This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Williams, J. E. Articles by Algra, A. Search for Related Content PubMed PubMed Citation Articles by Williams, J. E. Articles by Algra, A. Related Collections Behavioral/psychosocial - stroke Acute Cerebral Hemorrhage Acute Cerebral Infarction Acute Stroke Syndromes Epidemiology

(Stroke. 2002;33:13.)
© 2002 American Heart Association, Inc.

Original Contributions

The Association Between Trait Anger and Incident Stroke Risk

The Atherosclerosis Risk in Communities (ARIC) Study

Janice E. Williams, PhD, MPH; F. Javier Nieto, MD, PhD; Catherine P. Sanford, MSPH; David J. Couper, PhD Herman A. Tyroler, MD

From the Centers for Disease Control and Prevention, Atlanta, Ga (J.E.W.); the Department of Epidemiology, Johns Hopkins University Bloomberg School of Hygiene and Public Health, Baltimore, Md (F.J.N.); the Injury and Violence Prevention Unit, North Carolina Department of Health and Human Services, Raleigh, NC (C.P.S.); and the Department of Biostatistics (D.J.C.) and the Department of Epidemiology (H.A.T.), School of Public Health, University of North Carolina, Chapel Hill.

Correspondence to Janice E. Williams, PhD, MPH, Centers for Disease Control and Prevention, 4770 Buford Hwy, NE, Mailstop K-47, Atlanta, GA 30341–3717. E-mail jwill22{at}bellsouth.net

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
Background and Purpose— This study examined the relation between trait anger and incident stroke risk among participants without a history of stroke at the first follow-up examination of the Atherosclerosis Risk in Communities (ARIC) study.

Methods— The study sample included 13 851 black and white men and women, aged 48 to 67 years, who completed the Spielberger Trait Anger Scale. Median follow-up time was 77.3 months.

Results— In the full cohort, Cox proportional hazards regression analyses showed a modest increase in the risk for stroke among individuals with high trait anger, though the association did not remain statistically significant after multivariate adjustment. Participants 60 years of age who reported having high trait anger had a 2.82 (95% CI, 1.65 to 4.80) times greater risk for hemorrhagic and ischemic strokes combined (any) and a 2.93 (95% CI, 1.64 to 5.22) times greater risk for ischemic strokes alone than their counterparts who reported having low trait anger (hazard rate ratios adjusted for sex and race/ethnicity). Similarly, among participants with HDL cholesterol levels >47, the risk for any stroke was 2.86 (95% CI, 1.56 to 5.25) times greater for those who reported having high trait anger, whereas the risk for ischemic strokes alone was 2.98 (95% CI, 1.58 to 5.61) times greater (hazard rate ratios adjusted for age, sex, and race/ethnicity). These associations remained strong and statistically significant after further adjustment for several established biological and sociodemographic risk factors for stroke and were absent among older participants and those with lower HDL cholesterol values.

Conclusions— Trait anger was associated with an increased risk for incident stroke in the ARIC study among younger participants and those with higher HDL cholesterol levels.

Key Words: anger • prospective studies • risk factors • stroke, ischemic • survival analysis

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
Anger/hostility can have deleterious consequences for physical health, most notably coronary heart disease (CHD)^1–4 and cerebrovascular disease or stroke.^5–8 Although a considerable body of epidemiologic and clinical research data supports a positive association between anger and CHD, fewer data are available for stroke. The extant research literature on anger and stroke consists of only a handful of studies. Adler et al,⁵ in a retrospective case study, reported that among men in their sample, stroke was most often preceded by negative affective states, primarily hopelessness and anger. Similarly, Gianturco et al⁶ reported that compared with hospitalized control subjects, a greater proportion of stroke patients had anger just before their stroke. Although the full report from the Framingham Heart Study is unavailable, in an abstract summarizing its findings, Eaker and Feinleib⁷ reported that the 10-year incidence of stroke was significantly associated with anger among women and

See Editorial Comment, page 19

marginally associated with anger among men; however, the association was no longer statistically significant after risk factor adjustment. The most recent published work in this area is a prospective study by Everson et al,⁸ based on results from the Kuopio Ischemic Heart Disease Study. After examining the association between different anger expression styles and the incidence of stroke among middle-aged Finnish men, these investigators found that men who were most prone to frequent outward expressions of anger had twice the risk for stroke as those who were least prone to such expressions. Furthermore, among men with prevalent CHD, they found those who were most prone to outwardly expressed anger had almost 6 times the risk for stroke as those who were least prone.

