| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
REVIEWS |
From the Department of Epidemiology and Public Health, University College London, UK.
Address correspondence and reprint requests to Andrew Steptoe, DPhil, Department of Epidemiology and Public Health, University College London, 1–19 Torrington Place, London WC1E 6BT, UK. E-mail: a.steptoe{at}ucl.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
Method: Systematic review of the published literature from 1970 to 2004 of trigger events, defined as stimuli or activities occurring within 24 hours of the onset of acute coronary syndromes.
Results: There is consistent evidence that physical exertion (particularly by people who are not normally active), emotional stress, anger, and extreme excitement can trigger acute myocardial infarction and sudden cardiac death in susceptible individuals. Many triggers operate within 1 to 2 hours of symptom onset. There are methodologic limitations to the current literature, including sampling, retrospective reporting, and presentation biases, the role of memory decay and salience, and reverse causation because of silent prodromal events.
Conclusions: Behavioral and emotional factors are probable triggers of acute coronary syndromes in vulnerable individuals, and the pathophysiological processes elicited by these stimuli are being increasingly understood. The benefits to patients of knowledge to these processes have yet to accrue.
Key Words: acute coronary syndromes • myocardial infarction • sudden death • stress • triggers • physical exertion
Abbreviations: ACS = acute coronary syndromes; CHD = coronary heart disease; EKG = electrocardiogram; IL-6 = interleukin 6; MI = myocardial infarction; PAI-1 = plasminogen activator inhibitor-1; SCD = sudden cardiac death.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
What Is a Trigger?
A trigger can be defined as an external stimulus, emotional state, or activity that produces acute physiological or pathophysiological changes leading directly to onset of acute cardiovascular disease. There is no general agreement about how long before the onset of symptoms an activity can take place to be regarded as an acute trigger rather than a more general etiologic factor (7–9). Trigger studies typically assess activities in the period ranging from a few minutes to 24 hours before ACS onset. Most recently, attention has focused on a 1- to 2-hour period before the onset of symptoms. Triggering of ACS typically takes place against a background of long-term coronary artery disease and coronary heart disease (CHD). The behavioral and psychosocial factors that influence the development of CHD in the long-term such as cigarette smoking, inactivity, work stress, social isolation, anxiety, and depression may be different from those that act acutely as triggers of events in vulnerable individuals.
Acute coronary syndromes comprise ST segment elevation myocardial infarction (MI), nonST segment elevation MI, unstable angina, and sudden cardiac death (SCD). NonST segment MI and unstable angina are generally grouped together for the purposes of clinical management (10). The majority of work published to date has examined acute MI and SCD with little research on unstable angina.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
Studies of personal experiences such as physical exertion or emotional stress are more common. Unfortunately, triggering is a phenomenon that is difficult to investigate prospectively, so data collection is typically retrospective and susceptible to memory loss, social acceptability bias, and to patients' private beliefs about the causes of heart disease. Clinical studies of the triggering of fatal cardiac events are difficult, unless the circumstances are witnessed by others, and the triggers of fatal events may be different from those for nonfatal ACS. One method that avoids problems of retrospective bias is to couple behavioral diaries with ambulatory monitoring of the electrocardiogram (EKG) for the detection of transient ischemia or dysrhythmia. This technique has documented associations between emotional states and myocardial ischaemia (11–13), and similar procedures have been used in patients with implanted defibrillator devices to link activation of the defibrillator by dysrhythmia with stress and anger (14). Diary methods are suitable for the investigation of relatively frequent subclinical episodes of cardiac dysfunction but would be difficult to apply to ACS themselves.
