This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rozanski, A. Articles by Kaplan, J. Search for Related Content PubMed PubMed Citation Articles by Rozanski, A. Articles by Kaplan, J. Pubmed/NCBI databases Medline Plus Health Information Mental Health Stress Related Collections Chronic ischemic heart disease Behavioral/psychosocial - treatment Risk Factors

(Circulation. 1999;99:2192-2217.)
© 1999 American Heart Association, Inc.

Clinical Cardiology: New Frontiers

Impact of Psychological Factors on the Pathogenesis of Cardiovascular Disease and Implications for Therapy

Alan Rozanski, MD; James A. Blumenthal, PhD; Jay Kaplan, PhD

From the Division of Cardiology, Department of Medicine, St Luke's/Roosevelt Hospital Center, and the Department of Medicine, Columbia University College of Physicians and Surgeons, New York, NY (A.R.); the Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC (J.A.B.); and the Department of Pathology (Comparative Medicine) and Anthropology, Wake Forest University School of Medicine and Wake Forest University, Winston-Salem, NC (J.K.).

Correspondence to Alan Rozanski, MD, Division of Cardiology, St Luke's/Roosevelt Hospital Center, 114th Street and Amsterdam Avenue, New York, NY 10025. E-mail ar77{at}columbia.edu

Abstract

Abstract—Recent studies provide clear and convincing evidence that psychosocial factors contribute significantly to the pathogenesis and expression of coronary artery disease (CAD). This evidence is composed largely of data relating CAD risk to 5 specific psychosocial domains: (1) depression, (2) anxiety, (3) personality factors and character traits, (4) social isolation, and (5) chronic life stress. Pathophysiological mechanisms underlying the relationship between these entities and CAD can be divided into behavioral mechanisms, whereby psychosocial conditions contribute to a higher frequency of adverse health behaviors, such as poor diet and smoking, and direct pathophysiological mechanisms, such as neuroendocrine and platelet activation. An extensive body of evidence from animal models (especially the cynomolgus monkey, Macaca fascicularis) reveals that chronic psychosocial stress can lead, probably via a mechanism involving excessive sympathetic nervous system activation, to exacerbation of coronary artery atherosclerosis as well as to transient endothelial dysfunction and even necrosis. Evidence from monkeys also indicates that psychosocial stress reliably induces ovarian dysfunction, hypercortisolemia, and excessive adrenergic activation in premenopausal females, leading to accelerated atherosclerosis. Also reviewed are data relating CAD to acute stress and individual differences in sympathetic nervous system responsivity. New technologies and research from animal models demonstrate that acute stress triggers myocardial ischemia, promotes arrhythmogenesis, stimulates platelet function, and increases blood viscosity through hemoconcentration. In the presence of underlying atherosclerosis (eg, in CAD patients), acute stress also causes coronary vasoconstriction. Recent data indicate that the foregoing effects result, at least in part, from the endothelial dysfunction and injury induced by acute stress. Hyperresponsivity of the sympathetic nervous system, manifested by exaggerated heart rate and blood pressure responses to psychological stimuli, is an intrinsic characteristic among some individuals. Current data link sympathetic nervous system hyperresponsivity to accelerated development of carotid atherosclerosis in human subjects and to exacerbated coronary and carotid atherosclerosis in monkeys. Thus far, intervention trials designed to reduce psychosocial stress have been limited in size and number. Specific suggestions to improve the assessment of behavioral interventions include more complete delineation of the physiological mechanisms by which such interventions might work; increased use of new, more convenient "alternative" end points for behavioral intervention trials; development of specifically targeted behavioral interventions (based on profiling of patient factors); and evaluation of previously developed models of predicting behavioral change. The importance of maximizing the efficacy of behavioral interventions is underscored by the recognition that psychosocial stresses tend to cluster together. When they do so, the resultant risk for cardiac events is often substantially elevated, equaling that associated with previously established risk factors for CAD, such as hypertension and hypercholesterolemia.

Key Words: coronary disease • stress • psychology

Although the importance of psychosocial factors in the development and expression of coronary artery disease (CAD) has been debated, an extensive recent literature now establishes that psychosocial factors contribute significantly to the pathogenesis of CAD. Furthermore, by use of new technologies and animal models, elucidation of the basic pathophysiology underlying the relationship between psychosocial factors and CAD is expanding rapidly. However, because the literature relating psychosocial factors to CAD is multidisciplinary, there may be an underappreciation of the strength of some of the epidemiological and pathophysiological observations that have been reported. Accordingly, we will review the relationship between psychosocial stress and CAD development, with emphasis on the following psychosocial factors: (1) depression, (2) anxiety, (3) personality factors and character traits (eg, hostility), (4) social isolation, and (5) chronic and subacute life stress. Although these domains can overlap, epidemiological data for each domain will be reviewed separately, emphasizing studies that have used the "hard" cardiovascular end points of myocardial infarction and cardiac death as outcome variables (or all-cause mortality in the case of some early studies). In some instances, "alternative" cardiac end points will be considered, such as progression of atherosclerosis during serial carotid ultrasonography. Studies that relate psychosocial factors to "soft" cardiac end points, such as angina and static findings on coronary angiography, will not be reviewed. Information regarding the pathophysiology by which psychosocial factors promote CAD development will be examined for each psychological domain. These will include (1) behavioral mechanisms, whereby the given factor exacerbates lifestyles known to potentiate CAD (eg, smoking), and (2) direct pathophysiological effects, as delineated in experimental animal studies and/or investigations in humans. The pathogenic effects of 2 other phenomena, acute psychological stress and sympathetic nervous system hyperresponsivity, will then be reviewed. Finally, the implications of these findings relative to the prevention and treatment of CAD will be discussed.

Psychosocial Factors and CAD

Depression and Related Syndromes
Episodes of major depression are characterized by the presence of a depressed mood and markedly decreased interest in all activities, persisting for at least 2 weeks and accompanied by at least 4 of the following additional symptoms: changes in appetite, sleep disturbance, fatigue, psychomotor retardation or agitation, feelings of guilt or worthlessness, problems concentrating, and suicidal thoughts. The 1-month community-based prevalence of major depression episodes is 5%.¹ Among CAD patients, however, the prevalence of major depression is 3-fold higher. Also, depressive symptoms that are not sufficient in magnitude to meet the criteria for major depression occur at least as commonly among cardiac patients.^{2 3} Recent epidemiological studies evaluating the relationship between depression and CAD among healthy^{3 4 5 6 7 8 9 10 11} and CAD^{12 13 14 15 16 17 18 19} populations consistently demonstrate a significant prospective relationship between the occurrence of major depression episodes and the incidence of cardiac events (Table 1). Two additional findings are notable. First, the presence of depressive symptoms, in the absence of diagnosed major depression episodes, is also associated with an increased risk for cardiac events.⁴ Second, a number of studies support a gradient between the magnitude of depression and future cardiac events.^{4 7 9} Together, these data suggest that risk for CAD associated with depression exists along a continuum, according to the magnitude of depressive symptoms.