In the literature, anger is conceptualized in terms of the degree to which people have this emotion and the characteristic means by which they express it.⁹ The predisposition for frequent, intense, long-lasting anger is a relatively enduring and stable personality attribute known as trait anger. Anger expression, on the other hand, refers to how anger is managed, that is, whether it is expressed outwardly, held in, or controlled. In the current study, we examined the relation of trait anger to incident stroke risk in a large bi-ethnic US population of middle-aged men and women enrolled in the Atherosclerosis Risk in Communities (ARIC) study.¹⁰ Analyses of ARIC data and results of other studies have shown the relation between anger/hostility and cardiovascular disease (CVD) risk to be modified by hypertensive status,^3,11 prevalent CHD,⁸ and age.^12,13 Taken together, these findings suggest that anger/hostility may confer a greater CVD risk among people who otherwise are at low risk for the disease. The aims of this study were (1) to assess the relation between trait anger and stroke incidence and 2) to test whether this relation differed by levels of selected risk factors.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
Study Population
The study population consisted of participants in ARIC, a large, prospective study of cardiovascular disease and its risk factors carried out in the US communities of Washington County, Maryland; suburban Minneapolis, Minnesota; Forsyth County, North Carolina; and Jackson, Mississippi.¹⁰ ARIC study participants were recruited from random probability samples enumerated in these communities. Written informed consent to participate in the ARIC study was obtained from each participant at the baseline clinical examination (visit 1) and updated at each follow-up examination (visits 2, 3, and 4). The study was approved trienneally by the institutional review board at each participating university, as were the study protocols by the ARIC Steering Committee, the ARIC Policy Board, and the National, Heart, Lung, and Blood Institute. Blacks were oversampled in Forsyth County, North Carolina, and only blacks were sampled in Jackson, Mississippi. The ARIC study was designed with two distinct research arms: (1) community surveillance of residents 35 to 74 years of age and (2) cohorts of volunteers who were 45 to 64 years of age at recruitment. The baseline cohort evaluations were conducted in 1987 to 1989 and were followed by 3 triennial clinical examinations and ongoing morbidity and mortality surveillance.

At the second clinical examination (visit 2), conducted between 1990 and 1992, the Spielberger Trait Anger Scale⁹ was administered to 14 348 returning cohort members, 92.9% of the cohort members enrolled at baseline. Of these, 148 with incomplete responses to the anger questionnaire and 42 with a racial/ethnic identity other than black or white were excluded from analyses. An additional 307 were excluded because of a history of stroke at visit 2. After these sequential exclusions were made, a total of 13 851 participants remained for the current analyses.

Assessment of Trait Anger
The Spielberger Trait Anger Scale⁹ was used to assess the frequency and degree to which each participant had anger (see Appendix). This scale was a component of the Health-Life Profile that participants completed at visit 2. The Spielberger scale consists of 10 items endorsed on a 4-point anchor, including almost never=1, sometimes=2, often=3, and almost always=4. The sum of the response category for each of the individual items comprised the overall trait anger score, which ranged from 10 to 40. High trait anger was defined by scores of 22 to 40, moderate anger by scores of 15 to 21, and low anger by scores of 10 to 14. The chosen cut-points were comparable with those used in previously published work with the Spielberger scale.^14–16

Adequate internal consistency has been reported for the Spielberger scale (r=0.81 to 0.92).¹⁷ Correlations with the Buss-Durkee Hostility Inventory have been reported as ranging from 0.66 to 0.73 and 0.43 to 0.59 with the Cook-Medley Hostility Scale.¹⁷ A study in ARIC has been planned to determine the internal consistency and stability of the Spielberger scale in this cohort.