Overall Incidence of Triggers
Several large-scale studies in the early 1990s evaluated the incidence of triggers of acute MI by interviewing patients soon after their hospital admission. These studies have differed in measures, the interval between symptom onset and assessment of triggers, and in their national and cultural context, and this may account for some of the differences in the prevalence of possible triggers that have been reported. The Multicenter Investigation of Limitation of Infarct Size (MILIS) study interviewed 849 acute MI patients within 18 hours of symptom onset, of whom 48.5% reported a possible trigger (15). Emotional upset and moderate physical activity were the most common triggers, followed by lack of sleep and overeating, and 13% described multiple possible triggers. The pilot phase of the Triggers and Mechanisms of Myocardial Infarction (TRIMM) study involved 224 acute MI patients who were interviewed an average of 16.8 days after admission (16). Possible acute triggers were noted by 67%, with 52% reporting either emotional upset or stress within the 24 hours before infarction. By contrast, only 10% of 1818 patients in the Secondary Prevention Reinfarction Israeli Nifedipine (SPRINT) study reported possible external triggers of acute MI (17). Exceptionally heavy physical work, quarrels at work or home, and unusual mental stress were the most commonly mentioned triggers. In a sample of 1480 patients with anterior or inferior acute MI admitted to the hospital in Split, Croatia, the frequency of triggers was 44%, with physical exertion, emotional stress, cold and wet weather, and overeating being most common (18). Other large-scale studies such as the Determinants of Myocardial Infarction Onset Study (Onset), the Thrombolysis in Myocardial Infarction Phase II (TIMI II) study, and the onset study within the Stockholm Heart Epidemiology Program (SHEEP), have also interviewed large samples of patients soon after admission (19–21); however, they have not presented results for the overall incidence of triggers, but rather the occurrence of specific types of triggers, as detailed later in this review.
Younger patients, nondiabetics, and men were more likely to report triggers in the MILIS, SPRINT, and the Croatian studies (15,17,18). Reports of triggers were more common for MIs occurring during the hours of 6:00 am to 6:00 pm than between 6:00 pm and 6:00 am (16,19). The MILIS study identified no link between site of infarct and triggers, and a trend toward a lower likelihood of reporting physical or emotional stress as triggers (p = .08) for patients taking beta-blockers (15). Beta-blocking drugs have also been shown to attenuate the circadian and seasonal peaks of ACS (3).
Challenges to the Interpretation of Trigger Results
Two major problems limit the interpretation of studies that simply inquire about the occurrence of triggers. The first is that patients' reports may be influenced retrospectively by their attempts to make sense of their predicament. Studies of illness representations indicate that patients develop causal models of heart disease, with stress and physical activity featuring high on the list of presumed causes (22). By the time that individuals are interviewed, these views may be established and will influence responses to inquiries about the triggering period. The second problem is that control groups or control time periods are not tested. Physical exertion, emotional stress, and other factors may occur frequently in patients' lives, so their association with the onset of ACS may be coincidental. For example, the observation that triggered MIs are more common between 6:00 am and 6:00 pm than at other times of day could be because triggers themselves are more frequent between those hours. Case–control methods can be used, but as Maclure and Mittleman (23) have pointed out, there are difficulties in identifying appropriate control groups. General population controls have the limitation not only that healthy individuals will be most likely to participate, leading to a healthy volunteer bias, but also that control people are less likely to agree to be assessed on a stressful day. These factors may bias the comparison. Data from individuals hospitalized for other emergencies can be used but will be compromised by whatever caused their medical problems.
The investigators in the Onset study therefore developed the case–crossover design, in which the critical time periods are compared with control time periods on a within-subject basis (23,24). This method involves questioning patients about potential triggers during the hazard period and then for control periods. For example, an ACS patient might be questioned about emotional stress in the 2 hours before symptom onset and then about the corresponding 2 hours that occurred 24 hours earlier (pair-matched interval approach) or another time period such as the last week or the previous 12 months (usual frequency method). If a patient is habitually emotionally stressed, then he or she will report stress for both time periods. By comparing hazard and control periods, the relative risk that an episode of emotional stress is followed by an ACS can be calculated. This method has many advantages in the analysis of transient exposures, because self-matching eliminates selection and individual reporting biases. Any differences in chronic cardiovascular risk profile between cases and controls are eliminated, reducing the risk of residual confounding. Several of the results described in this review have used this method. The incidence of triggers in observational studies that have lacked comparison periods may be different from the rates obtained with these more elaborate statistical methods.
Behavioral Triggers
Physical Exertion
Physical exertion has an apparently paradoxic association with the triggering of ACS; people who are habitually physically active and are physically fit are at reduced risk (25), but vigorous activity has been found acutely to trigger both MI and SCD. The proportion of patients who reported moderate or vigorous physical activity before acute MI was 23% in the MILIS study, 18.7% in TIMI-II, and 27.1% in a case series in New Zealand (15,19,26). The levels of heavy exertion before acute MI were 4.4% in the Onset study (20) and 7.1% in the main TRIMM study in Augsberg, Germany (27). Using case–crossover methods, the relative risk of having engaged in vigorous activity in the hour before cardiac events has been estimated at 5.9 (95% confidence intervals [CIs], 4.6–7.7) and 2.1 (CI, 1.6–3.1) in different samples (20,27). In the SHEEP study, the relative risk was 3.3 (CI, 2.4–4.5) for physical exertion in the hour before acute MI, but this increased to 6.1 (CI, 4.2–9.0) in those who had no premonitory symptoms (21). There is also a strong protective effect of regular exercise (20). In the TRIMM cohort, the relative risks was 6.9 among those who exercised fewer than 4 times a week compared with 1.3 for those who exercised above this level (27).
