This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Möller, J. Articles by Ahlbom, A. Search for Related Content PubMed PubMed Citation Articles by Möller, J. Articles by Ahlbom, A. Related Collections Coronary Artery Disease Personality

Do Episodes of Anger Trigger Myocardial Infarction? A Case-Crossover Analysis in the Stockholm Heart Epidemiology Program (SHEEP)

From the Department of Public Health Sciences (J.M., J.H., F.D.), Division of Social Medicine, Karolinska Institutet, Stockholm; Swedish National Institute for Psychosocial Factors and Health (T.T.), Stockholm; Institute of Environmental Medicine (C.R., A.A.), Karolinska Institutet, Stockholm; Department of Occupational Medicine (C.R.), Karolinska Hospital, Stockholm; and National Institute for Working Life (C.R.), Solna, Sweden.

Address reprint requests to: Jette Möller, Karolinska Institutet, Department of Public Health Sciences, Division of Social Medicine, Norrbacka, S-171 76 Stockholm, Sweden. Email: jette.moller{at}phs.ki.se

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
OBJECTIVE: Our objectives were to study anger as a trigger of acute myocardial infarction (MI) and to explore potential effect modification by usual behavioral patterns related to hostility.

METHODS: This study was a case-crossover study within the Stockholm Heart Epidemiology Program. Exposure in the period immediately preceding MI was compared with exposure during a control period for each case. From April 1993 to December 1994, 699 patients admitted to coronary care units in Stockholm County were interviewed.

RESULTS: During a period of 1 hour after an episode of anger, with an intensity of at least "very angry," the relative risk of MI was 9.0 (95% CI, 4.4–18.2). In patients with premonitory symptoms, the time of disease initiation may be misclassified. When restricting the analyses to those without such symptoms, the trigger risk was 15.7 (95% CI, 7.6–32.4). The possibility of examining effect modification was limited by a lack of statistical power (eight exposed cases). Results of the analyses suggested, however, an increased trigger effect among subjects reporting nonhostile usual behavior patterns, nonovert strategies of coping with aggressive situations (not protesting when being treated unfairly), and nonuse of ß-blockers.

CONCLUSIONS: The hypothesis that anger may trigger MI is further supported, with an increased risk lasting for approximately 1 hour after an outburst of anger. It is suggested that the trigger risk may be modified by personal behavior patterns.

Key Words: myocardial infarction • trigger • anger • case-crossover study

Abbreviations: MI = myocardial infarction; Onset SHEEP =Stockholm Heart Epidemiology Program.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
Results of several studies show that hostility is a risk factor for coronary heart disease, although there are some contradictory results (1–11). Different models have been suggested to explain this relationship (1, 12). The triggering of acute MI by episodes of anger is one potential mechanism. Patients with MI often believe that emotions, such as anger, are involved in the onset of their disease (13–17). A new epidemiological tool, the case-crossover design, has been developed to facilitate the study of triggers of disease onset (18). In the first case-crossover study of anger as a trigger of MI, Mittleman et al. (19, 20) reported a relative risk of 2.3 (95% CI, 1.7–3.2) during a period of 2 hours after an episode of anger. The trigger effect was modified by educational status, use of ß-blockers or aspirin, and history of a previous MI. A case-control study (21) also showed an increased risk of MI during the first 24 hours after an emotionally upsetting event, with an odds ratio of 2.7 (95% CI, 1.1–6.6). Results of other studies also support the notion that emotions may cause cardiovascular ischemia with a short induction time. Gabbay et al. (22) found that in patients with coronary artery disease undergoing ambulatory electrocardiographic monitoring, the likelihood of myocardial ischemia was greatest during intense physical and stressful mental activity. In a case-crossover study, Gullette et al. (23) found an increased risk of silent myocardial ischemia (measured by electrocardiographic changes) after episodes of mental stress, such as sadness, frustration, and tension.