Assessment of Stroke Risk Factors
Technicians who were trained and certified in the use of standardized protocols made the physiological measurements and laboratory assessments during the clinical examination at visit 2. Three consecutive blood pressure measurements, assessed with a random zero sphygmomanometer, were taken after participants were seated for a 5-minute rest period. The blood pressure levels reported were the average of the second and third readings. Waist-to-hip ratio was the ratio of waist girth (abdominal circumference measured at the umbilicus) to hip girth (circumference measured at the maximal gluteal protrusion). Diabetes was defined as a fasting serum glucose level 126 mg/dL, a nonfasting serum glucose level >200 mg/dL, and/or a history of diabetes, insulin therapy, or oral hypoglycemic medication use. Data on alcohol consumption, cigarette smoking, and education were based on self-reports by participants. von Willebrand factor antigen was measured by the enzyme-linked immunosorbent assay method. Left ventricular hypertrophy was determined by the Cornell voltage electrocardiographic criteria. HDL cholesterol (HDL-C) levels were measured enzymatically after precipitation of LDL-containing lipoproteins with dextran-magnesium. LDL cholesterol (LDL-C) levels were calculated by means of the Friedewald formula. Prevalent CHD was defined as a history of myocardial infarction or cardiac revascularization procedures (eg, coronary artery bypass graft surgery or percutaneous transluminal angioplasty).

Ascertainment of Incident Stroke
Incident strokes were hospitalized events among members of the ARIC study cohort captured from the time of their second clinical examination (1990 to 1992) through December 31, 1997. In this study, strokes were classified as ischemic or hemorrhagic.

Potential strokes were identified through annual telephone interviews with cohort members and through ongoing morbidity and mortality surveillance in local hospitals. Hospital records (eg, medical records and hospital discharge summaries) were reviewed for the presence of a diagnosis consistent with cerebrovascular disease (eg, International Classification of Diseases, ninth revision, Clinical Modification codes 430–438), a CT or MRI image diagnostic of cerebrovascular disease, or a patient history of hospitalization on the neurological intensive care unit.¹⁸ Potentially eligible cases were subjected to classification and validation through the use of a computerized algorithm and physician review. Relevant data from hospital records used in the classification scheme included the type and duration of patients’ initial neurological symptoms, their medical history, results of medical procedures, medications, reports from imaging (CT or MRI), autopsy findings, and other supportive clinical evidence. The inclusion criterion for a diagnosis of stroke was evidence of sudden or rapid onset of neurological symptoms (eg, paralysis, numbness/weakness, aphasia, headache, vertigo, convulsions) lasting for >24 hours. Disagreements between computer-generated diagnoses and those derived by physician review were adjudicated by a second physician reviewer. Classification of stroke was conducted without knowledge of the participant’s status on the trait anger scale. A more detailed description of the procedures and definition for classification of stroke events in the ARIC cohort have been previously reported.¹⁸

Statistical Analyses
Anger classes (low, middle, high) were entered into Cox proportional hazards regression models as dummy variables in the main analyses. Anger was entered into the models as a continuous variable when assessing the presence of a linear trend in stroke risk. The association between trait anger and incident stroke risk was examined for effect modification by each of the following risk factors: age (continuous), HDL-C level (continuous), LDL-C level (continuous), sex (dichotomous), hypertensive status (dichotomous), diabetes (dichotomous), prevalent CHD (dichotomous), and cigarette years of smoking (continuous). These covariates were selected on the basis of their convergence with effect modifiers identified in previous studies of anger/hostility and CVD^3,8,11–13 and their convergence with the risk factors constituting the CHD risk score as set forth by Wilson et al in the Framingham study.¹⁹ Effect modification was assessed by entering into a Cox model a term consisting of the product of the continuous trait anger variable and the covariate. The significance of each term was determined by the -2 log likelihood ratio test. Age and HDL-C level were statistically significant effect modifiers, having probability values of <0.01 and 0.04, respectively. Therefore, subsequent analyses were stratified by age and HDL-C level. A cut-point of age 60 years was chosen a priori to be consistent with a previous study in which an age-dependent relation in the anger/hostility and CVD association was observed.¹² The median value of the HDL-C distribution for the cohort, 47, was chosen as the cut-point for HDL-C. Incrementally adjusted (eg, crude, age- [where appropriate], sex-, and race/ethnicity-adjusted, and multivariate-adjusted) models were fit for hemorrhagic and ischemic strokes combined (any) and for ischemic strokes alone. Crude probabilities of stroke-free survival were determined by the Kaplan-Meier product-limit method.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
The distribution of trait anger was as follows: among persons 60 years of age and younger, mean trait anger scores were 15.6 (SD=3.9) for black men, 16.1 (SD=3.9) for black women, 16.5 (SD=3.8) for white men, and 16.1 (SD=3.5) for white women. Correspondingly, among those >60 years of age, mean trait anger scores were 15.7 (SD=4.6) for black men, 15.8 (SD=4.2) for black women, 6.1 (SD=3.7) for white men, and 15.5 (SD=3.3) for white women.