Vigorous activity is a trigger for SCD as well as nonfatal events. An analysis was carried out of deaths in the Physician's Health Study using case–crossover comparisons with habitual exercise levels measured previously by questionnaire (28). Of the 122 SCD recorded over a 12-year period, 23 were preceded by vigorous exertion within the previous 30 minutes. The relative risk was 10.9 (CI, 4.5–26.2) in those who exercised regularly, increasing to 74.1 (CI, 22.0–249) in those who were sedentary.
These studies therefore show a consistent pattern despite the variation in relative risks. Heavy exertion is an acute trigger for ACS, particularly in otherwise sedentary people. It should also be noted that the absolute risk of nonfatal ACS or SCD during physical exertion is very low. In the Physicians' Health Study, it was estimated that the absolute risk after any individual bout of physical activity was 1 death per 1.51 million episodes (28).
Several studies indicate that triggering by exertion is more common in men than women (19,26,29), but this may be in part because middle-aged men engage in more leisure-time physical activity than women. There is some evidence that individuals who present with exertional triggering have less severe clinical risk profiles than others. In the TIMI-II study, exertional triggering was associated with not smoking, absence of recent chest pain, and fewer stenosed vessels than in other patients (19). In the New Zealand case series, the physical exertion-triggering group was younger and was less likely to have a history of cardiovascular disease than others (26). They also had less heart failure and a lower inhospital mortality risk. By contrast, Giri et al. (29) observed that patients with exertional onset were more likely to be smokers, hyperlipidemic, have heart failure, single-vessel disease, and a large thrombus in the infarcted artery. However, these data were collected from patients referred to a tertiary-care hospital, so they may not be as representative as other cohorts.
Space prevents detailed discussion of underlying mechanisms. However, the majority of exertion-related deaths are associated with plaque rupture. An autopsy study of 141 men with advanced coronary artery disease who died suddenly showed plaque rupture in 68%, compared with 23% in those dying at rest, and hemorrhage into the plaque was also frequent (30). Myocardial ischemia during exercise testing is a common phenomenon, and physical activity also triggers ischemia during everyday life in patients with coronary artery disease (11). Using case–crossover methodology for analyzing asymptomatic (silent) electrocardiographic (EKG) ischemia during ambulatory monitoring of patients with coronary artery disease, Gullette et al. (12) reported a relative risk of 13.2 (CI, 7.4–23.6) for heavy physical activity in the hour before onset after adjusting for time of day. Both ischemia on treadmill testing and ambulatory EKG ischemia predict cardiac events prospectively (31).
It is important to bear in mind that physical activity may coincide with periods of acute mental stress, especially in traumatic settings (natural disasters, assault, and so on). Depression and poor emotional well-being are inversely associated with regular physical activity (32). It is possible that people under stress or with depression are less likely to exercise regularly and so would be more at risk from the effects of sudden unaccustomed physical exertion. The influence of these different triggers may be difficult therefore to disentangle in many clinical cases.
Sexual Activity
The role of recent sexual activity as a trigger of acute MI has been studied in both the Onset and SHEEP studies, with similar results. In the Onset study, sexual activity was reported within 2 hours of MI by 3.0%, with a relative risk using the case–crossover method of 2.5 (CI, 1.7–3.7) (33). In the SHEEP study, 1.3% reported sexual activity over this time period, with a relative risk of 2.1 (CI, 0.7–6.5) (34). The absolute risk was very low, estimated at approximately 1 in a million. In both studies, risk was greater in patients who were sedentary, suggesting that the mediating mechanisms may be similar to those operating with physical exertion. Studies using ambulatory EKG monitoring indicate that heart rate increases to 118 to 127 beats per minute during sexual intercourse (35). In a study of 88 men with documented coronary artery disease, EKG signs of myocardial ischemia occurred in 31% during sexual intercourse, and in 78% of these individuals, ischemia was silent (asymptomatic) (35).
Sexual activity is often carried out as part of a loving intimate relationship. Because social support and good marital relationships are strongly protective for cardiovascular and other health end points, the benefits of sexual activity may offset costs (36). There is limited evidence that abstinence from or lack of sexual activity is associated with increased risk of all-cause and cardiovascular mortality (37), although confounding with other risk factors may account for this relationship.