The first objective of this case-crossover study was to examine whether an outburst of anger can trigger the onset of an acute nonfatal MI. The straightforward interpretation of such a finding would be that one should refrain from getting angry to avoid an acute MI. This message may be too simplistic, however, for several reasons. Anger is sometimes an adequate response to external sources of irritation. Submissive people who usually do not express their aggressive feelings openly might become more tense when feelings of anger develop, implying an increased trigger effect of the anger. Furthermore, the frequency of instant aggressive reactions partly defines more stable behavioral or personality traits of the individual, which in turn might modify the trigger effect. People with a hostile behavioral pattern, who face an increased risk of MI over the long term, might, on the other hand, have a lower risk with each episode of anger because they are more accustomed to such feelings or because their emotional fluctuation is not as drastic. The second objective was therefore to explore whether hostile behavioral patterns and strategies of coping with aggressive situations modify the triggering effect of episodes of anger.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
The Swedish Onset study is a case-crossover analysis nested in a population-based case-referent study of causes of MI, entitled the Stockholm Heart Epidemiology Program (SHEEP). In SHEEP, all cases of first events of MI in Stockholm County from 1992 to 1994 were identified and registered at the 10 emergency hospitals through a special organization. Missed cases were later identified by reviewing hospital discharge registers and by screening death certificates. MI was diagnosed according to standardized criteria developed by the Swedish Association of Cardiologists (see Appendix 1). Cases were included at the time of disease detection, and a referent was randomly selected from the study base after stratification by age, sex, and hospital catchment area. The study base included all Swedish citizens aged 45 to 70 years living in Stockholm County who had no history of MI. Exposure information was obtained by postal questionnaires, telephone interviews, and health examinations.

The Onset study population consists of all patients in the SHEEP study base from April 1993 to December 1994 with a first event of nonfatal acute MI. The total number of identified cases in SHEEP during this study period was 1266, of which 699 (70% of all nonfatal cases) were interviewed for the Onset study. After exclusion of cases with unreliable information or with a high percentage of absent or obviously inaccurate answers about the time of disease onset and exposure information, 660 cases remained. Characteristics of the study population are shown in Table 1.

Cases in the Onset study were interviewed during the period of hospitalization or soon after discharge. The interviews, which were conducted by nurses trained in the interview technique through group sessions, were followed by pilot interviews, in which a training manual and individual instructions were used. Several meetings were held with all interviewers during the study period to reinforce consistent interview and coding practices. To diminish recall bias, the interviewers were not informed about the assumptions regarding the length of induction periods (the time from exposure to disease initiation). Instead, they were asked to pay equal attention to all 1-hour periods during the 26 hours before MI. Throughout the interview, detailed information was also obtained on all episodes of pain (time, type, duration, etc.) and other symptoms and circumstances during the 4 days before MI to determine the precise time of disease onset.

Premonitory symptoms were defined as any symptom fulfilling the criteria of a diagnostic symptom (see Appendix 1) occurring in the 4-day period before the onset of disease. In cases of angina pectoris, the patient made a subjective evaluation of whether a "premonitory" episode of chest pain should be regarded as ordinary angina. Once the group of patients without premonitory symptoms was subsequently determined, coding decisions aiming at high specificity were made to define a group in which misclassification concerning the time of disease onset was minimal.

In the Onset study, episodes of anger were identified during the interview through use of the anger scale, which was developed by the Boston team (19). The scale measures anger at seven different levels (see Table 2 for a description of the four highest levels). The interviewers were instructed to stress the intensity concepts of the scale (moderately angry, very angry, furious, and enraged), and the physical expressions of anger were introduced only as examples of what the patient could have felt like doing when feeling anger at a particular level. Subjects were asked to give detailed information about each of the four highest levels.

To check the validity of the case-crossover estimate of the trigger effect, a matched case-referent analysis was also conducted. The cases in the Onset study had referents in the SHEEP study, and 96% of these participated in a telephone interview concerning trigger exposures before a specific point in time. The control information referred to the week before the referent interview and was matched on hour of day and day of week of the corresponding cases’ MI onset.