Selected characteristics of the study population at the first follow-up examination, shown in Table 1, indicate that participants who scored in the high trait anger range were slightly younger, more likely to be male, and more likely to have fewer years of formal education than those who scored in the moderate- or low- anger range. High trait anger scorers also had higher von Willebrand factor values, lower HDL-C levels, and higher waist-to-hip ratios than moderate- or low-anger scorers. Furthermore, participants who scored in the high trait anger range smoked more cigarettes, consumed more alcohol, and were more likely than their counterparts to have a history of CHD.

View this table: [in this window] [in a new window]   Table 1. Distribution of Population Characteristics by Level of Trait Anger: ARIC Study, 1990 to 1992

Of the 257 participants who had strokes in the study’s 96.2-month follow-up period (median period, 77.3 months), 31 scored in the high anger range, 136 in the moderate anger range, and 90 in the low anger range. Of these strokes, 226 were ischemic and 31 hemorrhagic.

In the full cohort, there was a modest association between trait anger and incident stroke (any) risk, though the multivariate-adjusted hazard rate ratio (HRR) was not statistically significant. The crude association between trait anger (high versus low) and any stroke was 1.60 (95% CI, 1.07 to 2.41). The HRRs for any stroke were 1.73 (95% CI, 1.15 to 2.60) adjusting first for age; 1.71 (95% CI, 1.14 to 2.58) adjusting second for age, race/ethnicity, and sex; and 1.19 (95% CI, 0.75 to 1.89) adjusting third for all the covariates. Similar results were obtained for ischemic strokes alone, with the HRRs being slightly higher in general and the HRRs in the fully adjusted model again not reaching statistical significance.

Among the younger participants (age 60 years), the risk of incident stroke increased monotonically with increasing levels of trait anger (Table 2). Without adjustment, the risk for any stroke among high-anger scorers was 2.67 times greater than that of low-anger scorers (P value for linear trend=0.004). After adjustment for sex and race/ethnicity, the risk was 2.82 times greater (P value for linear trend=0.001). Furthermore, the HRRs for ischemic stroke in the crude model (2.78) and in the model adjusted for sex and race/ethnicity (2.93) were both higher than those for any stroke when the high trait anger group was compared with the low (Table 2). These associations remained strong and statistically significant after multivariate adjustment; they were absent among the older participants (age >60 years). Crude Kaplan-Meier probability curves for the two age groups are displayed in the Figure (Panels 1 and 2).

View this table: [in this window] [in a new window]   Table 2. Hazard Rate Ratios (95% CI) for Association Between Trait Anger and Incident Stroke by Age: ARIC Study, 1990 to 1997

View larger version (31K): [in this window] [in a new window]   Crude stroke-free survival probabilities by trait anger scores: ARIC Study, 1990 to 1997. Panel 1, Age 60 years; Panel 2, age >60 years; Panel 3, HDL-C levels >47; Panel 4. HDL-C levels 47.

The pattern of risk among participants in the two HDL-C strata was similar to that observed in the two age strata, as described above. Among participants with HDL-C values above the median (>47), the risk for any stroke increased monotonically with levels of trait anger (Table 3). However, these associations were not observed among participants with HDL-C levels at or below the median (47). Additionally, among participants with higher HDL-C levels (>47), each HRR for ischemic strokes alone was higher than that for any stroke. Crude Kaplan-Meier probability curves for the two HDL-C groups are displayed in the Figure (Panels 3 and 4).

View this table: [in this window] [in a new window]   Table 3. Hazard Rate Ratios (95% CI) for Association Between Trait Anger and Incident Stroke by HDL Cholesterol Levels: ARIC Study, 1990 to 1997

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
In the full cohort, trait anger was associated with only a modest increased stroke risk. This personality trait, however, was associated with a significant increase in the risk for stroke, especially ischemic stroke, among younger men and women and those with higher HDL-C levels. Younger participants (60 years) who reported having high trait anger were at nearly 3 times greater risk for ischemic stroke than those who reported having low trait anger. Among participants with higher HDL-C levels (>47) and high trait anger, the magnitude of risk was similar. These associations remained strong and statistically significant after further adjustment for several established biological and sociodemographic risk factors for stroke and were absent among older persons and participants with lower HDL-C levels.