Sleep Disturbance
A substantial number of cases of ACS and SCD occur during the night when patients are asleep. Sleep is associated with dynamic changes in autonomic tone, neuroendocrine function, and inflammatory cytokine release. The occurrence of cardiac events during sleep is not uniform, with a high frequency of acute MI, SCD, and discharge from implantable cardioverter–defibrillators early in the night, followed by a trough before waking (38). Whether there are triggers for ACS in the night is not known, because no studies have been reported concerning acute sleep disturbances on the night before onset of cardiac events. Poor sleep has been associated prospectively with risk of CHD mortality (39), and sleep apnea is positively related to cardiovascular disease (40). A polysomnographic study of acute MI patients revealed that those with sleep apnea were more likely to have had onset during morning hours (41). Disturbed sleep is associated with shifts in the pattern on interleukin 6 (IL-6) production (42), and with compensatory increases in sympathetic tone and cortisol output in the evening (43). No controlled evaluations of the risk of triggering by lack of sleep have been described.
Tobacco Smoking
Smoking increases risk of CHD, and smokers have higher plasma IL-6 and C-reactive protein, more procoagulant hemostatic profiles, and poorer endothelial function than nonsmokers (44). Acutely, tobacco smoking increases catecholamine and cortisol levels, heart rate, and blood pressure (45), impairs endothelial function (46), and increases leukocyte counts (47). Smoking causes symptomless regional myocardial perfusion abnormalities in patients with coronary artery disease (48), and stimulates epicardial coronary artery vasoconstriction despite an increase in myocardial oxygen demand (49).
Despite these adverse effects, there is little evidence to date of direct triggering of ACS by tobacco smoking. This is partly the result of the nature of cigarette smoking; unless symptoms begin during sleep, a heavy smoker is likely to have smoked tobacco within 1 or 2 hours of symptom onset, but this may be no different from periods that do not precede ACS. There have been few reports of triggering by tobacco smoking from most of the large-scale studies reviewed here. In their study of ambulatory ischemia during EKG monitoring, Gabbay et al. (11) reported that the likelihood of an ischemic episode was 27% when patients had been smoking compared with 5% when they had not. Unfortunately, only 6 smokers were included in the study.
Alcohol Consumption
There is a well-established J-shaped relationship between alcohol consumption and total mortality, with the lowest levels in moderate drinkers, and this pattern has been established for CHD as well (50). Moderate consumption is also protective in patients with diagnosed coronary artery disease (51). It is possible that heavy drinking is a trigger for ACS, but definitive proof is lacking. Prospectively, binge drinking of beer increased risk of death in a sample of middle-aged men in Finland, independently of total alcohol consumption (52). Clinical studies have produced variable results, although some clear cases of alcohol triggering have been described (53). McElduff et al. (54) reported a protective effect of drinking up to 8 drinks in men and 4 drinks in women on acute MI or cardiac death over the previous 24 hours, compared with not drinking in a case–control study in Australia. However, the drinkers who abstained from consumption in the previous 24 hours may have been experiencing prodromal symptoms or cardiac problems so may not be an appropriate comparison group. There was some indication in this study of an increased risk in men and women who drank above these levels, but the number of people in these categories was very small. A case–control study of patients less than 65 years old in New Zealand showed that drinking in the 24 hours before onset was less likely in victims of fatal and nonfatal CHD than in community controls (55). However, individuals with premonitory symptoms may reduce intake before ACS, and the rather low consumption levels suggest that reporting biases may have been operating. Also pertinent is a case–control study of patients admitted for acute brain infarction (56). The relative risk for patients who had consumed a large amount of alcohol in the previous 24 hours was 4.19 (CI, 2.24–7.81), whereas light drinking was not associated with risk. Risk was particularly high in patients who were also hypertensive.
The acute biologic response to alcohol differs with the level of intake. A moderate dose has been shown to reduce platelet aggregation, whereas large doses increase platelet activation (57). Animal studies have shown that alcohol increases coronary artery vasoconstriction acutely (58), whereas in patients with coronary artery disease, an increase in ectopic beats has been observed (59). Large doses of alcohol increase factor VII and plasminogen activator inhibitor-1 (PAI-1) and impair fibrinolysis, effects that persist for several hours (60,61).
