A Prospective Study of Anger and Coronary Heart Disease
The Normative Aging Study
Background Recent laboratory and epidemiological studies have suggested that high levels of anger may increase the risk of coronary heart disease (CHD).
Methods and Results We examined prospectively the relationship of anger to CHD incidence in the Veterans Administration Normative Aging Study, an ongoing cohort of older (mean age, 61 years) community-dwelling men. A total of 1305 men who were free of diagnosed CHD completed the revised Minnesota Multiphasic Personality Inventory (MMPI-2) in 1986. Subjects were categorized according to their responses to the MMPI-2 Anger Content Scale, which measures the degree to which individuals have problems controlling their anger. During an average of 7 years of follow-up, 110 cases of incident CHD occurred, including 30 cases of nonfatal myocardial infarction (MI), 20 cases of fatal CHD, and 60 cases of angina pectoris. Compared with men reporting the lowest levels of anger, the multivariate-adjusted relative risks among men reporting the highest levels of anger were 3.15 (95% confidence interval [CI]: 0.94 to 10.5) for total CHD (nonfatal MI plus fatal CHD) and 2.66 (95% CI: 1.26 to 5.61) for combined incident coronary events including angina pectoris. A dose-response relation was found between level of anger and overall CHD risk (P for trend, .008).
Conclusions These data suggest that high levels of expressed anger may be a risk factor for CHD among older men.
A small but growing number of epidemiological studies have suggested that anger and hostility are related to incidence of CHD.1 2 3 4 5 In a 3-year follow-up study of 3750 Finnish men aged 40 to 59, high ratings of hostility (assessed from self-ratings of irritability and ease of anger arousal) were associated with increased CHD mortality, although the excess risk was confined to men with preexisting heart disease.2 In the Multiple Risk Factor Intervention Trial,3 men with high “potential for hostility” (a behavioral rating correlated with high verbal and physical expression of anger4 ) had an adjusted relative risk of incident CHD of 1.5 (P=.03). In the Determinants of Myocardial Infarction Onset Study,5 a case-crossover study involving 1623 patients, episodes of anger were found to be potent triggers of acute MI. In a case-crossover design, each individual's exposure to a transient risk factor (such as an episode of anger) is contrasted with exposure during a “control” period (eg, at the same time the previous day).5 Using this design, the relative risk of MI in the 2 hours after an episode of anger was found to be 2.3 (95% CI: 1.7 to 3.2).
In addition to these epidemiological studies, laboratory investigations of patients with established CHD have found that recall of anger produced significant decreases in left ventricular ejection fraction as measured by radionuclide ventriculography,6 as well as coronary artery vasoconstriction in already narrowed segments.7
In this report, we examined prospectively the relationship between anger and CHD risk in the Normative Aging Study, an ongoing cohort of older community-dwelling men.
The Normative Aging Study is a longitudinal study of aging established by the Veterans Administration in 1961.8 The study cohort consists of 2280 community-dwelling men from the Greater Boston area who were aged 21 to 80 years at the time of entry. Volunteers were screened at entry according to health criteria8 and were free of known chronic medical conditions (including diabetes mellitus) at the start of follow-up.
Assessment of Anger
In 1986, the MMPI-29 was administered by mail to all active cohort members (n=1881). One thousand five hundred fifty men responded (82.4% response rate), in whom complete and valid questionnaire data were available in 94% (n=1459). One hundred fifty-four men with preexisting CHD (angina pectoris or history of MI) were excluded, resulting in a study population of 1305 men. The mean age of the study population was 61.8 years (SD, 8.3 years; range, 40 to 90 years).
The Anger Content scale of MMPI-2 is made up of 16 items that require true or false responses to questions about expression or control of anger (see “Appendix”). High scores on the anger scale suggest anger-control problems. These individuals report being irritable, hotheaded, and sometimes feeling like swearing or smashing things. Individual items on the MMPI-2 anger scale are quite similar to those of other scales, such as the Spielberger Anger Expression Inventory scale,10 which measures aspects of anger such as frequency of the expression of anger directed to other people or objects (Anger-Out subscale) and the extent to which a person works to control the experience and expression of anger (Anger-Control subscale).
Responses to the anger subscale were categorized a priori into three levels on the basis of distribution of scores in this population: 0 or 1, 2 to 4, and 5 to 14. The anger scale was also analyzed as a continuous variable.