The mechanism by which trait anger increases ischemic stroke risk is unknown. The research on anger and CHD may help explain the association, since both ischemic stroke and the acute coronary syndromes (eg, myocardial infarction, sudden death, angina pectoris) result from an occlusive arterial process. In the anger-CHD literature, anger is viewed as a trigger of hemodynamic (eg, increased arterial pressure), vasoconstrictive, and hemostatic (eg, increased platelet adhesion and aggregability, increased plasma coagulation, and fibrinolysis) forces that are key in the pathogenesis of the acute coronary syndromes.^2,20 They contribute to the atherogenic process by disrupting vulnerable atherosclerotic plaques in the artery wall, leading to the formation of lesions and to occlusive thrombosis. The triggering hypothesis may be a useful conceptual framework from which to understand the link between trait anger and incident stroke; that is, anger may trigger similar physiological changes in the vessels supplying the brain with blood, causing occlusive thrombosis, blockage, and a resulting stroke.

Another pathway by which anger may be associated with ischemic stroke is through the impact of sympathetic arousal and neuroendocrine activation in the initiation and progression of atherosclerosis. There can be direct injury to the blood vessel walls in response to increased blood pressure and increased flow velocity and in response to an exaggerated discharge of catecholamines (eg, epinephrine and norepinephrine).^21,22 Furthermore, catecholamines can initiate other vascular and prothrombotic events that are associated with atherosclerosis such as increased platelet adhesion and aggregation,²³ vascular lipid uptake,²⁴ and activation of macrophages.²⁵ Parenthetically, in this study, the prevalence of hypertension was only slightly higher in persons in the high-anger group compared with that in the low- or moderate-anger groups. In view of the sympathetic arousal hypothesis, a much higher prevalence of hypertension in the high-anger group would have been expected. Future studies in this cohort are needed to assess the relation of anger to blood pressure levels as well as to the development and progression of hypertension.

A third hypothesis suggests that persons who have higher levels of anger/hostility, compared with those who have lower levels, may be more likely to engage in adverse health behaviors (eg, sedentary lifestyles, cigarette smoking, greater consumption of alcohol, and overeating) that place them at increased risk for CHD.^4,26 These behaviors and the resulting risk factors are viewed as potential mediators on the causal pathway from anger/hostility to CHD. Given the similarities between CHD and stroke risk factors, this hypothesis may help explain the mechanism linking trait anger to stroke. If a biobehavioral mechanism is chiefly operative in the trait anger-stroke relation, then the multivariate estimates of risk reported in this study may underestimate the strength of the association.

We observed neither a statistically significant nor a consistently positive association between trait anger and incident stroke risk among older participants (age >60 years) in the cohort. This finding is consistent with research on hostility and CHD in which the investigators have observed heterogeneity of effect by age and explained the association from the perspective of selective survival.^12,13 This finding is also consistent with that from the Framingham study,²⁷ in which the influence of smoking on CHD risk decreased with age. Alternatively, the lack of a positive association between trait anger and incident stroke risk among participants over age 60 years may be explained, at least in part, by the fact that as a group they were at 2.59 times greater risk for stroke, and any further risk attributable to anger was negligible.

Another main finding of the current study was the lack of a positive association between trait anger and incident stroke risk among participants with lower HDL-C levels. In this study, there was an inverse association between incident stroke risk and higher HDL-C values (relative risk, 0.59; 95% CI, 0.45 to 0.76). Therefore, the lack of a statistically significant or a consistently positive association between trait anger and incident stroke risk among participants with lower concentrations of HDL-C, a known antiatherogenic lipoprotein, may be explained by the fact that in this context, the addition of an anger-prone personality confers little or no additional risk for stroke.

We are aware of only one other published prospective study of anger and stroke⁸ and none involving a bi-ethnic sample of white and black men and women. Our study contained a relatively large number of strokes that were validated by standardized diagnostic criteria. Additionally, this study, like a study reported recently by Everson et al,⁸ was based on data gathered with a well-known, reliable, and valid anger assessment instrument. Thus, in both exposure and outcome assessment, we were able to overcome some of the methodological weaknesses that plagued the earliest studies in this area. However, even though the process of ascertaining and validating strokes was standardized, there could have been errors in the classification of disease resulting from the incompleteness or inaccuracy of data abstracted from medical records and discharge summaries.