View this table: [in this window] [in a new window]   Table 1. Depression and Coronary Artery Disease

One particular aspect of depression, the absence of hope, has received particular attention. Hopelessness has been linked to sudden death, both in observational studies^{20 21} and in animal models of hopelessness.²² Recently, prospective epidemiological studies have also reported a relationship between symptoms of hopelessness and the development of CAD.^{4 7} In one study, for example, a positive answer to the question "(During the last month) have you felt so sad, discouraged, hopeless, or had so many problems that you wondered if anything was worthwhile?" more than doubled the risk of CAD.⁴ It has also been demonstrated that men experiencing hopelessness develop significantly more carotid atherosclerosis over time.²³ A related phenomenon is "vital exhaustion." This syndrome, measured by the 37-item Maastricht questionnaire,²⁴ focuses on a triad of symptoms: fatigue, irritability, and demoralized feelings. The presence of vital exhaustion has also been reported to predict future CAD and/or cardiac events in healthy²⁵ and CAD^{26 27} populations.

Pathophysiological Mechanisms
Considerable evidence indicates that depression has both behavioral and direct pathophysiological effects. With respect to behavioral mechanisms, depression is associated with both unhealthy lifestyle behaviors,^{28 29} such as smoking, and poor patient compliance.^{29 30} Direct pathophysiological effects of depression involve at least 3 mechanisms. First, depression is accompanied by hypercortisolemia.^{31 32 33} Associated findings include attenuation of the adrenocorticotropin hormone response to corticotropin-releasing factor administration,³² nonsuppression of cortisol secretion after dexamethosone administration,³⁴ and elevated corticotropin-releasing factor concentrations in the cerebrospinal fluid of depressed patients.³⁵ Second, depressed individuals may develop significant impairments in platelet function, including enhanced platelet reactivity and release of platelet products such as platelet factor 4 and ß-thromboglobulin.^{36 37} The combination of hypercortisolemia and enhanced platelet function establishes the theoretical basis for explaining the proatherogenic effects of depression. In addition, reduced heart rate variability³⁸ and impaired vagal control³⁹ have been reported among depressed patients. These findings suggest that depressed patients may also be subject to enhanced arrhythmogenic potential.

Anxiety Syndromes
Until recently, evidence linking anxiety to CAD was limited to demonstrations of elevated mortality rates among psychiatric patients with anxiety disorders.⁴⁰ Increasing evidence now links anxiety disorders to development of cardiac events in general populations (Table 2). Most notably, 3 large-scale community-based studies, including one involving 34 000 men, have now reported a significant relationship between anxiety disorders and cardiac death.^{41 42 43} Moreover, a dose-dependent relationship has been noted between anxiety levels and the occurrence of cardiac death.^{42 43} Anxiety has not been associated with myocardial infarction in these studies. Rather, the excess mortality appears to be confined to sudden (versus nonsudden) cardiac death.^{42 43} Notably, these community-based studies did not include women,^{41 42 43} even though anxiety disorders are more common among women.⁴⁴

View this table: [in this window] [in a new window]   Table 2. Anxiety Disorders

Prospective positive associations between CAD and panic disorder⁴⁵ and between CAD and "worry" (a subcategory of generalized anxiety disorder)⁴⁶ have also been noted in 2 recent studies; however, more studies are needed to establish whether these findings are indeed valid. The epidemiological investigation of anxiety disorders among CAD patients has also been quite sparse. Because 4 small studies have each noted a relationship between anxiety and a constellation of hard and soft cardiac events among CAD patients (Table 2),^{17 18 47 48} more large-scale epidemiological studies among CAD patients now appear to be warranted.

Pathophysiological Mechanism
The association between anxiety and sudden death, but not myocardial infarction, suggests that ventricular arrhythmias may be the mechanism for cardiac death among individuals with anxiety disorders. In support of this hypothesis, it has been observed that individuals with anxiety disorders have reduced heart rate variability.⁴⁹ Hence, there may be a pathological alteration in cardiac autonomic tone. This alteration could involve either increased sympathetic stimulation, which has been linked to the occurrence of arrhythmias and sudden death,⁵⁰ or impaired vagal control, which has also been linked to increased cardiac mortality.^{51 52} With respect to the latter possibility, reduced vagal control has been linked to impaired, vagally mediated baroreflex control of the heart.⁵³ Such impairment appears to be a particularly important risk factor for sudden death.^{54 55} Along these lines, a recent study reported reduced baroreflex cardiac control in patients with anxiety,⁵⁶ but prospective work is needed to determine whether this is a common operative mechanism for sudden deaths among patients with anxiety syndromes.

Individuals with anxiety disorders are prone to more unhealthy lifestyle behaviors^{42 43} ; however, the lack of correlation between anxiety syndromes and myocardial infarction (a sign of underlying atherosclerosis) suggests that, at least among initially healthy individuals, this behavioral association is not a significant pathogenic mechanism. It is conceivable, nonetheless, that this behavioral association could be of importance among CAD patients manifesting anxiety.

Personality and Character Traits
After the identification of the type A behavior pattern by Friedman and Rosenman in the late 1950s,⁵⁷ a syndrome characterized by competition, hostility, and exaggerated commitment to work, many studies have investigated whether personality patterns or individual character traits promote the development of CAD. Other personality types have included "type D" personality,¹⁷ "social dominance,"⁵⁸ and a "hardy personality" construct⁵⁹ ; these latter personality types have not been widely studied as potential risk factors for CAD. Interest in type A behavior accelerated after the Western Collaborative Group Study, which reported that type A behavior was associated with a 2-fold increased risk of CAD and 5-fold increased risk of recurrent MI over an 8.5-year follow-up.⁶⁰ Although type A behavior continues to receive attention,⁶¹ a series of studies have reported no correlation between type A behavior and CAD risk.^{62 63 64 65 66} This lack of consistency has cast doubt on the potential robustness of the type A behavior as a clinical syndrome. Potential confounders have been suggested. For instance, animal model studies⁶⁷ and some human studies^{68 69} suggest that social support is a potential confounding variable. Furthermore, suspecting that not all components of type A behavior are pathogenic, investigators have examined the components of this behavior pattern.

Hostility, a major attribute of the type A behavior pattern, has received considerable attention as a potential "toxic" element in this personality construct. Hostility is a broad psychological construct, encompassing negative orientations toward interpersonal relationships, and includes such traits as anger, cynicism, and mistrust. Table 3 lists 10 studies assessing the relationship between hostility and CAD in healthy subjects.^{70 71 72 73 74 75 76 77 78 79} The results of these prognostic studies are mixed, with both positive and negative studies. However, the studies are of uneven quality. For instance, in one negative study, 57% of the individuals were lost to follow-up,⁷⁵ whereas the follow-up of healthy individuals in another study was only 3 years.⁷⁶ Of note, 2 large studies that used tailored scales to focus on cynical mistrust⁷⁷ and anger⁷⁹ have yielded positive associations with cardiac events. In the anger study, a gradient was noted between anger levels and the frequency of subsequent cardiac events.⁷⁹ Thus, it is possible that certain components of the hostility construct are more pathogenic.