Statistical Analysis
Episodes of anger were assumed to be point exposures and infrequent enough for the hazard periods not to overlap. The number of exposed hours per year was calculated as the length of the hazard period multiplied by the usual annual frequency. The number of unexposed hours was calculated by subtracting the exposed hours per year from the total number of hours per year (8760 hours). The case-crossover study can be viewed as a pooled analysis of 660 retrospective, self-matched follow-up studies, each with a sample size of one. Methods for highly stratified data are applicable (24, 30, 31). Standard Mantel-Haenszel methods for follow-up studies with sparse data in each stratum are used for the statistical analysis (24, 32). The effect is measured as a relative risk, estimated by the average incidence rate ratio (the ratio between the observed and expected odds). Through stratification by potential effect modifiers, group-specific relative risks were assessed, and differences between the groups were tested by the ² test (30). Conditional logistic regression was used in the matched case-referent analyses. Calculations were made using SAS software, version 6.12, and Excel 7.0.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
Table 3 shows that during and up to 60 minutes after an episode of anger, the relative risk of MI was 9.0 (95% CI, 4.4–18.2). During the period 61 to 120 minutes, we found a nonsignificantly increased relative risk of 2.3 (95% CI, 0.5–9.1). The risk in the next period, 121 to 360, minutes was not increased (1.2; 95% CI, 0.4–3.1). This implies that the induction time is probably 1 hour or less and definitely less than 2 hours.

The Cronbach coefficients were 0.8 for the scale of overt coping and 0.7 for the five items measuring a hostile behavioral pattern. The crude relative risk of nonfatal MI in SHEEP due to increased hostility as a risk factor with a long induction time was 1.4 (95% CI, 1.2–1.6). Among those normally using an overt coping strategy, the crude long-term risk was 1.2 (95% CI, 1.0–1.5). The analyses of effect modification were restricted to those without premonitory symptoms, and anger at level 5 or higher during the 0- to 60-minute period before MI was defined as exposure. Table 4 shows that those with infrequent outbursts of anger (once every second week or less) had a significantly higher trigger risk than those with more frequent outbursts. Regardless of usual hostile behavioral patterns according to the five-item scale, there was an increased risk of triggering by episodes of anger. However, the risk seemed to be higher among the usually nonhostile cases. Subjects with an overt coping pattern seemed to have a lower trigger risk than those who usually do not express their anger openly. Use of ß-blockers seemed also to alter the risk substantially. There was a reversed association between ß-blocker medication and the risk of being triggered due to anger. The results support our hypotheses, but, even though some of the differences in relative risk estimates were considerable, the power in the study was not sufficient to always prove significance. The percentage of acetylsalicylic acid users was as low as 7.7%, which is explained by the fact that we exclusively studied first MI events.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
In the first case-crossover study of anger as a trigger of MI, Mittleman et al. (19) reported a relative risk of 2.7 within 2 hours after an episode of anger among those without previous MI. The hypothesis that anger may trigger MI is further supported by our study. The relative risk of 5.7 (95% CI, 3.0–10.6) (Table 3) is directly comparable with the relative risk reported by Mittleman et al. In the Boston study, 2.4% of the patients had been exposed to anger during the 2-hour period before MI, compared with 2.3% of the patients without premonitory symptoms in this study. The Stockholm study was population-based, with an age range of 45 to 70 years, which was a narrower interval than the 20- to 92-year range in the Boston study.

To reduce bias, we restricted further analyses to those without premonitory symptoms. If such symptoms indicate the start of the biological process in disease initiation (eg, plaque rupture or beginning of thrombosis), then the period before the premonitory symptom should be of interest. Because exposure information was available only for hazard periods before the diagnostic pain, case exposure may be misclassified in patients with premonitory symptoms, leading to dilution of the effect. Premonitory symptoms also imply the possibility of reversed causation. Patients with premonitory symptoms may feel weak and try to avoid situations that may lead to feelings of anger. The result would be an underestimation of the risk. Conversely, if the premonitory symptoms are irritating, the patients may be more easily teased and therefore more likely to become angry than usual, resulting in an overestimation of the trigger effect. The consequences of our restriction to those without premonitory symptoms are diminished power and increased effect estimates. However, the conclusions drawn from the study would not have been different had the material been analyzed unstratified. We also believe that the results from the group without premonitory symptoms can be generalized, because there are no biological reasons for assuming different mechanisms between the two groups. In future studies, it would be interesting to also collect exposure information related to the hazard periods before premonitory symptoms.