There can be difficulties in the estimation of the true effect of factors such as tobacco smoking and alcohol, because patients may not accurately report actions that they might consider socially unacceptable. It is also important to note that many behavioral risk factors such as smoking and alcohol consumption can increase in frequency or intensity at times of emotional stress.
High-Fat Meals
It has been hypothesized that high-fat meals could be triggers of ACS, because of their effects on endothelial function, prothrombic responses, and platelet activation (62). However, the current literature is inconclusive. For example, the acute impairments of endothelial function after consumption of high-fat meals reported by some investigators (63) have not been replicated by others either in healthy individuals or patients with coronary artery disease (64,65). The evidence from interview studies that high-fat meals or excessive food intake trigger ACS is modest. Overeating in the period before symptom onset was reported by a small proportion of patients in the MILIS study, but in the absence of control data, it is not possible to determine whether this was an unusual or common event in the patients' lives (15).
Emotional Triggers
A growing body of evidence attests to the influence of emotional and stress-related psychosocial factors in the etiology of CHD. Strong support comes from large-scale prospective epidemiologic studies involving the assessment of psychosocial and standard risk factors in disease-free individuals, and the identification of predictors of fatal and nonfatal cardiac events. Systematic reviews have concluded that the firmest evidence to date is for an influence of low socioeconomic status, work stress, social isolation, and depression on CHD (36,66). Findings related to other forms of chronic stress, anxiety, and anger/hostility have been less consistent. These factors have also been associated with subclinical markers of atherosclerosis as well as manifest disease and may act through accelerating atherogenesis (67,68). It is of great interest therefore to discover whether similar factors also operate acutely as triggers.
Earthquakes
A number of studies have examined the rates of cardiovascular events after the experience of an earthquake. Natural disasters of this type lead to widespread devastation, with long-term as well as acute effects on cardiac health. In studies of acute effects, it is important to distinguish between cardiac events resulting from the severe stress of living through an earthquake from those resulting from sudden physical exertion (eg, running away from buildings) or direct injury and trauma. Several earthquakes have been studied over recent years, but effects on ACS have not been uniform. The most thorough analyses have been those carried out after the Northridge earthquake in the Los Angeles area in January 1994 (69–71). A postal survey of more than 100 hospitals in the area showed that the number of admissions for acute MI increased from 149 in the week before to 201 in the week after the earthquake (70). Examination of the coroners records for Los Angeles County revealed that the number of sudden deaths from cardiac causes increased from an average of 4.6 per day in the week before to 24 on the day of the earthquake (69). Only 3 of these cases were associated with unusual physical exertion, and the vast majority had elevated cardiovascular risk factors or clinical histories. A further analysis examined all deaths in the county and confirmed the increase in deaths from CHD on the day of the earthquake (71). There was no increase in deaths from other cardiovascular diseases such as cardiomyopathy or from noncardiovascular causes, suggesting a specific association with coronary artery disease. Deaths from CHD and SCD were less frequent than average in the first or second week after the earthquake, suggesting that the acutely traumatic event had caused fatal cardiac events in people who were highly vulnerable to ACS (69,71).
Analyses of the Hanshin-Awaji earthquake in 1995 in the Kobe region of Japan are consistent with these findings, with a substantial increase in the number of patients admitted with acute MI on the day of the event (72). Somewhat smaller increases in rates were recorded after earthquakes in Greece and Australia (73,74). One exception to this association between earthquakes and triggering was seen in the period after the 1989 Loma Prieta earthquake in the San Francisco Bay area. A possible explanation may lie in the timing of this event. Brown (75) compared the cardiac consequences of the Loma Prieta and Northridge earthquakes, and noted that the Northridge earthquake struck at 4:31 am, whereas the Loma Prieta earthquake (an event of similar magnitude) struck at 5:04 pm. This suggests that risk is greater during the vulnerable morning period. The Northridge earthquake occurred on a Monday in January, whereas the Loma Prieta earthquake occurred on a Tuesday in October. There is a greater susceptibility to acute MI in winter months, and from other information about timing of infarctions, a Monday morning in the winter is among the most lethal times for an event such as this to take place (3). The Hanshin-Awaji earthquake struck early on a Tuesday morning on a winter's day.