View this table: [in this window] [in a new window]   Table 3. Hostility and the Pathogenesis of Coronary Artery Disease

To date, there have been no large-scale epidemiological studies evaluating hostility among CAD patients. Four small epidemiological studies among CAD patients, however, have been positive, as noted in Table 3.^{76 80 81 82} In addition, studies have reported that CAD patients with high levels of hostility have a greater rate of restenosis after angioplasty,⁸³ experience more rapid atherosclerosis progression during serial carotid ultrasonography,^{84 85} and manifest more ischemia during stress testing than other CAD patients.⁸⁶

Pathophysiological Mechanisms
Hostility may affect atherogenic activity by behavioral mechanisms. Hostility is associated with a higher concentration of unhealthy lifestyle behaviors, including smoking, poor diet, obesity, and alcoholism.^{77 79 87} Hostile individuals are also more likely to manifest other psychosocial factors associated with CAD, such as social isolation.⁸⁸ An accumulating body of evidence also suggests multiple pathophysiological mechanisms by which hostility may be linked to CAD. For example, compared with nonhostile individuals, hostile subjects manifest higher heart rate and blood pressure responses to physiological stimuli, such as mental tasks,⁸⁹ as well as higher ambulatory blood pressure levels during daily-life activity.⁹⁰ Also, evidence suggests that hostile individuals are more likely to exhibit hypercortisolemia and high levels of circulating catecholamines,^{91 92} as well as diminished mononuclear leukocyte ß-adrenergic receptor function.⁹³ Preliminary data suggest that hostile individuals may also manifest diminished vagal modulation of heart function^{94 95} and increased platelet reactivity.^{96 97}

Social Isolation and Lack of Social Support
Since the late 1970s, a series of prospective community-based studies have examined the influence of social factors on the development of CAD. Initial studies focused on quantitative aspects of social support, such as the presence of family affiliations, number of friends, and the extent of one's participation in group and organizational activities. This domain of measurement has been called one's "social network." Within this domain, some studies evaluated the influence of partner status (living alone, marital status, and/or marital disruption), and others have assessed aspects of "instrumental" (ie, tangible) support, such as access to guidance and practical community services. Over time, however, the qualitative nature of one's social support system (eg, amount of perceived emotional support) has also been increasingly subject to study. Fifteen studies examining the impact of social factors on the future incidence of CAD in initially healthy populations are summarized in Table 4.^{98 99 100 101 102 103 104 105 106 107 108 109 110 111 112} A relatively small network has been found, on average, to be associated with a 2- to 3-fold increase in the incidence of CAD over time. Similarly, low levels of perceived emotional support confer an even greater increased risk for future cardiac events.⁹⁹ Table 5 lists 11 studies evaluating the relationship between social factors and prognosis in patients with preexisting CAD.^{113 114 115 116 117 118 119 120 121 122 123} Significant prognostic relationships are present in most of these studies, and the risk ratios are substantial. For instance, Berkman et al¹¹⁹ observed a nearly 3-fold increase in subsequent cardiac events in post-MI patients reporting a low level of emotional support, and Williams et al¹¹⁸ observed a similar 3-fold increase in mortality over 5 years among CAD patients who were unmarried or had no significant confidant in their life.

View this table: [in this window] [in a new window]   Table 4. Social Influences and the Pathogenesis of CAD in Healthy Subjects

View this table: [in this window] [in a new window]   Table 5. Social Influences and the Occurrence of Cardiac Events in CAD Patients

In addition to the consistency and magnitude of these findings, the cause-and-effect relationship between social factors and CAD development is also supported by other evidence. First, an inverse gradient has been reported between the magnitude of social support and the incidence of CAD and/or future cardiac events, as summarized in Table 6.^{98 100 101 103 119} Moreover, acculturation independently influences CAD development. For instance, in one study, 3809 Japanese-Americans in California were classified according to the degree to which they retained a traditional Japanese culture.¹²⁴ The most traditional group of Japanese-Americans had a CAD prevalence as low as that observed in Japan, whereas the group that was most acculturated had a 3- to 5-fold excess in CAD prevalence. Major CAD risk factors did not account for these differences. In another study,¹²⁵ temporal rates of CAD development were assessed in Roseto, Pa, and an adjacent town. Initially, CAD incidence was significantly lower in Roseto, despite shared medical resources. At that time, Roseto was a cohesive and homogeneous community of 3-generation households, descendants of Italian immigrants. As the distinguishing social characteristics of the Roseto community disappeared over time, its lower incidence of CAD vanished. Finally, animal studies have also implicated social factors in the promotion of atherogenesis. For instance, Ratcliffe and Cronin¹²⁶ described the potential importance of social disruption among animals, noting that crowding and social disruption were the apparent causal factors for a 10-fold increase in atherosclerotic lesions that occurred among birds and mammals over a 20-year period at the Philadelphia Zoo. Ratcliffe et al¹²⁷ also studied social support experimentally by deliberately assigning swine to various social situations (alone, pairs, groups). At postmortem examination, coronary arteriosclerosis was most advanced in isolated females, intermediate in isolated males, and least advanced in animals sustained in groups. More recently, the extent of atherosclerosis was compared at autopsy among 39 cynomolgus female monkeys exposed to 2 different housing conditions: 15 monkeys housed in single cages and 24 housed in groups.¹²⁸ The extent of atherosclerosis was 4 times greater, on average, in the females that were housed alone than in those housed in social groups. This difference occurred in the absence of significant differences in plasma lipids. In combination, these data provide strong evidence that social factors relating to grouping and isolation can promote atherogenesis.

View this table: [in this window] [in a new window]   Table 6. Social Support and Clinical Outcome: Evidence of a Gradient Effect1

Socioeconomic Status
Aside from social factors, low socioeconomic status is a significant contributor to increased risk in healthy persons and a contributor toward poor prognosis in patients with established CAD, in graded fashion.¹²⁹ The gradient between socioeconomic status and cardiac outcome is observable whether measured by education, income, or occupation. Low socioeconomic status is associated with increased levels of high-risk behaviors¹³⁰ and psychosocial risk factors.¹³¹ Interestingly, some evidence also suggests that low socioeconomic status may be an independent risk factor in its own right,^{118 132 133} but this possibility requires prospective validation and, if true, delineation of the operative pathophysiological mechanism(s). In any case, when low socioeconomic status is clustered together with other psychosocial risk factors, the risk of cardiac events is often magnified.^{129 134}

Pathophysiological Mechanisms
Like other psychosocial factors, social support influences the extent to which individuals engage in such high-risk behaviors as smoking, fatty diet intake, and excessive alcohol consumption. In addition, social factors may exert direct pathophysiological effects, including hypercortisolemia. Animal studies have reported an association between social isolation and hypercortisolemia^{135 136} and reversible increases in resting heart rates among cynomolgus monkeys, depending on the presence or absence of social separation.¹³⁷ Similarly, human studies have demonstrated an inverse relationship between the quality of social relationships and urinary levels of epinephrine¹³⁸ and between the degree of social support and resting heart rates.¹³⁹ Elevated resting heart rates may constitute a sign of altered autonomic arousal. The presence of social support may also attenuate blood pressure and heart rate responses to stressful stimuli in humans.^{140 141} In summary, these data suggest that social factors promote atherogenesis through activation of the autonomic nervous system.