Persons who have high levels of anger may benefit from interventions designed to help them better handle anger-provoking situations. There is some evidence that anger management is effective for lowering anger/hostility levels, improving CHD risk factor profiles, and forestalling the recurrence of CHD events.^28,29 The efficacy of anger reduction for the prevention of stroke awaits evaluation.

In conclusion, for the entire ARIC study cohort, trait anger was modestly associated with the risk for incident stroke. However, among younger persons and those with higher HDL-C levels, these associations were stronger and statistically significant even after adjusting for potential confounders. These results provide additional epidemiologic evidence of a positive association between anger and stroke. Although there are tenable hypotheses linking anger and ischemic stroke, additional studies are needed to identify the operative biological and behavioral mechanisms and to explore the effect modification of this association by age and HDL-C level.

	Acknowledgments

 
Support for this research was provided by the National Heart, Lung, and Blood Institute contracts NO1-HC-55015, NO1-HC-55016, NO1-HC-55018, NO1-HC-55019, NO1-HC-55020, NO1-HC-55021, and NO1-HC-55022.

The authors thank the staff and participants in the ARIC Study for their important contributions.

Received March 6, 2001; revision received August 23, 2001; accepted September 4, 2001.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
1. Kawachi I, Sparrow D, Spiro AIII, Vokonas P, Weiss ST. A prospective study of anger and coronary heart disease: the Normative Aging Study. Circulation. 1996; 94: 2090–2095.[Abstract/Free Full Text]

2. 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.[Abstract/Free Full Text]

3. Williams JE, Paton CC, Siegler IC, Eigenbrodt ML, Nieto FJ, Tyroler HA. Anger proneness predicts coronary heart disease risk: prospective analysis from the Atherosclerosis Risk in Communities (ARIC) Study. Circulation. 2000; 101: 2034–2039.[Abstract/Free Full Text]

4. Everson SA, Kauhanen J, Kaplan GA, Goldberg DE, Julkunen J, Tuomilehto J, Salonen JT. Hostility and increased risk of mortality and acute myocardial infarction: the mediating role of behavioral risk factors. Am J Epidemiol. 1997; 146: 142–152.[Abstract/Free Full Text]

5. Adler R, MacRitchie K, Engel GL. Psychologic processes and ischemic stroke (occlusive cerebrovascular disease), I: observations on 32 men with 35 strokes. Psychosom Med. 1971; 33: 1–29.[Abstract/Free Full Text]

6. Gianturco DT, Breslin MS, Heyman A, Gentry WD, Jenkins CD, Kaplan B. Personality patterns and life stress in ischemic cerebrovascular disease, I: psychiatric findings. Stroke. 1974; 5: 453–460.[Abstract/Free Full Text]

7. Eaker ED, Feinleib M. Psychosocial factors and the ten-year incidence of cerebrovascular accident in the Framingham Heart Study. Psychosom Med. 1983; 45: 84.Abstract.

8. Everson SA, Kaplan GA, Goldberg DE, Lakka TA, Sivenius J, Salonen JT. Anger expression and incident stroke: prospective evidence from the Kuopio Ischemic Heart Disease Study. Stroke. 1999; 30: 523–528.[Abstract/Free Full Text]

9. Spielberger CD, Krasner SS, Solomon EP. The experience, expression, and control of anger.In: Janisse MP, ed. Health Psychology: Individual Differences and Stress. New York, NY: Springer-Verlag; 1988: 89–108.

10. The ARIC Investigators. The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. Am J Epidemiol. 1989; 129: 687–702.[Abstract/Free Full Text]

11. Williams JE, Nieto FJ, Sanford CP, Tyroler HA. Effects of an angry temperament on coronary heart disease risk: the Atherosclerosis Risk in Communities (ARIC) Study. Am J Epidemiol. 2001; 154: 230–235.[Abstract/Free Full Text]

12. Siegman AW, Dembroski TM, Ringel N. Components of hostility and the severity of coronary artery disease. Psychosom Med. 1987; 49: 127–135.[Abstract/Free Full Text]

13. Meesters CMG, Smulders J. Hostility and myocardial infarction in men. J Psychom Res. 1994; 38: 727–734.

14. Denollet J, Brutsaert DL. Personality, disease severity, and the risk of long-term cardiac events in patients with a decreased ejection fraction after myocardial infarction. Circulation. 1998; 97: 167–173.[Abstract/Free Full Text]

15. Lopez FG, Thurman CW. High-trait and low-trait angry college students: a comparison of family environments. J Counseling Dev. 1993; 71: 524–527.