Given the potential similarity between anger and hostility, an attempt was made to reproduce the Cook-Medley Hostility Scale, which has been shown previously11 12 to predict risk of CHD. However, this goal was not possible because of the substantial differences between the original MMPI (from which the Cook-Medley Scale was derived) and the MMPI-2 (which was the version used in the present study). Instead, we chose to examine the MMPI-2 Cynicism scale, whose items most closely resemble the original Cook-Medley Hostility Scale. The cynicism scale is comprised of 23 items that inquire about misanthropic beliefs. Individuals who score high on this scale expect hidden, negative motives behind the acts of others, for example, believing that most people are honest simply because they fear being caught. These individuals are likely to hold negative attitudes toward those close to them, including fellow workers, family, and friends.13
Measurement of Other Cardiovascular Risk Factors
Every 3 to 5 years, participants in the Normative Aging Study are assessed by physical examination, updating of medical history, and measurement of a variety of biochemical values including serum cholesterol. Cigarette-smoking status (current, former, never) is ascertained by a trained interviewer. Current smokers are defined as men who smoke ≥1 cigarette per day. Weight and height are measured with the participants wearing only socks and underpants. Body mass index (weight/height2) is then calculated. Blood pressure is measured by an examining physician with a standard mercury sphygmomanometer with a 14-cm cuff. With the subject seated, systolic blood pressures and fifth-phase diastolic blood pressures are measured in each arm to the nearest 2 mm Hg. The average systolic and diastolic blood pressures in both arms were used in analyses. Only eight individuals in the study population were receiving oral hypoglycemic agents or insulin.
Assessment of Morbidity and Mortality
The average length of follow-up in the present study was 7.0 years (SD, 2.3 years). The present study includes all confirmed CHD end points (angina pectoris, MI, and fatal CHD) that occurred during the average 7 years of follow-up. Individuals were censored either at the time of developing a coronary end point (or death) or from the time of their most recent follow-up visit.
A medical history was obtained from each participant at his regular follow-up visit every 3 to 5 years. The hospital records were obtained for every report of a possible CHD event and reviewed by a board-certified cardiologist (P.V.). The criteria for MI and angina pectoris were those used in the Framingham Heart Study.14 MI was diagnosed only when documented by unequivocal ECG changes (ie, pathological Q waves), by a diagnostic elevation of serum enzymes (serum glutamic-oxaloacetic transaminase and lactate dehydrogenase) together with chest discomfort consistent with MI, or by autopsy. Angina pectoris was diagnosed when the subject reported recurrent chest discomfort lasting up to 15 minutes, which was distinctly related to exertion and relieved by rest or nitroglycerin. If any individual developed more than one event (eg, angina, then later, nonfatal MI), he was censored at the time of the earlier event. This was done to minimize the misclassification bias introduced when individuals change their behavior after the first event (eg, seek care for symptoms of anger).
Death from CHD was designated when a death certificate (coded according to the Eighth Revision of the International Classification of Diseases15 ) indicated an underlying cause of death coded to rubric 410-414. The medical records in each instance of CHD death were reviewed by a board-certified cardiologist (P.V.) to ensure accurate coding. Most deaths occurring in this cohort are reported by next of kin or postal authorities. Every year, birthday cards are mailed to participants in the cohort, at which point news of a participant's death is likely to be reported back to the investigators by the next of kin. Additional opportunities to ascertain the vital status of participants occur when supplemental questionnaires are mailed to participants on an approximately annual basis. Finally, we routinely search the state vital records as well as the records of the Department of Veterans Affairs to pick up deaths that may have gone unreported. Thus, our ascertainment of fatal events is both systematic and comprehensive.
We ran proportional hazards models using the SAS16 to estimate the relative risks of CHD according to different levels of anger, controlling for a range of potential confounding variables ascertained in 1986, including age (years); body mass index (kg/m2); smoking status (never, former, current); systolic and diastolic blood pressures (mm Hg); serum cholesterol level (mg/dL); family history of heart disease (yes/no); and whether the participant drank two or more drinks of alcohol per day (yes/no). The multivariate probability value for linear trend in the relative risk was estimated by entering the anger score as a continuous variable in the regression models.