Chronic and Subacute Life Stress
Work-related stress is the most widely studied chronic life stress relative to CAD. Although many aspects of one's work environment relative to the development of CAD have been studied, much interest has focused on models of inherent "tension" at work. One such model has been the "job strain" model, defined by Karasek et al¹⁴² as jobs with high demand but low decision latitude. In one prospective study of 1928 male workers followed up for 6 years, job strain was associated with a 4-fold increase in the risk of cardiovascular system–related death. Subsequent studies have supported the relationship between job strain and CAD risk,^{143 144} but negative studies have also been reported.^{145 146}

More recently, research has begun to focus on other forms of work-related stress. For example, one model views work stress as the outcome of high work demand and low reward.¹⁴⁷ This model both predicts cardiac events^{146 147} and has been correlated with progression of carotid atherosclerosis.¹⁴⁸ Also, low job control, per se, predicts future cardiac events.¹⁴⁹ Taken together, the studies regarding presence of stress at work and subsequent CAD development have been largely positive, suggesting a strong causal relationship between this form of chronic stress and development of atherosclerosis.

Because many observational studies have reported psychological prodromata in the months preceding development of acute MI,¹⁵⁰ interest has also been focused on the potential pathogenicity of "subacute" life stress (defined as an accumulation of stressful life events over a duration of months). In one of the earliest attempts to quantify the relationship between subacute psychological stress and CAD, Holmes and Rahe¹⁵¹ developed a "Recent Life Change Questionnaire," with a predetermined weighting assigned to different life events, ranging from high numbers for such events as the death of a spouse, divorce, or loss of a job to low weightings for vacations and holidays. In one study, marked elevations in Recent Life Change scores were seen for most cases of MI or sudden cardiac death during the 6-month period preceding these events.¹⁵² Similar confirmations of increased life stress before cardiac events can be extended to specific cohorts, ranging from healthy middle-aged men¹⁵³ to patients pre- senting with acute myocardial infarction.¹⁵⁴

Pathophysiological Mechanisms
Like other psychosocial factors, chronic stress appears to exert direct pathophysiological effects, including elevation of arterial blood pressure^{155 156} and neurohumoral arousal.¹⁵⁷ Evidence of neurohumoral arousal has also been noted in situations associated with subacute stress.^{158 159}

"Clustering" of Psychosocial Variables
Although psychosocial stresses have been reviewed here as individual entities, generally, these stresses tend to cluster together. When they do so, risk ratios for cardiac events often rise substantially. For example, in one study of post-MI patients, the presence of high levels of life stress and social isolation were each associated with an 2-fold increase in subsequent events.¹¹⁴ But when the 2 factors occurred together, the rate of subsequent events was 4-fold higher. A similar synergy between these 2 factors has also been reported among healthy individuals.¹⁶⁰ Similarly, the combination of anxiety and depression compounds cardiac risk in post-MI patients,¹⁶¹ and many other examples can be found within the psychosocial literature. These data indicate that psychological factors occurring in combination substantially magnify risk associated with individual psychological factors, resulting in risk elevations that are comparable to those associated with hypercholesterolemia, hypertension, and other major risk factors for CAD. Furthermore, psychosocial factors also interact synergistically with conventional CAD risk factors to heighten the risk for cardiac events. For example, depressed patients who smoke have a substantially higher risk of cardiac events than depressed patients who do not smoke.⁴ These findings could provide impetus for developing algorithms that integrate psychosocial factors and conventional risk factors into the Bayesian analysis of CAD, as schematized for a patient with nonanginal chest pain in Table 7.

View this table: [in this window] [in a new window]   Table 7. Bayesian Probability of CAD in a 47-Year-Old Man With NACP According to Risk Factors and Depressive Symptoms

Pathophysiological Mechanisms
From a pathophysiological point of view, the increase in cardiac events associated with clustering of psychosocial stresses suggests that this clustering compounds the health-damaging effects of individual psychosocial stresses. However, because psychosocial stresses and behavioral risk factors in humans change over time and cluster together in variable fashion, it is difficult to study the potential mechanisms by which they exert their pathophysiological effects. In contrast, adequate experimental control can be achieved in animal models, especially monkeys. In this regard, cynomolgus monkeys (Macaca fascicularis) provide a potentially relevant model for studying the interaction of multiple psychological factors in a controlled setting. Like humans, cynomolgus monkeys develop coronary atherosclerosis when fed fatty diets and manifest similarities to people in the development of coronary lesions and coronary vasodilator abnormalities. Notably, cynomolgus monkeys resemble human beings in both the organization and expression of their social behavior. For example, dominance and nurturance are sometimes considered to be the 2 major dimensions that define the content of interpersonal human behavior.¹⁶² Cynomolgus monkeys are characterized by well-defined social status hierarchies in which some animals (dominants) reliably defeat others (subordinates) in competitive interactions, as well as by elaborate, generation-spanning, networks of affiliation, alliance, and mutual support. Humans and monkeys also use similar facial expressions and postures to communicate an antagonistic or combative mood, and both rely extensively on visual cues to signal moods quickly and unambiguously in complex social settings. These behavioral similarities suggest that monkeys might be especially useful for modeling the human expression of anger or hostility.

In a set of investigations designed to evaluate the interaction between personality factors and a stressful social environment, 30 male monkeys were fed a moderately atherogenic diet while housed in 5-member social groups and assigned to 1 of 2 social conditions⁶⁷ : (1) an "unstable" environment in which animals were switched among groups on a regular basis so that animals periodically had to reestablish their dominance and affiliative relationships or (2) a "stable" environment in which initial group memberships were maintained without disruption throughout a 22-month period. Repeated behavioral observations permitted identification of individuals as relatively more dominant or subordinate in their social groups. The index of coronary artery atherosclerosis in this and all other cited monkey experiments was the average lesion extent as measured in 15 cross sections of pressure-perfused coronary arteries. Five sections each were taken from the left circumflex, left anterior descending, and right coronary arteries for these determinations. At the end of this study, quantitative evaluation of the coronary arteries of these animals revealed that dominant male monkeys in an unstable environment had significantly more coronary artery atherosclerosis than the other 3 subgroups (Figure 1, top). These results were independent of variations in serum lipid concentrations and blood pressure. Thus, this animal model revealed that it was the interaction between 2 psychosocial factors that proved pathogenic in cynomolgus monkeys: the trait of "dominance," coupled with environmental stress.

View larger version (35K): [in this window] [in a new window]   Figure 1. A, Measurements of coronary artery intimal extent (±SEM) among dominant and subordinate monkeys living in unstable (left) or stable (right) groups. Atherosclerosis was significantly more extensive in dominant, unstable animals compared with monkeys in other 3 conditions. B, Coronary artery atherosclerosis extent in dominant and subordinate monkeys, all housed in unstable groups and either untreated (left) or treated with propranolol HCl (right). Exacerbated atherosclerosis of dominant animals living in unstable groups was completely inhibited by ß-adrenergic blockade. A was adapted from Reference 67; B, from Reference 163.

Subsequently, to test the causal role of sympathetic activation in promoting atherogenesis in these predisposed animals, a number of male monkeys were housed in unstable social groups and fed an atherogenic diet for 26 months; however, half of the monkeys also received a ß-adrenergic antagonist, propranolol, throughout the study.¹⁶³ As shown in Figure 1, bottom, untreated dominant monkeys again developed substantial atherosclerosis; however, pretreatment with propranolol abolished the excess atherosclerosis that develops among dominants housed in an unstable environment. These data provide strong confirmatory evidence that the atherogenic effect of chronic psychological stress in these monkeys is dependent on concomitant sympathetic activation.