In the analysis of hazard period length among those without premonitory symptoms, the relative risk was 15.7 during the first 1-hour period before MI. The relative risk was 2.0 in the period 61 to 120 minutes before MI, but the confidence interval was wide. We conclude that the maximum induction period is likely to be 1 hour or less. The excess risk may, however, remain for another hour, as reported by Mittleman et al. (19), but perhaps at a lower level, which is in agreement with the stepwise risk function proposed by Maclure (18).

The study of effect modification in subgroups of patients suffered from a lack of statistical power because there were only eight exposed cases. The tentative results have, however, interesting interpretations. Patients using ß-blockers seemed to be at lower risk than those who did not. Mittleman et al. (19, 20) made a similar observation in their studies. The explanation may be that the ß-blocker cuts off the peak of a sudden rise in blood pressure, thereby reducing the risk of plaque rupture.

The triggering effect of anger also seemed to be modified by the individual’s usual pattern of aggressive behavior (Table 4). Anger was a trigger in all groups, yet the effect was less pronounced in patients who were more accustomed to outbursts of anger. The long-term effects of a hostile behavioral pattern may, on the other hand, also increase the risk of MI (1–11). The result suggests that it is inappropriate to simply conclude that all anger should be avoided to reduce the risk of MI. There may be a balance between short- and long-term effects of anger and perhaps also an important distinction between the accumulation of hostile tension and appropriate angry reactions.

Hostility was measured by five items selected from the immersion scale of Siegrist (29), and they were chosen because they included statements on aggressive behaviors or feelings (Appendix 2). The other measure of hostility, which is based on the usual weekly number of anger outbursts, is crude but in accordance with the use of annual frequency of physical exertion as a proxy measure of physical fitness in earlier trigger studies (25). No standard instrument for measuring the hostility trait was available. The consequence of using an ad hoc instrument based on face validity should be an underestimation of the exposure contrast regarding hostility.

The self-matching feature of the case-crossover design eliminates the problems of control selection and confounding from risk factors with long induction periods. There was no evidence of any circadian patterns in anger occurrence; therefore, confounding by time of day should not pose any problem. Confounding from other triggers might introduce a bias. If a subject often gets angry while engaging in physical exertion (or vice versa), the MI could be triggered by either the outburst of anger or the physical exertion. Two patients were exposed to heavy physical exertion and anger during the same hazard period. No information about the usual frequency of this cooccurrence was available; as a result, accurate expected exposure odds could not be calculated. A less powerful test was conducted in which all cases were excluded from the analysis of anger if the patient had been exposed to physical exertion during the same hazard period before MI. The relative risk then decreased from 15.7 (Table 3) to 12.2 (95% CI, 5.3–28.2), indicating that some of the excess risk we found could be due to coexposure to anger and physical exertion. If we had the correct information on usual frequency of the coexposure, then the denominator of the relative risk would be expected to be lower, under the assumption that every outburst of anger is not associated with physical exertion, and, as a result, the relative risk in the test should be somewhat higher.

Misclassification of the usual frequency information constitutes an important validity problem (31). The average frequency of anger at level 5 or higher was about once a month, and this usual frequency can be difficult to estimate because anger is an exposure that seldom has a repeatable pattern. It is perhaps more likely that episodes of anger are forgotten rather than overreported, because unpleasant experiences are not kept in memory. This would lead to an overestimation of the relative risk. Mittleman et al. (19) found consistent results with the usual frequency approach and the pair-matched approach (using control information from the corresponding 2-hour period 24 hours before the MI), which is an argument against the significance of this bias. We could not perform the pair-matched analysis because we had no exposed cases during the control period (ie, the 25th hour before MI), which is a limitation in this study. Instead, we conducted a matched case-referent analysis using information from the referents of the SHEEP study, in which we found a relative risk of 3.0 (95% CI, 0.3–28.8) for an episode of anger during the 0 to 60 minutes preceding the MI.