The mean anger score among 1305 subjects was 4.59 (SD, 3.11; range, 0 to 14). The distribution of responses to the MMPI-2 Anger Content scale in the entire cohort is shown in Fig 1⇓: 15.2% of the cohort scored 0 or 1 on the MMPI-2 Anger scale, 41.9% scored 2 to 4, and 42.8% scored 5 to 14. Differences were noted in the distribution of coronary risk factors by level of anger (Table 1⇓). Individuals who scored high on anger were younger, heavier, more likely to be current smokers, and somewhat more likely to drink at least two drinks of alcohol per day. There were no statistically significant differences in the distribution of other cardiovascular disease risk factors, including mean systolic or diastolic blood pressure and serum cholesterol level. All potential confounding variables were adjusted for in proportional hazards models.
During the follow-up period, 110 new coronary events occurred in the 1305 men: 30 incident cases of nonfatal MI, 20 cases of fatal CHD, and 60 cases of angina pectoris. Cases of nonfatal MI and fatal CHD were combined to form the category of total CHD (n=50).
Men who reported higher levels of anger had increased risks of total CHD (Table 2⇓). Compared with men scoring 0 or 1, the age-adjusted relative risks of total CHD were 2.76 (95% CI: 0.82 to 9.26) in men scoring 2 to 4 and 3.58 (95% CI: 1.08 to 11.9) in men scoring 5 to 14. Adjusting for multiple risk factors did not appreciably alter these point estimates of risk, although the lower 95% CI now overlapped 1.0 in the highest category of anger (Table 2⇓). Being in the highest category of anger was associated with ≈60% excess risk of nonfatal MI. No cases of fatal CHD occurred among men scoring 0 or 1, so that we could not separately estimate the relative risks of fatal CHD (ie, the relative risks were infinite).
Higher levels of anger were also associated with increased risks of incident angina pectoris (multivariate P for linear trend=.04). When we combined all end points (total CHD plus angina pectoris), a score of 2 to 4 was associated with a twofold increased risk, while a score of 5 to 14 was associated with a nearly threefold increased risk (multivariate P for linear trend=.008; Table 2⇑). The multivariate relative risks of a 1.0-SD increment in the anger scale were 1.27 (95% CI: 0.83 to 1.95) for total CHD and 1.27 (95% CI: 1.06 to 1.52) for combined angina and CHD. Survival curves were drawn for the three categories of anger, using the Kaplan-Meier product-limit method (Fig 2⇓).17 Men with higher levels of anger had shorter CHD-free survival, defined as the cumulative probability of remaining free of any coronary event (P=.04, using the log-rank test to compare the three survival curves).
When we examined the responses to the Cynicism Content scale of MMPI-2, the mean score among 1305 participants was 9.22 (SD, 4.98; range, 0 to 23). The correlation between anger score and cynicism score in this cohort was .496 (P<.0001). However, no association was evident between cynicism score (modeled as a continuous variable) and any CHD end point (Table 3⇓). Similarly, no association was apparent between cynicism categorized into tertiles and combined CHD/angina.
Since responses to the anger scale and cynicism scale were moderately correlated, we repeated all the analyses of anger and CHD, adjusting for cynicism score (modeled as a continuous variable) in addition to all the other coronary risk factors. In these analyses, the associations between anger and CHD became somewhat stronger for all the end points examined. For example, the multivariate-adjusted relative risks of total CHD were 2.77 and 3.33, respectively, for men scoring 2 to 4 and 5 to 14 on the anger scale. The corresponding relative risks for angina were 1.76 and 2.74. For all end points combined (total CHD plus angina), the multivariate-adjusted relative risks were 2.11 (95% CI: 0.98 to 4.52) for men scoring 2 to 4 and 2.90 (95% CI: 1.32 to 6.35) for men scoring 5 to 14 on the anger scale.
Finally, we examined potential modification of the relative risks by medication use. Following the suggestive findings of a recent study,5 we stratified our analyses according to the use of two classes of drugs: aspirin and β-adrenergic antagonists. Use of these medications was assessed at the same time as anger symptoms were assessed. Because of the limited number of individuals on regular medication, we examined the risk of total CHD combined with angina. The use of β-blockers did not appear to modify the risk of CHD: 9 events occurred among 92 β-blocker users, yielding a relative risk of 1.30 (95% CI: 0.62 to 2.72) per SD increase in the anger score. This value compared with a relative risk of 1.27 (95% CI: 1.05 to 1.52) among 1202 nonusers of β-blockers (based on 99 CHD events). The relative risk of CHD associated with anger appeared to be lower among aspirin users (0.85; 95% CI: 0.53 to 1.34) than among nonusers (1.41; 95% CI: 1.16 to 1.72). Twenty-seven CHD events occurred among 157 aspirin users compared with 81 events among 1137 nonusers.