Cynomolgus monkeys have also been used to study the influence of chronic psychosocial stress on coronary endothelial integrity. First, it was demonstrated that the stress model cited above induces atherosclerosis in dominant male monkeys in the absence of hypercholesteremia, albeit with smaller lesions than those noted for monkeys concomitantly fed a high-cholesterol diet.¹⁶⁴ Thus, under conditions of chronic psychological stress, endothelial injury can occur even without dietary provocation. Subsequently, quantitative coronary angiography was used to demonstrate that psychosocial stress in cynomolgus monkeys can also lead to impairment of endothelial function in the presence of underlying coronary atherosclerosis.¹⁶⁵ Specifically, arterial responses in nonatherosclerotic controls (which always consumed a low-cholesterol diet and were housed in stable groups) were compared with those in monkeys that consumed a high-cholesterol diet for 1 year and were subsequently assigned to 1 of 3 experimental conditions: (1) continued consumption of a high-cholesterol diet plus exposure to periodic social disruption, (2) consumption of a low-cholesterol diet plus exposure to periodic social disruption, and (3) consumption of a low-cholesterol diet and housed in stable social groups. As shown in Figure 2, coronary vascular responses to acetylcholine differed across groups in a manner consistent with the exposure to psychosocial stress. Thus, chronic psychosocial stress can impair endothelium-dependent vascular responses in a manner that is not necessarily dependent on extent of underlying atherosclerosis or diet.

View larger version (17K): [in this window] [in a new window]   Figure 2. Mean coronary vascular responses during intracoronary infusion of acetylcholine (individual values and means) measured as percent change in diameter from control (y axis, left). Left bar represents vascular response in 6 monkeys consuming a low-cholesterol (CHOL) monkey chow and sustained in a nonstressed environment for full 36 months; this group served as control for this study. Other 3 groups of monkeys were experimental groups, all fed a high-cholesterol diet for first 18 months of experiment and then subjected to the following interventions for duration of experiment: (1) 9 monkeys continued on atherogenic diet and were housed in unstable social conditions (HI CHOL/UNSTABLE); (2) 8 monkeys consumed cholesterol-lowering diet and were housed in unstable psychosocial environment (LOW CHOL/UNSTABLE); and (3) 10 monkeys were fed cholesterol-lowering diet and placed into stable psychosocial environment (LOW CHOL/STABLE). Vascular responses to infusion of acetylcholine (mean±SEM) are represented by solid bars: vasodilatory responses are represented by increases above middle horizontal line and vasoconstrictor responses by decreases below this line. Average plaque sizes in 3 experimental groups are shown in gray bars. Notably, the 2 experimental groups in unstable social environment (2 middle bars) had similar coronary vasoconstrictor responses, despite different cholesterol diets and differing magnitudes of underlying atherosclerosis. By contrast, the 2 experimental groups fed a cholesterol-lowering diet for second half of experiment (last two bars) had opposite coronary vascular responses to acetylcholine infusion, despite similar diets and similar magnitude of coronary atherosclerosis. Thus, presence or absence of ongoing psychosocial stress in this study modulated endothelial function.

Sex Differences
The relative sparing of premenopausal women in relation to men of similar age is a prominent feature of CAD, ischemic stroke, and atherosclerosis.¹⁶⁶ Although this phenomenon is sometimes referred to as "female protection," it is more accurately characterized as a delay in disease onset, with the incidence curve for women lagging behind that of men by 10 years.¹⁶⁷ The various effects of estrogen are believed to account for most of this sex difference, at least with respect to CAD incidence and atherosclerosis.¹⁶⁸ Not only are premenopausal women "protected" from atherosclerosis, but the provision of estrogen replacement to initially healthy postmenopausal women is associated with a significant reduction in CAD risk.^{168 169} Nevertheless, because atherosclerosis progresses over decades, it is likely that the clinical events occurring in postmenopausal women have their beginnings in the premenopausal years. This conclusion is supported by a recent study showing the presence of relatively extensive focal atherosclerosis in premenopausal women.¹⁷⁰ Ovarian abnormalities or failure, by reducing the amount of endogenous estrogen, could accelerate atherogenesis in premenopausal women, thereby predisposing these individuals to CAD (and possibly ischemic stroke) in later years.

Notably, studies in premenopausal monkeys suggest that psychosocial stress reliably induces ovarian impairment in the half of socially housed females that occupy subordinate status in their social groups.^{171 172} Subordinate monkeys have estradiol concentrations of 60 pg/mL, contrasting significantly with dominants at 130 pg/mL.¹⁷³ Furthermore, these subordinate, ovary-impaired females, compared with their dominant counterparts, develop exacerbated atherosclerosis and abnormalities in coronary reactivity.^{171 173 174} Such females are also typically hypercortisolemic and have exaggerated heart responses to stress, characteristics that are themselves risk factors for atherosclerosis in both humans and monkeys.

These animal findings establish the possibility that behavioral stressors influence the development of CAD in women during the premenopausal period, through effects on estrogenic and neuroendocrine activity. In fact, several lines of evidence are consistent with the suggestion that ovarian impairment, possibly stress-induced, potentiates atherogenesis before menopause in women. First, 2 studies have linked a lifelong history of menstrual irregularity with a significantly increased risk for acute myocardial infarction.^{175 176} A third study has shown that irregularly menstruating women, in comparison with normally cycling control women, have elevated plasma fibrinogen concentrations (a risk factor for CAD) and a thickened arterial intima.¹⁷⁷ The possibility that these observations might reflect the reduced concentrations of endogenous estrogen characteristic of ovary-impaired women is supported by the finding that premenopausal women with angiographically confirmed CAD have significantly lower plasma estradiol concentrations than do control subjects and that such levels resemble those observed in subordinate female monkeys.^{173 178}

Notably, many premenopausal women may experience ovarian compromise at some time during their reproductive years.¹⁷⁹ The general term for this compromise is functional hypothalamic hypogonadism, the manifestations of which range from subclinical luteal-phase defects with regular menstrual intervals to irregular cycles to amenorrhea.^{179 180} Psychogenic stress is often linked to functional hypothalamic hypogonadism in women.^{172 181 182 183} One particular expression of this syndrome, functional hypothalamic amenorrhea, is associated with abnormal luteinizing hormone pulse generator activity and is accompanied by hypercortisolemia and other neuroendocrine and behavioral indicators of stress.^{184 185} Furthermore, recent data suggest that subclinical ovarian abnormalities sufficient to cause premenopausal bone loss may affect a substantial number of women.^{179 186} For example, disturbed luteal phase function characterized 29% of the menstrual cycles recorded from a sample of 66 premenopausal women thought to be cycling normally; decreases in spinal density at both 1 and 5 years after measurement were significantly associated with this degree of abnormality.¹⁷⁹ If ovarian hormones are indeed cardioprotective, women with functional hypothalamic hypogonadism and subclinical ovarian dysfunction could share with subordinate monkeys a predilection for accelerated atherosclerosis and an increased risk of CAD, especially because a relatively modest impairment of ovarian function is sufficient to cause marked exacerbation of atherosclerosis in monkeys.¹⁷²

Ovarian impairment probably has eluded detection as a risk factor for CAD because it is often occult and because premenopausal women have a low incidence of CAD.¹⁷⁴ Nonetheless, surrogate measures (eg, a history of menstrual irregularity) are associated with premature CAD or elevated CAD risk factors.^{176 177} Hence, the percentage of premenopausal women who experience accelerated atherosclerosis may be much larger than the number diagnosed as amenorrheic or otherwise ovary-impaired. At present, however, atherosclerosis progression has not been studied prospectively in premenopausal women in conjunction with ovarian function.