Not surprisingly, there have been few direct studies of acute changes in the pathophysiological mechanisms that might underlie triggering by earthquakes. Fortuitously, however, a research group in Taiwan was carrying out Holter monitoring on 15 patients with suspected coronary artery disease when the island was struck by a major earthquake in 1999 (76). Power spectrum analysis documented a marked increase in low-frequency to high-frequency ratio for approximately 40 minutes after the earthquake, indicative of vagal withdrawal. ST-segment depression occurred in several patients and was correlated with the increase in low-frequency power. Parati et al. (77) have described a single patient who was undergoing ambulatory blood pressure monitoring during a moderate-severity earthquake in central Italy. There was a marked increase in systolic and diastolic pressure and in heart rate, which persisted for an hour, and the next 6 hours were characterized by high blood pressure variability. Less acute effects that have been reported include increased blood viscosity, fibrinogen, and D-dimer levels (78), an increase in deep, negative T waves without Q waves, and abnormal cardiac sympathetic function as revealed with metaiodobenzyl guanidine imaging (79).
Sporting Events
The quarter-final of the 1996 European football (soccer) championships between the French and Dutch teams resulted in a draw at the end of extra time and so went to a penalty shoot out (sudden death) that the French won. The impact of this exciting event was studied in an analysis of mortality in the complete population of Dutch men and women aged 45 years or more. There was a relative risk of death from acute MI or stroke of 1.51 (CI, 1.08–2.09) for men on the day of the match, compared with the 5 days on either side, with no effect on women. No such effect was observed among French men (80). A pattern of increased hospital admissions for acute MI in England on the day of the country's 1998 World Cup match against Argentina has also been described (81). This match again involved a penalty shoot out, and England lost. Another study from The Netherlands failed to observe any increase in ACS on the days of other international matches, possibly because these did not involve penalty shoot outs, which are particularly tense (82). On the other hand, a more recent study of football matches played by teams from the northeast of England over a 5-year period demonstrated a modest increase in deaths attributable to acute MI and stroke in men on the days on which these teams lost (83). The circumstances surrounding these deaths are not known, and it is possible that physical exertion, emotional stress, and alcohol consumption all contributed to triggering.
War
In the initial phases of the Gulf War in 1991, the incidence of acute MI and sudden death increased in an area near Tel Aviv that was close to but not directly affected by missile attacks (84). This finding was subsequently confirmed in a national survey in Israel, in which there was a 58% increase in total mortality on the day of the first missile strikes that was largely attributable to acute MI and SCD (85). No excess mortality over the next 16 days of attacks was observed, suggesting either that susceptible individuals had succumbed or that the population adapted to the stress. Chi et al. (86,87) have examined cardiac death rates and admissions to acute coronary care in New York City in the period surrounding September 11, 2001. No excess rates were observed. Studies in other zones of war or civil unrest have produced mixed results, but the quality of data collected under such difficult circumstances is very variable (88,89).
Emotional Distress and Anger
The TRIMM study was primarily concerned with physical exertion as a trigger, but also assessed emotional factors (27). Emotional upset was reported by 4.4% of patients in the 24 hours before onset, with a relative risk in the case–control analysis of 2.7 (CI, 1.1–6.6). The associations with emotional upset and stress at work in the 4 weeks before acute MI were also significant. In the Onset study, participants completed the anxiety subscale from the State-Trait Personality Inventory for the 2 hours before acute MI and for a pair-matched control period 24 hours earlier (90). The relative risk of having a score above the 75th percentile during the 2-hour hazard period was 1.6 (CI, 1.1–2.2).
A recent analysis of the SHEEP study assessed exposure to work-related stressors as acute triggers (91). Using case–crossover methodology, it was found that certain work stressors in the previous 24 hours such as having high-pressure deadlines were associated with substantial increases in risk (odds ratio, 6.0; CI, 1.8–20.4), in comparison with the immediately preceding 24 hours. This result is interesting in view of the evidence that high job demands, either alone or in conjunction with low control, predict future CHD (36).
The role of anger as a possible trigger of ACS has been assessed in 2 studies. An analysis of the Onset study indicated that 39 of 1623 patients reported being very angry or furious in the 2 hours before acute MI, an incidence of 2.4% (90). In comparison with usual levels of anger, the relative risk of acute MI in the 2 hours after an episode of anger was 2.3 (CI, 1.7–3.2), and in comparison with a pair-matched control period 24 hours earlier, it was 4.0 (CI, 1.9–9.4). This effect was independent of age, sex, cardiovascular risk factors, and the use of beta-blockers, although aspirin had a protective effect. Interestingly, the risk of anger triggering was inversely related to socioeconomic status as indexed by educational attainment; the relative risk varied from 3.3 (CI, 2.0–5.4) in patients with less than high school attainment, to 1.6 (CI, 0.9–2.9) in those with college education (92). The association between anger and symptom onset was confirmed in the SHEEP study, in which the absolute incidence of intense anger in the hour before onset was 1.2% (93). The relative risk (compared with usual levels of anger) was 9.0 (CI, 4.4–18.2), but increased to 15.7 (CI, 7.6–32.4) when analysis was limited to patients who had no premonitory symptoms.