Another misclassification problem in the control information was due to occasional difficulties in distinguishing between a very low annual frequency of anger and missing answers. The reason for this was incomplete coding by the interviewers, and it could imply some degree of underreporting that would lead to overestimation of the relative risk. A sensitivity analysis showed that the usual frequency must have been underestimated by approximately 160 times per year and person to eliminate the observed relative risk, which seems unlikely. Furthermore, if we confine the analysis to those who clearly stated a usual frequency, keeping the low-frequency/high-trigger risk group out of the analysis, the relative risk becomes 4.5 (95% CI, 1.6–12.1) for all patients and 7.8 (95% CI, 2.8–21.8) for those without premonitory symptoms.

Survival bias implies that if cases being exposed to anger have a better prognosis for surviving MI than those not exposed to anger, a study of only nonfatal cases would overestimate the relative risk of MI. Likewise, if cases exposed to anger right before their MI are less inclined to participate, this would result in an underestimation. Case selection due to survival effects or differences in participation rate is, however, unlikely.

Finally, it should be noted that although the relative risks are substantial, the absolute risk per hour, and likewise the risk difference, is low. In Stockholm, the absolute risk of MI was 0.3 per million person-hours in the age group 45 to 70. This means that in a group of 4 million people monitored for 1 hour, there will be one case of MI. If they were all very angry during this hour, the number of MI cases would increase to 16. However, the finding in controlled studies showing that anger may trigger MI is important because it opens up a number of interesting research questions related to the interaction between the long- and short-term cardiovascular effects of aggressive behavior.

	APPENDIX 1

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
Diagnostic Criteria for Acute MI
This diagnostic criteria was developed in 1991 by the Swedish Association of Cardiologists. A diagnosis of acute myocardial infarction requires that either two of criteria 1 to 3 or criterion 4 be fulfilled.

1. Any of the following patient histories:
   • pain/pressure and/or a burning sensation located centrally in the chest, starting less than 24 hours before diagnosis, with a duration of more than 15 minutes and without association with breathing, body position, or food intake;
   • recurrent attacks at rest of the same symptoms as above but with a duration of less than 15 minutes during the last 24 hours;
   •pulmonary edema; shock without suspicion of hemorrhage, hypovolemia, sepsis, or intoxication; or
   • fainting during the last 24 hours.

2. Increase in enzyme levels with time according to specific criteria.

3. Emergence of a pathologic Q wave in at least two electrocardiographic channels.

4. Necrosis of the myocardium at autopsy with an age corresponding to the onset of symptoms.

	APPENDIX 2

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 
Questions From Which the Index Variables Were Constructed
Responses to the following statements were used in the assessment of hostile behavior:

1. Even the slightest disturbance bothers me;

2. I enjoy proving others wrong;

3. I get upset at others more often than seems appropriate;

4. I can get furious if others do not understand instantly;

5. I get furious if, when cooperating, a colleague doubts my competence.

The answers were placed on a scale with the following response categories:

correct

almost correct

hardly correct

not correct at all

The alternatives were given scores of 1 to 4. The index then was constructed as the sum of these scores. A high sum indicates hostile behavior. [fr]Responses to the following statements were used in the assessment of overt coping pattern: How do you usually react when you are treated in an unfair way or get into a conflict with a) a superior or b) a work mate?

1. You protest directly.

2. You talk to the person right away.

3. You yell at the person right away.

4. You speak to the person later when things have calmed down.

The answers were placed on a scale for each situation (a and b) with the following categories:

yes, mostly

yes, sometimes

no, seldom

no, never

The alternatives were given scores of 1 to 4. The index was then constructed as the sum of these scores. A high sum indicates overt behavior.

Received for publication March 12, 1998.