Acute Life Stress

Anecdotal reports and case studies^{187 188 189} have long reported a relationship between acute stress and the development of cardiac disease. In addition, the effects of acute stress on heart disease are well supported by epidemiological studies regarding natural life stressors. An acute stressor associated with increased rates of cardiac events is bereavement.^{190 191} For example, in one study of 95 647 individuals followed up for 4 to 5 years, the highest relative mortality occurred immediately after bereavement, with a >2-fold higher risk for men and 3-fold higher risk for women.¹⁹⁰ After the first month, mortality rates returned to normal population levels. Cardiac event rates also increased in the immediate aftermath of other acute life stressors, such as earthquakes and terrorist activities. For instance, during the massive Los Angeles earthquake of 1994, the number of sudden cardiac deaths due to CAD rose sharply, from a daily average of 4.6 in the preceding week to 24 on the day of the earthquake.¹⁹² Similarly, there was also a sharp increase in the number of deaths on the first day of missile strikes on Israeli cities during the Gulf War of 1991.¹⁹³ Finally, a retrospective interview format has been used to examine the effect of anger as an acute trigger of myocardial infarction among 1623 post-MI patients.¹⁹⁴ A 7-point self-report anger scale (1=calm, 7=enraged) was used, with anger episodes defined as scores 5. After an episode of anger, the relative risk of myocardial infarction was increased >2-fold.

Pathophysiological Mechanisms
In contrast to chronic stress, acute stress is easier to model and can be studied under controlled laboratory conditions in both humans and animals. In recent years, as increasingly sophisticated techniques, such as radionuclide imaging techniques and the measurement of coronary endothelial function, have been applied to the laboratory study of acute stress, an understanding has emerged as to how such acute stress causes deleterious effects in CAD patients, as schematized in Figure 3 and elucidated below.

View larger version (39K): [in this window] [in a new window]   Figure 3. Schematic of pathophysiological effects of acute psychosocial stress. Sympathetic nervous system (SNS) stimulation emanating from acute stress leads to a variety of effects, ranging from heart rate and blood pressure stimulation to direct effects on coronary vascular endothelium. Clinical consequences of these effects include development of myocardial ischemia, cardiac arrhythmias, and fostering of more vulnerable coronary plaques and hemostatic changes. These changes form substrate for development of acute myocardial infarction and sudden cardiac death.

Induction of Myocardial Ischemia
The ability of acute psychological stress to induce myocardial ischemia has been assessed in the laboratory with modeled forms of stress (eg, mental arithmetic and speaking tasks) and sensitive imaging techniques able to detect the extent and severity of mental stress–induced myocardial ischemia. These techniques include those used to measure left ventricular function (radionuclide ventriculography^{195 196 197 198 199} [Figure 4], echocardiography,²⁰⁰ assessment of left ventricular changes by either a stationary probe²⁰¹ and ambulatory VEST²⁰² ) and those used to assess myocardial perfusion (positron emission tomography²⁰³ and ^99mTc-sestamibi myocardial perfusion tomography²⁰⁴ ). Approximately half of CAD patients with exercise-induced myocardial ischemia also manifest inducible ischemia during mental stress testing in the laboratory, as identified by these techniques. Mental stress–induced ischemia, however, is not common among CAD patients without exercise-induced ischemia. Mental stress–induced ischemia is usually electrocardiographically and clinically "silent" and generally occurs at relatively low heart rate elevations compared with exercise testing. In addition, the frequency and magnitude of mental stress–induced ischemia varies according to the type of mental stressor. Specifically, stress that is more emotionally laden and/or personally relevant, such as a speaking assignment concerning personal faults, results in a significantly greater frequency and magnitude of inducible left ventricular wall motion abnormalities than does more nonspecific mental stress, such as the performance of mental arithmetic or the Stroop Color-Word task.¹⁹⁵ Because the mean heart rate increases of 15 to 20 bpm during laboratory-modeled public speaking²⁰¹ are far less than those of real-life speaking experiences,^{205 206} this and similar tasks may underestimate the potential potency of mental stress in certain real-life situations. Recall of angry events is also a laboratory trigger of myocardial ischemia,¹⁹⁶ which supports the epidemiological study of anger by Mittleman et al.¹⁹⁴

View larger version (37K): [in this window] [in a new window]   Figure 4. Comparative still-frames of left anterior oblique scintigrams in patient who underwent radionuclide ventriculography at rest and then during a mental stress task involving a speech task given in imaging laboratory. Patient had worsening of left ventricular segmental wall motion while speaking about feelings of personal stress concerning his problems in caring for his family. Images shown for rest are on top and those for the speaking task on bottom. Shown are end-diastolic (ED) images (left), end-systolic (ES) images (middle), and superimposed ED and ES edges (right). During speech, frank dyskinesis (abnormal outward motion during systole) developed in septum. From Reference 195.

These laboratory studies are complemented by ambulatory ECG studies that demonstrate an association between psychological stress and/or negative emotions and the occurrence of myocardial ischemia during daily-life circumstances (Figure 5).^{207 208 209} Like ischemia induced during mental stress testing in the laboratory, transient ischemic episodes out of hospital are overwhelmingly silent and occur at relatively low heart rate elevations.²¹⁰ Furthermore, patients who manifest myocardial ischemia during laboratory mental stress are also more likely to manifest myocardial ischemia during ambulatory ECG monitoring of daily-life activity.^{200 209 211} Just as CAD patients who demonstrate myocardial ischemia during daily-life activity manifest a significantly increased likelihood of subsequent cardiac events,²¹² recent studies also suggest that mental stress–induced ischemia in the laboratory setting also predicts cardiac events (Figure 6).^{197 198}

View larger version (21K): [in this window] [in a new window]   Figure 5. A, On x-axis, feelings of two positive emotional states, being in control (far left) and happiness, and 3 negative emotional states are represented. y-axis represents percent of hours with ischemia during 2760 hours of monitoring among 58 patients with CAD. Low levels of each state are represented by gray bars and high levels by solid bars. Percentage of ischemic hours increased during negative emotional states. Relative risk ratios for ischemia during high vs low levels of tension, frustration, and sadness, respectively (after adjustment for time of day and physical activity level) were as follows: 2.2 (1.1–4.5), 2.2 (0.7–6.4), and 2.2 (1.1–4.3). Control hours are all nonischemic hours during 48 hours of monitoring. Adapted from Reference 209. B, Linear relationship between magnitude of sadness and duration of myocardial ischemia during ambulatory ECG monitoring was also noted. Relative risk (95% CI) for ischemia by level of sadness was as follows: 1.8 (1.0–3.4) for level 1; 2.3 (1.0–5.3) for level 2; 3.4 (1.0–11.7) for level 3; and 2.4 (0.3–16.6) for level 4.