Studies of acute mental states to date have centered on stress and anger. Acute depression has yet to be thoroughly investigated, but might be important in view of the evidence that depression is relevant both to the long-term development of CHD and to prognosis after cardiac events (66).
Emotional Stress and Sudden Cardiac Death
There is extensive evidence that psychosocial factors are associated with the development of SCD over the long-term (94), with particularly striking effects for anxiety (95,96). However, there have been few studies of acute emotional triggers of SCD over recent years, although earlier work suggested a connection (97,98). One of the most thorough studies was carried out with the relatives of 100 men aged 70 years or less who died suddenly and who were interviewed within 10 days about the circumstances surrounding the death. Results were compared with those collected from 100 MI admissions (99). Coronary risk profiles did not differ, but the men who suffered a sudden death were more likely to have consumed alcohol within 3 hours of symptoms. They were also significantly more likely to have experienced moderate or severe stress in the 30 minutes before onset (23% vs. 8%). Although the study involved a selected sample of cases meriting necropsy by the coroner's office, and neither the informants nor investigators were blind to group, the results are consistent with emotional stress being a trigger in many cases of SCD as well as nonfatal ACS. Similarly, acute stress was said by relatives of friends to precede SCD in 19% of patients in a case series from Finland (98).
Methodologic Issues in Studying Triggers
The use of case–control and case–crossover designs has greatly increased the confidence with which conclusions can be drawn about behavioral and emotional triggers, because these methods overcome many of the biases present in studying acute causes. Nevertheless, a cautious attitude to interpretation is necessary because of several factors in the design of these studies.
Retrospective Reporting Bias
Neither the case–crossover nor case–control designs involving hospitalized patients can eliminate biases in reporting. Although the same patients provide information for hazard and control periods in the case–crossover design, individuals with ACS may emphasize emotional or other triggers in the presymptomatic period as they develop their cognitive representation of the cardiac event (22). One method of helping to address this issue would be to seek verification of the patient's behavior or emotional state from bystanders or relatives who witnessed the circumstances. This has not been done in the major studies of triggers of nonfatal events, and data have derived almost completely from patients themselves. Studies of transient silent ischemic events occurring during Holter monitoring do not suffer from this limitation, because patients are not aware of the timing of cardiac events (11,12).
Memory Decay and Salience
In the case–crossover design, the control period is more distant in time than the hazard period (23) and may lack the salience of the hours before symptom onset. Memory for these control periods, even for a pair-matched period 24 hours earlier, may not therefore be as accurate as for the hazard period. This could result in an overestimation of relative risk, particularly with emotionally salient triggers. Additionally, a common control period selected in these designs is the usual level of exposure in the week, 6 months, or year before ACS onset. This requires patients to estimate how often they were usually angry, physically active, and so on, over an extended period of time. Mittleman et al. (100) have modeled various different control periods and have shown that the relative efficiency of the pair-matched control approach is far inferior to using usual frequency as the comparison condition. In an analysis of patients with Ménière's disease involving repeated measures of stress, it was found that memory decay did not influence classification of exposure to triggers (101). However, 1 of the cardinal features of behavioral assessment is that estimates are often less accurate for extended time periods than for specific intervals that can be accessed in memory through prompts or cues. Because susceptibility to triggers may vary with time of day as well, control periods that are matched in time will also improve precision. It may be, therefore, that the relative risks for hazard periods compared with usual levels are less reliable in these analyses than are comparisons with specific defined periods of the same length (such as the same 1 or 2 hours on the preceding day).
Premonitory and Prodromal Symptoms
A particularly important issue is whether the trigger events precede acute symptoms, or whether there were premonitory signs or silent prodromal events. The presence of premonitory symptoms may alter the hypothesized causal chain, and reverse causation may operate. For example, there are large and rapid increases in inflammatory markers such as IL-6 and C-reactive protein at the early stages of ACS (102). Inflammatory stimuli have been shown to induce acute negative emotional states, even when the stimulus is relatively mild (103). Hence, it is conceivable some "triggers" are actually consequences of early manifestations of plaque disruption, or direct responses to earlier pain or fatigue.