Revision received June 9, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

 

1. Smith TW. Hostility and health: current status of a psychosomatic hypothesis. Health Psychol 1992; 11: 139–50.[Medline]
2. Shekelle RB, Gale M, Ostfeld AM, Oglesby P. Hostility, risk of coronary heart disease, and mortality. Psychosom Med 1983; 45: 109–14.[Abstract/Free Full Text]
3. Dembroski TM, MacDougall JM, Williams RB, Haney TL, Blumenthal JA. Components of type A, hostility, and anger-in: relationship to angiographic findings. Psychosom Med 1985; 47: 219–33.[Abstract/Free Full Text]
4. Markovitz JH, Matthews KA, Kiss J, Smitherman TC. Effects of hostility on platelet reactivity to psychological stress in coronary heart disease patients and in healthy controls. Psychosom Med 1996; 58: 143–9.[Abstract/Free Full Text]
5. Meesters CMG, Smulders J. Hostility and myocardial infarctions in men. J Psychosom Res 1994; 38: 727–34.[Medline]
6. Haynes SG, Feinleib M, Kannel WB. The relationship of psychosocial factors to coronary heart disease in the Framingham study. III. Eight-year incidence of coronary heart disease. Am J Epidemiol 1980; 111: 37–58.[Abstract/Free Full Text]
7. Barefoot JC, Dahlstom GW, Williams RB. Hostility, CHD incidence, and total mortality: a 25-year follow-up study of 255 physicians. Psychosom Med 1983; 45: 59–63.[Abstract/Free Full Text]
8. McCraine EW, Watkins LO, Brandsma JM, Sisson BD. Hostility, coronary heart disease (CHD) incidence, and total mortality: lack of association in a 25-year follow-up study of 478 physicians. J Behav Med 1986; 9: 119–25.[Medline]
9. Hearn MD, Murray DM, Luepker RV. Hostility, coronary heart disease, and total mortality: a 33-year follow-up study of university students. J Behav Med 1989; 12: 105–21.[Medline]
10. Leon GR, Finn SE, Murray D, Bailey JM. Inability to predict cardiovascular disease from hostility scores or MMPI items related to type A behavior. J Consult Clin Psychol 1988; 56: 597–600.[Medline]
11. Whiteman MC, Fowkes FG, Deary IJ. Hostility and the heart [editorial]. BMJ 1997; 315: 379–80.[Free Full Text]
12. Verrier RL, Mittleman MA. Life-threatening cardiovascular consequences of anger in patients with coronary heart disease. Cardiol Clin 1996; 14: 289–307.[Medline]
13. Tofler GH, Stone PH, Maclure M, Edelman E, Davis VG, Robertson T, Antman EM, Muller JE. Analysis of possible triggers of acute myocardial infarction (the MILIS study). Am J Cardiol 1990; 66: 22–7.[Medline]
14. Willich SN, Löwel H, Lewis M, Arntz R, Baur R, Winther K, Keil U, Schröder R, for the TRIMM Study Group. Association of wake time and the onset of myocardial infarction: triggers and mechanisms of myocardial infarction (TRIMM) pilot study. Circulation 1991; 84 (suppl 6): VI-62–7.
15. Behar S, Halabi M, Reicher-Reiss H, Zion M, Kaplinsky E, Mandelzweig L, Goldbourt U, for the SPRINT Study Group. Circadian variation and possible external triggers of onset of myocardial infarction. Am J Med 1993; 94: 395–400.[Medline]
16. Gelernt MD, Hochman JS. Acute myocardial infarction triggered by emotional stress. Am J Cardiol 1992; 69: 1512–3.[Medline]
17. Sumiyoshi T, Haze K, Saito M, Fukami K, Goto Y, Hiramori K. Evaluation of clinical factors involved in onset of myocardial infarction. Jpn Circ J 1986; 50: 164–73.[Medline]
18. Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol 1991; 133: 144–53.[Abstract/Free Full Text]
19. Mittleman MA, Maclure M, Sherwood JB, Mulry RP, Tofler GH, Jacobs SC, Friedman R, Benson H, Muller JE. Triggering of acute myocardial infarction onset by episodes of anger. Circulation 1995; 92: 1720–5.[Abstract/Free Full Text]
20. Mittleman MA, Maclure M, Nachnami M, Sherwood JB, Muller JE. Educational attainment, anger, and the risk of triggering myocardial infarction onset. Arch Intern Med 1997; 157: 769–75.[Abstract]
21. Willich SN, Lewis M, Löwel H, Arntz H-R, Schubert F, Schröder R. Physical exertion as a trigger of acute myocardial infarction. N Engl J Med 1993; 329: 1684–90.[Abstract/Free Full Text]
22. Gabbay FH, Krantz DS, Kop WJ, Hedged SM, Klein J, Gottdiener JS, Rozansji A. Triggers of myocardial ischemia during daily life in patients with coronary artery disease: physical and mental activities, anger and smoking. J Am Coll Cardiol 1996; 27: 585–92.[Abstract]
23. Gullette ECD, Blumenthal JA, Babyak M, Jiang W, Waugh RA, Frid DJ, O’Connor CM, Morris JJ, Krantz DS. Effects of mental stress on myocardial ischemia during daily life. JAMA 1997; 277: 1521–6.[Abstract]
24. Mittleman MA, Maclure M, Robins JM. Control sampling strategies for case-crossover studies: an assessment of relative efficiency. Am J Epidemiol 1995; 142: 91–8.[Abstract/Free Full Text]
25. Mittleman MA, Maclure M, Tofler GH, Sherwood JB, Goldberg RJ, Muller JE. Triggering of acute myocardial infarction by heavy physical exertion: protection against triggering by regular exertion. N Engl J Med 1993; 329: 1677–83.[Abstract/Free Full Text]
26. Harburg E, Erfurt JC, Hauenstein LS, Chape C, Schuli WJ, Schork MA. Socio-ecological stress, suppressed hostility, skin color, and black-white male blood pressure: Detroit. Psychosom Med 1973; 35: 276–96.[Abstract/Free Full Text]
27. Knox SS, Theorell T, Svensson JCH, Waller D. The relation of social support and working environment to medical variables associated with elevated blood pressure in young men: a structural model. Soc Sci Med 1985; 21: 525–31.
28. Theorell T, Alfredsson L, Westerholm P, Falck B. Coping with unfair treatment at work: how does the coping pattern relate to risk of developing hypertension in middle-aged men and women? Psychother Psychosom. In press.
29. Siegrist J. Adverse health effects of high-effort/low-reward conditions. J Occup Health Psychol 1996; 1: 27–41.[Medline]
30. Rothman KJ. Modern epidemiology. Boston: Little, Brown, and Company; 1986.
31. Marshall RJ, Jackson RT. Analysis of case-crossover designs. Stat Med 1993; 12: 2333–41.[Medline]
32. Greenland S, Robins JM. Estimation of a common effect parameter from sparse follow-up data. Biometrics 1985; 41: 55–68.[Medline]