View larger version (20K): [in this window] [in a new window]   Figure 6. After performance of mental tasks during radionuclide ventriculography, patients were followed up over time (x-axis). Probability of event-free survival is plotted as a function of mental stress–induced EF change plotted at 2 prototypical values, 1 SD below (EF change=-12.40%) and 1 SD above (EF change=+1.05%) mean of entire sample (EF change= -6.73%). Curves are adjusted for baseline EF, history of myocardial infarction, and age. Relative risk associated with lower curve compared with higher curve is 2.40 (P=0.024). From Reference 197.

Mechanisms for mental stress–induced myocardial ische- mia.

Even though heart rate elevations during laboratory-induced mental stress are relatively small, blood pressure elevations during mental stress are substantial, paralleling those noted with exercise.^{195 200} Thus, oxygen demand is increased during mental stress testing. However, because the double product threshold for the induction of ischemia during mental stress testing is substantially lower than that associated with exercise testing, other mechanisms must also be involved. One mechanism is mental stress–induced coronary vasoconstriction.^{213 214} During mental stress, significant coronary vasoconstriction may occur in CAD patients at sites manifesting vasoconstriction during acetylcholine infusion.²¹³ Because acetylcholine is used to test for endothelium-dependent coronary vasoconstriction, these findings suggest that neurohumoral stimulation during mental stress induces coronary vasoconstriction through an endothelium-dependent mechanism. In addition, it has been reported that the coronary microcirculation fails to dilate during mental stress.²¹⁵ However, significant coronary vasoconstriction also occurs during exercise in CAD patients²¹⁶ through the same endothelium-dependent mechanism as noted for mental stress. Thus, other mechanisms must be considered to explain why mental stress ischemia is induced at relatively low double-product thresholds compared with exercise test thresholds for ischemia.

One potential factor may relate to the presentation of stress. Mental stress testing in the laboratory is a "sudden" stressor, presenting without warm-up. Maximal heart rate and blood pressure responses are generally observed at the near onset of mental stress, and ischemic abnormalities are induced relatively rapidly during laboratory mental stress.²⁰¹ By contrast, laboratory exercise is always presented to patients in a graded fashion. However, when healthy individuals were exercised to a high workload without warm-up, ST-segment depression was induced in a significant number of subjects.²¹⁷ Similarly, Eschar et al²¹⁸ compared 6 CAD patients during graded versus sudden exercise stress; during the graded stress protocol, 3 patients had the indication of chest pain, and only 1 patient developed exercise-induced ST-segment depression of 1 mm. By contrast, during sudden stress, all 6 patients both had the induction of chest pain and developed >2 mm of ST-segment depression. Because CAD patients may commonly experience short bursts of strenuous physical and mental stress without warm-up during daily life, further study of this issue is warranted.

Autonomic factors may also be operative in regulating myocardial ischemia. For instance, myocardial ischemia during daily life shows a characteristic circadian rhythm.²¹⁹ Even though exogenous factors (ie, the amount of physical and mental activity) trigger episodes of myocardial ischemia,^{208 209 219} Krantz et al²¹⁹ demonstrated that the circadian rhythm of myocardial ischemia also appears to have an endogenous component, as demonstrated in Figure 7. Understanding the factors driving this apparent endogenous component could help further elucidate the pathophysiology of mental stress–induced ischemia.

View larger version (27K): [in this window] [in a new window]   Figure 7. Figure 7 A, Hour-by-hour circadian variation of myocardial ischemia in 63 CAD patients manifesting myocardial ischemia during ambulatory ECG monitoring. Left, Total ischemic time (minutes) in each hour of day; right, number of ischemic episodes in each hour. There was a morning peak in ischemic time/hour between 6 and 11 AM, with a secondary rise between 2 and 6 PM, before tapering of ischemia in evening. ANOVAs revealed a significant overall temporal variation in ischemic time during course of day. A similar pattern was noted for number of ischemic episodes. B, To assess contribution of an endogenous component to diurnal variation of ischemia, hourly duration of ischemia was controlled for concomitant activities, according to a statistical formula that regressed maximal physical and mental activity levels for each hour on corresponding ischemic time. This yielded "activity-adjusted residual scores," which reflect measurements of ischemia not accounted for by activity. Shown are 3 separate curves for adjusted measurements of predicted ischemic time per hour of day, including those adjusted for (1) physical activity, (2) mental activity, and (3) heart rate. Adjusted ischemic times >0 on y axis indicate ischemic times that were longer than expected on the basis of concurrent activity levels or heart rate. Conversely, values <0 indicate ischemic times that were shorter than expected. Thus, if there had been no ischemic effect independent of activity, flat curves would be present. Rather, analyses showed a significant diurnal variation for all 3 curves (P<0.01), proving presence of a circadian rhythm that is endogenous in nature. Further statistical analysis of hour-by-hour changes in ischemic time revealed that for each curve, there was a significant increase in ischemic time at 6 AM vs 5 AM. Afternoon peak in ischemia noted in A, however, is now abolished, suggesting that this afternoon rise in ischemia is of an exogenous nature, driven by physical and mental activity levels. From Reference 220.

Other factors that may potentially help regulate myocardial ischemia may be elucidated by the evolving study of factors that modify hemodynamic responses to exercise and mental stress. This includes the study of systems modulating systemic vascular tone, such as the endothelial system. In the Psychophysiological Investigations of Myocardial Ischemia (PIMI) Study, systemic vascular resistance increased during mental stress and decreased during exercise testing among CAD patients.¹⁹⁹ In that study, increased systemic vascular resistance during mental stress testing was reported to be the most significant hemodynamic feature associated with mental stress–induced myocardial ischemia. Interestingly, in a subsequent recent study, increases in systemic vascular resistance during mental stress were directly related to compromised peripheral endothelial function in a group of 40 healthy men and women.²²⁰ These data raise the possibility that excessive systemic vascular resistance responses to mental stress may be a potential marker of peripheral endothelial dysfunction. However, prospective work is clearly indicated to determine the reproducibility and validity of these new observations and their potential pathophysiological import.

Promotion of Arrhythmogenesis
Investigators have consistently noted an interrelationship between behavioral factors and arrhythmogenesis in humans.^{50 221 222 223 224} According to Lown et al,²²¹ 3 sets of conditions contribute to the occurrence of such arrhythmias: (1) myocardial electrical instability, most often due to CAD; (2) an acute triggering event, frequently related to mental stress; and (3) a chronic, pervasive, and intense psychological state, often including depression and hopelessness. To study the pathophysiological interrelationship between behavioral factors and ventricular fibrillation, Verrier et al²²⁵ assessed myocardial electrical instability in conscious, freely moving dogs by placing a catheter in the right ventricular apex, scanning the ECG, and delivering repeated electrical stimulation during the vulnerable part of the cardiac cycle. The amount of electrical stimulation required to produce repetitive extrasystoles (REs) was used to define the ventricular fibrillatory threshold. Inasmuch as the animal does not perceive this stimulation, investigators can more directly study the effects of experimentally produced behavioral states on the ventricular fibrillatory threshold. In an initial series of studies, dogs were exposed to either an undisturbed environment or one in which the animal was held in a sling and given periodic transthoracic shocks over 3 successive days.⁵⁰ Behaviorally, the animals seemed relaxed in the undisturbed environment and agitated and autonomically aroused in the sling. The RE threshold was reduced by >40% when animals were moved from the benign to the stressed environment. Similarly, a natural emotion, an anger-like state provoked by denial of access to food, similarly reduced the RE threshold.²²⁵