These prospective data, though based on relatively small numbers of end points, suggest that anger is associated with a twofold to threefold increase in the risk of total CHD and angina pectoris, even after adjusting for potential confounding by behaviors such as cigarette smoking. Our findings extend the recent epidemiological finding that anger may trigger the onset of acute MI5 by demonstrating a more long-term increase in risk of CHD among older males.
Several studies have examined the association between hostility and CHD, with mixed results.11 12 18 19 20 21 Hostility has often been assessed by the Cook-Medley Hostility Scale,22 which was made up of 50 items derived from the original MMPI. The Cook-Medley Scale was unavailable in the MMPI-2. Despite its name, the Cook-Medley Hostility Scale contains items that have more in common with the MMPI-2 cynicism scale than with the anger scale. Based on factor analysis of responses from 1002 men and women, Costa et al23 identified two subscales of the Cook-Medley Scale, which they termed “Cynicism” and “Paranoid Alienation.” The authors concluded that, “Despite Cook and Medley's choice of the term ‘hostility’ for the full scale, neither of its subscales measure anger, irritability, or aggression. . . . the term ‘hostility’ is somewhat misleading as a description of this trait: a better label might be ‘Cynical Mistrust.’”23
The Cook-Medley Scale and the MMPI-2 Anger scale share only one common item (ie, the statement that “I have at times had to be rough with people who were rude or annoying” [see “Appendix”]). Thus, our findings with respect to anger and CHD stand apart from most previous reports of hostility and CHD that used the Cook-Medley Scale.
Cynicism and anger are conceptually distinct. Cynicism is characterized by a tendency to hold misanthropic beliefs, including the attribution of selfish motives to other people's acts.9 Anger, by contrast, refers to the unpleasant emotional state ranging in intensity from mild irritation or annoyance to rage and fury, usually in response to perceived mistreatment or provocation.24 The present study found no evidence of an association between cynicism and CHD. Barefoot et al25 identified three subscales derived from the Cook-Medley Scale that predicted mortality in a 29-year follow-up study of 118 individuals. The subscales were labeled “Cynicism” (13 items), “Hostile Affect” (5 items), and “Aggressive Responding” (9 items). However, none of these subscales shared any items in common with the MMPI-2 Anger scale.
The biological mechanisms by which anger may increase the risk of CHD include discharge of circulating catecholamines, increased myocardial oxygen demand, vasospasm, and increased platelet aggregability. Verrier and colleagues26 reported a dog model in which anger elicited increased coronary vascular resistance and ST-segment changes. In a study of 12 patients with symptomatic myocardial ischemia, a task involving recall of a recent event that produced anger resulted in acute vasoconstriction in previously narrowed coronary arteries.7 Reported anger was significantly correlated with decreases in both mean and minimal diameters of narrowed coronary arteries, as demonstrated by quantitative angiographic techniques. The highest correlations with arterial narrowing were demonstrated with the Anger-Out subscale of the Anger Expression scale,10 which measures higher outward expression of anger (such as swearing, exploding, and wanting to hurt objects or people) and which has much in common with the MMPI-2 Anger scale used in the current study.
Although the exact mechanisms by which anger might raise the risk of CHD remain uncertain, a recent study5 reported that aspirin appeared to mitigate the effects of anger on CHD risk: regular users were at approximately half the risk compared with nonusers. The authors speculated that the protective effect of aspirin was due to the prevention of platelet aggregation leading to thrombus formation.5 A moderate protective effect of aspirin in the presence of anger was also suggested by the present study (relative risk for cardiovascular events of 0.85 in users versus 1.41 among nonusers). Further studies are needed to confirm whether this protective effect is real.
A limitation of the present study is that it was based on a predominantly white male cohort. Future studies need to examine this association in women and in nonwhite populations. Anger may be related to other confounding factors, such as diet, which were not taken account of in these analyses. On the other hand, there was no relationship of anger level to serum cholesterol levels. Also, the magnitude of the relative risks makes it less likely that the observed associations were caused by confounding.
The clinical implications of an elevated anger score require further clarification. In the present study, even men who scored 2 to 4 on the anger scale appeared to be at increased risk of CHD relative to men who scored 0 or 1. Yet the average anger score for the whole cohort was 4.6 (SD, 3.1). In other words, we observed a gradient in CHD risk, with evidence of increased risk even among men with apparently “average” levels of anger.