This article has been cited by other articles:

				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. Nordin, A. Knutsson, and E. Sundbom Is Disturbed Sleep a Mediator in the Association between Social Support and Myocardial Infarction? J Health Psychol, January 1, 2008; 13(1): 55 - 64. [Abstract] [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]

				L. Warner, M. Macaluso, H. D. Austin, D. K. Kleinbaum, L. Artz, M. E. Fleenor, I. Brill, D. R. Newman, and E. W. Hook III Application of the Case-Crossover Design to Reduce Unmeasured Confounding in Studies of Condom Effectiveness Am. J. Epidemiol., April 15, 2005; 161(8): 765 - 773. [Abstract] [Full Text] [PDF]

				Y. Gidron, H. Gilutz, R. Berger, and M. Huleihel Molecular and cellular interface between behavior and acute coronary syndromes Cardiovasc Res, October 1, 2002; 56(1): 15 - 21. [Abstract] [Full Text] [PDF]

				R. Lampert, T. Joska, M. M. Burg, W. P. Batsford, C. A. McPherson, and D. Jain Emotional and Physical Precipitants of Ventricular Arrhythmia Circulation, October 1, 2002; 106(14): 1800 - 1805. [Abstract] [Full Text] [PDF]

				J Moller, A Ahlbom, J Hulting, F Diderichsen, U de Faire, C Reuterwall, and J Hallqvist Sexual activity as a trigger of myocardial infarction. A case-crossover analysis in the Stockholm Heart Epidemiology Programme (SHEEP) Heart, October 1, 2001; 86(4): 387 - 390. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Möller, J. Articles by Ahlbom, A. Search for Related Content PubMed PubMed Citation Articles by Möller, J. Articles by Ahlbom, A. Related Collections Coronary Artery Disease Personality