Clinical Presentation Bias
A proportion of acute cardiac events do not lead to hospitalization, but either to preadmission death or to myocardial damage that is clinically undetected and only comes to light at a later stage, if at all (104,105). It is well established that psychologic, social, and contextual factors influence the presentation of bodily complaints (106). It is possible, therefore, that triggers do not have a direct effect on the occurrence of ACS, but rather influence symptom reporting and health-seeking behavior. Perhaps individuals with ACS in the absence of triggering phenomena are less likely to attend emergency departments or contact their physicians, leading to a preponderance of trigger cases in interview studies. This factor is difficult to discount in the absence of population-based survey methods, but some control may be provided by covarying for negative affect and health anxiety.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
Received for publication April 8, 2004; revision received October 6, 2004.
DOI:10.1097/01.psy.0000155663.93160.d2
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSCONCLUSIONSNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. W. Davidson, M. Burg, and D. Shimbo Endothelin-1 Release and Stimulation of the Inflammatory Cascade: Is Acute Coronary Syndrome Triggered by Watching Spectator Sports? J. Am. Coll. Cardiol., February 16, 2010; 55(7): 643 - 644. [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gautam, J. Menachem, S. K. Srivastav, P. Delafontaine, and A. Irimpen Effect of Hurricane Katrina on the Incidence of Acute Coronary Syndrome at a Primary Angioplasty Center in New Orleans Disaster Med Public Health Preparedness, October 1, 2009; 3(3): 144 - 150. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Williams and K. W. Davidson Psychological Stress and Cardiovascular Disease With Emphasis on Acute Coronary Syndromes J. Am. Coll. Cardiol., April 14, 2009; 53(15): 1339 - 1339. [Full Text] [PDF]
|
|
|
|
|
|
Y. Chida and A. Steptoe The association of anger and hostility with future coronary heart disease: a meta-analytic review of prospective evidence. J. Am. Coll. Cardiol., March 17, 2009; 53(11): 936 - 946. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Burg, B. Graeber, A. Vashist, D. Collins, C. Earley, J. Liu, R. Lampert, and R. Soufer Noninvasive Detection of Risk for Emotion Provoked Myocardial Ischemia Psychosom Med, January 1, 2009; 71(1): 14 - 20. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Walters, G. Rait, I. Petersen, R. Williams, and I. Nazareth Panic disorder and risk of new onset coronary heart disease, acute myocardial infarction, and cardiac mortality: cohort study using the general practice research database Eur. Heart J., December 2, 2008; 29(24): 2981 - 2988. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B Chaix, M Lindstrom, M Rosvall, and J Merlo Neighbourhood social interactions and risk of acute myocardial infarction J Epidemiol Community Health, January 1, 2008; 62(1): 62 - 68. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kivimaki, J. E. Ferrie, M. Shipley, D. Gimeno, M. Elovainio, R. de Vogli, J. Vahtera, M. G. Marmot, and J. Head Effects on Blood Pressure Do Not Explain the Association Between Organizational Justice and Coronary Heart Disease in the Whitehall II Study Psychosom Med, January 1, 2008; 70(1): 1 - 6. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Smoller, M. H. Pollack, S. Wassertheil-Smoller, R. D. Jackson, A. Oberman, N. D. Wong, and D. Sheps Panic Attacks and Risk of Incident Cardiovascular Events Among Postmenopausal Women in the Women's Health Initiative Observational Study Arch Gen Psychiatry, October 1, 2007; 64(10): 1153 - 1160. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Dickens, L. McGowan, C. Percival, B. Tomenson, L. Cotter, A. Heagerty, and F. Creed Depression Is a Risk Factor for Mortality After Myocardial Infarction: Fact or Artifact? J. Am. Coll. Cardiol., May 8, 2007; 49(18): 1834 - 1840. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. H. Tofler and J. E. Muller Triggering of Acute Cardiovascular Disease and Potential Preventive Strategies Circulation, October 24, 2006; 114(17): 1863 - 1872. [Full Text] [PDF]
|
|
|
|
 |
|
 P C Strike, L Perkins-Porras, D L Whitehead, J McEwan, and A Steptoe Triggering of acute coronary syndromes by physical exertion and anger: clinical and sociodemographic characteristics Heart, August 1, 2006; 92(8): 1035 - 1040. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Strike, K. Magid, D. L. Whitehead, L. Brydon, M. R. Bhattacharyya, and A. Steptoe Pathophysiological processes underlying emotional triggering of acute cardiac events. PNAS, March 14, 2006; 103(11): 4322 - 4327. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