Given the difficulties of modifying the social context in which individuals experience anger, one approach to reduce CHD risk might be to suggest a program of anger management. However, further etiologic studies of anger and CHD in different populations are warranted before recommending intervention trials to establish whether certain individuals could benefit from such programs.
Selected Abbreviations and Acronyms
|CHD||=||coronary heart disease|
|MMPI-2||=||revised Minnesota Multiphasic Personality Inventory|
The Anger Content Scale of MMPI-2
Indicate true or false to the following statements, as applied to you:
1. At times I feel like swearing.
2. At times I feel like smashing things.
3. Often I can't understand why I have been so irritable and grouchy.
4. At times I feel like picking a fist fight with someone.
5. I easily become impatient with people.
6. I am often said to be hotheaded.
7. I am often so annoyed when someone tries to get ahead of me in a line of people that I speak to that person about it.
8. I have at times had to be rough with people who were rude or annoying.
9. I am often sorry because I am so irritable and grouchy.
10. It makes me angry to have people hurry me.
11. I am very stubborn.
12. Sometimes I get so angry and upset I don't know what comes over me.
13. I have gotten angry and broken furniture or dishes when I was drinking.
14. I have become so angry with someone that I have felt as if I would explode.
15. I've been so angry at times that I've hurt someone in a physical fight.
16. I almost never lose self-control.*
*Item is reverse-scored; for all others, true=1; false=0. Range of possible scores: 0 to 16.
This study was supported by the National Heart, Lung, and Blood Institute (grant HL45089) and the National Institute on Aging (grant AG02287), Bethesda, Md, and by the Medical Research and Health Services Research and Development Services of the Department of Veterans Affairs.
- Received February 13, 1996.
- Revision received May 13, 1996.
- Accepted May 20, 1996.
- Copyright © 1996 by American Heart Association
Koskenvuo M, Kaprio J, Rose RJ, Kesaniemi A, Sarna S, Heikkila K, Langinvainio H. Hostility as a risk factor for mortality and ischemic heart disease in men. Psychosom Med. 1988;50:153-164.
Dembroski TM, MacDougall JM, Costa PT Jr, Grandits GA. Components of hostility as predictors of sudden death and myocardial infarction in the Multiple Risk Factor Intervention Trial. Psychosom Med. 1989;51:514-522.
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;92:1720-1725.
Butcher JN, Dahlstrom WG, Graham JR, Tellegen A, Kaemmer B. MMPI-2: Minnesota Multiphasic Personality Inventory-2, Manual for Administration and Scoring. Minneapolis, Minn: University of Minnesota Press; 1989.
Spielberger CD. Professional Manual for the State-Trait Anger Expression Inventory. Research ed. Tampa, Fla: University of South Florida; 1988.
Williams RB Jr, Janey TL, Lee KL, Kong Y, Blumenthal JA, Whalen RE. Type A behavior, hostility, and coronary atherosclerosis. Psychosom Med. 1980;42:539-549.
Shekelle RB, Gale M, Ostfeld AM, Paul O. Hostility, risk of coronary heart disease, and mortality. Psychosom Med. 1983;45:109-114.
Butcher JN, Graham JR, Williams CL, Ben-Porath YS. Development and Use of the MMPI-2 Content Scales. Minneapolis, Minn: University of Minnesota Press; 1990.
Shurtleff D. Some Characteristics Related to the Incidence of Cardiovascular Disease and Death: Framingham Study, 18-Year Follow-up. Bethesda, Md: 1974. US Dept of Health, Education, and Welfare publication 74-599.
Department of Health, Education, and Welfare. International Classification of Diseases Adapted for Use in the United States. 8th revision. Washington, DC: 1963. US Dept of Health, Education, and Welfare publication PHS 1693.
SAS/STAT User's Guide. Version 6, 4th ed. Cary, NC: SAS Institute Inc; 1990.
McCranie EW, Watkins L, Brandsma J, 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. 1985;9:119-125.
Costa PT Jr, Zonderman AB, McCrae RR, Williams RB Jr. Cynicism and paranoid alienation in the Cook Medley HO scale. Psychosom Med. 1986;48:283-285.
Barefoot JC, Dodge KA, Peterson BL, Dahlstrom G, Williams RB Jr. The Cook-Medley Hostility Scale: item content and ability to predict survival. Psychosom Med. 1989;51:46-57.
Verrier RL, Hagestad EL, Lown B. Delayed myocardial ischemia induced by anger. Circulation. 1987;75:249-254.