We contend that, precisely because the psychological dimension forms the basis of other health factors, evaluating and addressing it is of paramount importance.

In the hectic environments of cardiology and primary care clinics, conducting a comprehensive assessment of psychological risk factors might be impractical. Although several short questionnaires for psychological distress exist (such as scales of perceived stress, anxiety and depression, alone or in combination) and are predictive of cardiovascular events^29,30, there is lack of agreement on which aspects should be targeted and on the best practices for screening implementation¹⁸³. More data from randomized clinical trials are needed on the utility of screening modalities and on appropriate clinical decision support systems for the evaluation and management of psychosocial stress. Nonetheless, using brief composite screening tools can help clinicians to risk stratify patients and identify those who might need referral for biobehavioural interventions²⁹. The use of a composite measure of psychological distress resulted in a significant improvement in risk prediction among patients with cardiac disease beyond the effect of traditional risk indicators, suggesting its potential applicability in a clinical setting²⁹.

Participation in stress management training, including group support, education, cognitive behavioural therapy and meditation, might mitigate perceived stress, depression, and anxiety and improve cardiovascular risk factors^184–188. Physical activity training¹⁸⁹, yoga¹⁹⁰ and biofeedback^191,192 can also be useful, as can collaborative care approaches that incorporate psychological management¹⁹³. Collectively, the data suggest that most stress-reduction management strategies can reduce psychological distress and potentially ameliorate cardiovascular risk factors and improve health behaviours. However, trials have been small and with limited follow-up, and evidence for an improvement in CVD end points is scarce¹⁸³. Medications should be considered only in cases of mental disorders such as depression, severe anxiety and PTSD. For most of these patients, if behavioural interventions are not sufficient, the cardiologist or primary care physician can start a trial of medication, for example, a selective serotonin reuptake inhibitor, without the need for immediate referral to a psychiatrist.

In summary, although robust clinical trials are required to evaluate the efficacy of screening modalities and optimal management of psychological stress for the prevention of CVD, the existing evidence underscores the importance of recognizing and addressing psychosocial factors in patient-centred clinical care to improve patient quality of life and promote a healthy lifestyle.

Future directions

The next frontiers in biobehavioural medicine involve innovative molecular, translational and clinical studies to uncover mechanisms for individual vulnerability and to test new screening modalities and therapeutic interventions that can mitigate adverse physiological responses to stress.

First, longitudinal studies through the life course are needed to demonstrate causal relationships and mechanisms linking stress and CVD, beginning at a young age. Ideally, investigations should consider the total effect of environmental exposures through the lifetime, using an exposome paradigm. This approach should facilitate the evaluation of a combination of multilevel exposures in the natural, built and social environments and their interactions with individual factors^34,35.

Second, differences in susceptibility between individuals and groups might ultimately reside in complex molecular pathways mediating stress response physiology and its connection to cardiovascular risk. These pathways, which are largely untested, can be explored using modern omics technologies and precision medicine approaches that involve changes at the cellular, molecular and organ levels and their relationship with vulnerability to CVD.

Third, techniques for studying mental stress should transition beyond laboratory settings to encompass real-life situations. Dynamic perturbations in autonomic, immune and vascular response systems with mental stress in the laboratory have been linked to CVD events in people with pre-existing CVD³. However, whether these risk mechanisms also apply to chronic, recurring or cumulative stressors in everyday life and to individuals without CVD is unknown. With progress in wearable technologies, ambulatory devices for assessing autonomic function or peripheral vasoconstriction could prove valuable in monitoring daily responses to stress in an unobtrusive way and in assessing changes achieved with interventions. The standardization, validity and reproducibility of these techniques is an area ripe for future research.

Fourth, effective screening and therapeutic strategies should be developed, ensuring not only their cost-effectiveness but also their feasibility and acceptance in a clinical setting and their ability to decrease vulnerability to CVD. Research is needed into whether information on stressful exposures can contribute to risk precision and risk stratification of individual patients (for example, with inclusion into algorithms for risk calculation) and help to identify people who are at risk and might benefit the most from interventions.

To address this research agenda effectively, it will be essential to leverage multidisciplinary teams that integrate diverse expertise such as cardiology, psychology, epidemiology, molecular sciences, engineering and bioinformatics. Given the current state of the field and the projected future directions, this multidisciplinary collaboration represents the cornerstone to advancing the science of stress and CVD.

Conclusions

Psychological stress has long been associated with negative health consequences, particularly CVD, but results have varied in terms of how stress is measured and how strong the association is. Additionally, the mechanisms and potential causal links have remained speculative despite decades of research. Physiological responses to stress are emerging as key pathways for the risk of CVD, and haemodynamic, vascular and immune perturbations triggered by stress are especially implicated. More data are needed from prospective studies and from assessments in real-life situations, taking advantage of mobile health technologies and wearable devices. Psychological assessment should be integrated with clinical care and prevention. Stress-reduction interventions might mitigate perceived stress levels and potentially reduce cardiovascular risk, although more data from large, randomized clinical trials are needed.

Key points.

• Psychological stress has long been associated with negative health consequences, particularly cardiovascular disease, but the mechanisms and potential causal links have remained speculative.

• Physiological responses to stress are emerging as key pathways for the risk of cardiovascular disease, especially haemodynamic, vascular and immune perturbations triggered by stress.

• More data are needed from prospective studies and from assessments in real-life situations.

• Mobile health technologies and wearable devices might be helpful in the prospective assessment of stressful exposures and stress responses in everyday life.

• Psychological assessment should be integrated with clinical care and prevention.

• Stress-reduction interventions might mitigate perceived stress levels and potentially reduce cardiovascular risk, although more data from large, randomized trials are needed.