Taken together, these studies show that behavioral stress, whether produced by aversive conditioning or a more naturalistic conflict, significantly decreases the electrical stability of the heart. Furthermore, ß-adrenergic blockade prevents the effects of either aversive conditioning or induced anger on the RE threshold, suggesting that these effects are mediated, in part, by sympathetic arousal.²²⁶ Other work demonstrated that when dogs were first predisposed to arrhythmia by acute myocardial ischemia (a 10-minute period of coronary artery occlusion followed by reperfusion), exposure to the stress of an aversive sling environment significantly increased the incidence of ventricular fibrillation.²²⁵ In general, stimuli that elicit anger-like responses are especially likely to provoke abnormalities in rhythm.²²⁵ Researchers have also related behavioral factors to arrhythmia in other experimental animals.²²⁷ Thus, the data relating behavioral factors to arrhythmias are impressive in reliability of the effects and in the identification of excessive sympathetic activation as a major precipitating factor.

Deleterious Endothelial Effects
In animal model studies, acute stress also causes coronary endothelial abnormalities, which range from endothelial dysfunction to frank endothelial injury and necrosis. For example, borderline hypertensive rats exposed to air-jet stress in the face (2 h/d for 10 days) display impaired arterial dilation in response to acetylcholine.²²⁸ Furthermore, the stressed animals also have a reduced sensitivity to nitroprusside, indicating an attenuated response to exogenous nitric oxide. The authors concluded that behavioral stress impairs endothelium-independent and nitric oxide–mediated coronary relaxation, but without causing visible endothelial damage (as evaluated by scanning electron microscopy). With respect to endothelial injury, investigators have directly tested the hypothesis that sympathetic stimulation alters endothelial integrity by exposing rabbits to chloralose anesthesia, a manipulation that produces persistent and reproducible increases in heart rate, blood pressure, and plasma norepinephrine concentrations (all indicators of sympathetic activation).^{229 230} Compared with conscious controls, chloralose-treated animals developed marked endothelial injury (as indicated by IgG incorporation) at both unbranched and circumostial aorta. In contrast, pretreatment with a ß-blocking agent attenuated heart rate and blood pressure increases, and endothelial injury was completely inhibited. Finally, male monkeys exposed to a clear-cut acute psychological stressor, 72 hour introduction to social strangers, had a significantly higher frequency of IgG-positive (injured) endothelial cells in the circumostial areas of the descending thoracic aorta than did control animals not exposed to this stressor.²³¹ Again, pretreatment with ß-adrenergic blocking agents prevented this stress-induced arterial damage (Figure 8). Importantly, the animals in this study consumed a low-cholesterol diet, indicating that the endothelial response to behavioral stimulation and adrenergic blockade was independent of dietary stimulation. Studies such as these clearly link behavioral factors and neuroendocrine activation to disruption of endothelial integrity and thus the earliest stages of atherosclerosis. On the basis of these animal data, it is reasonable to ask whether acute or subacute psychological stressors might similarly induce transient endothelial dysfunction in human beings as well. If so, it could help explain why prodromal psychological symptoms frequently precede the occurrence of acute myocardial infarction.

View larger version (27K): [in this window] [in a new window]   Figure 8. Frequency of injured endothelial cells in circumostial areas of descending thoracic aorta (y axis). Percentage of IgG-positive cells on Hautchen preparations was taken as index of arterial injury. Psychosocial stress (72-hour exposure of male monkeys to social strangers) caused a significant (P<0.02) increase in number of injured cells in unmedicated animal group (controls) (left bars). ß-Blockade, either with metoprolol (lipophilic) or atenolol (hydrophilic) (right bars) significantly (P<0.01) inhibited injurious effect of stress. 95% CIs are in parentheses. Adapted from Reference 231.

Coagulation Effects
The ability of emotional stress to induce coagulation abnormalities in human beings has been known for many years.^{232 233} Recently, the effects of acute psychological stress on indices of coagulation have been studied during laboratory mental stress tasks^{234 235} and during various naturalistic stressors.^{236 237} In both situations, significant, but not always consistent, platelet abnormalities have been observed. Acute laboratory mental stress also causes hemoconcentration through stress-induced decreases in plasma volume, as observed in various studies.^{238 239} This latter finding is significant in light of recent observations linking blood viscosity to cardiac events.²⁴⁰ Animal studies further confirm the ability of acute stress to induce coagulation abnormalities.²⁴¹

The duration of coagulation abnormalities induced by acute natural stresses is not yet known. However, because blood samples were fortuitously obtained in 42 hypertensive patients before the large Hanshin-Awaji earthquake, investigators obtained repeat blood samples 7 to 14 days afterward to assess the effects of this acute stressor on coagulation parameters.²³⁶ The earthquake induced transient increases in blood pressure; blood viscosity determinants, such as hematocrit and fibrinogen levels; and various hemostatic factors, including fibrin turnover and plasmin-₂–plasmin inhibitor complex, an activation marker of fibrinolysis. These parameters returned to normal levels by 4 to 6 months after the earthquake. These findings suggest that coagulation abnormalities may persist for weeks after a single stressful event. A more chronic hypercoagulable profile has also been reported among individuals subjected to forms of subacute stress.^{242 243}

It is not yet clear whether mental stress–induced hemoconcentration and mental stress–induced platelet activation are the result of the same or different mechanisms. A recent laboratory study, however, suggests that different mechanisms may be operative,²⁴⁴ because platelet activation in this particular study correlated only with changes in serum catecholamine levels during mental stress, whereas mental stress–induced hemoconcentration correlated only with changes in mean arterial blood pressure, but not serum catecholamine levels, during mental stress.

Sympathetic Nervous System Hyperresponsivity

Sympathetic nervous system hyperreactivity (also called cardiovascular reactivity) has been defined as a dispositional tendency to exhibit exaggerated heart rate and blood pressure responses when encountering behavioral stimuli experienced as engaging, challenging, or aversive. On a theoretical basis, it has been postulated that individuals manifesting more elevated heart rate and blood pressure responses to such physiological challenge (ie, "hot reactors") may experience more substantial sympathetic nervous system responses over time than "cold reactors" and that this may in turn promote the development of atherosclerosis. The first findings to support this hypothesis were reported by Keys et al,²⁴⁵ who evaluated 20 clinical variables in 275 men followed up for 20 years. Among these clinical variables, the diastolic blood pressure response to cold pressor stimulation was the most predictive variable for future CAD development. Subsequently, however, negative prognostic studies were also reported.^{116 246} In a recent development, the potential atherogenic effect of sympathetic hyperresponsivity has been assessed in 4 studies that used serial carotid Doppler measurements.^{247 248 249 250} In each study, progression of carotid atherosclerosis was more rapid among individuals manifesting more pronounced heart rate and blood pressure responses to physiological challenge.

Cynomolgus monkeys consistently display significant individual differences in heart rate response dur