16. Deffenbacher JL, Oetting ER, Thwaites GA, Lynch RS, Baker DA, Stark RS, Thacker S, Eiswerth-Cox L. State-trait anger theory and the utility of the trait anger scale. J Counseling Psych. 1996; 43: 131–148.

17. Spielberger CD, Jacobs G, Russell S, Crane R. Assessment of anger: the State-Trait Anger Scale.In: Butcher JN, Spielberger CD, eds. Advances in Personality Assessment. Vol 2. Hillsdale, NJ: Erlbaum; 1983: 161–189.

18. Rosamond WD, Folsom AR, Chambless LE, Wang C, McGovern PG, Howard G, Cooper LS, Shahar E. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke. 1999; 30: 736–743.[Abstract/Free Full Text]

19. Wilson PWF, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation. 1998; 97: 1837–1847.[Abstract/Free Full Text]

20. Muller JE, Abela GS, Nesto RW, Tofler GH. Triggers, acute risk factors and vulnerable plaques: the lexicon of a new frontier. J Am Coll Cardiol. 1994; 23: 809–813.[Abstract]

21. Kaplan JR, Pettersson K, Manuck SB, Olsson G. Role of sympathoadrenal medullary activation in the initiation and progression of atherosclerosis. Circulation. 1991; 84 (suppl VI): VI-23–VI-32.

22. Pauletto P, Scannapieco G, Pessina AC. Sympathetic drive and vascular damage in hypertension and atherosclerosis. Hypertension. 1991; 17 (suppl III): III-75–III-81.

23. Anfossi G, Trovati M. Role of catecholamines and platelet function: pathophysiological and clinical significance. Eur J Clin Invest. 1996; 26: 353–370.[CrossRef][Medline] [Order article via Infotrieve]

24. Born GVR. Recent evidence for the involvement of catecholamines and of macrophages in atheroslcerotic processes. Ann Med. 1991; 23: 569–572.[Medline] [Order article via Infotrieve]

25. Adams DO. Molecular biology of macrophage activation: a pathway by whereby psychosocial factors can potentially affect health. Psychosom Med. 1994; 56: 316–327.[Abstract/Free Full Text]

26. Hemingway H, Marmot M. Evidence-based cardiology: psychosocial factors in the aetiology and prognosis of coronary artery disease: systemic review of prospective cohort studies. BMJ. 1999; 318: 1460–1467.[Free Full Text]

27. Dawber TR. The Framingham Study: The Epidemiology of Atherosclerotic Disease. Cambridge, Mass: Harvard University Press; 1980.

28. Friedman M, Thoresen CE, Gill JJ, Ulmer D, Powell LH, Price VA, et al. Alteration of type A behavior and its effect on cardiac recurrences in post myocardial infarction patients: summary results of the recurrent coronary prevention project. Am Heart J. 1986; 112: 653–665.[CrossRef][Medline] [Order article via Infotrieve]

29. Lavie CJ, Milani RV. Effects of cardiac rehabilitation and exercise training programs on coronary patients with high levels of hostility. Mayo Clin Proc. 1999; 74: 959–966.[Abstract]

Editorial Comment

The Atherosclerosis Risk in Communities (ARIC) Study

Ale Algra, MD, Guest Editor

Department of Neurology and the Julius Centre for, General Practice and, Patient Oriented Research, University Medical Centre, Utrecht, the Netherlands

	Adjustment in the Anger-Stroke Relationship: How Far Should One Go?

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
The Atherosclerosis Risk in Communities (ARIC) Study Group continues to be productive with contributions on cerebrovascular disease for Stroke from a large multi-ethnic, population-based study. In this issue, Williams et al report on the relationship between trait anger and the risk of stroke. Trait anger was evaluated by means of the Spielberger Trait Anger Scale and subdivided into 3 categories: high (scores 22 to 40), moderate (15 to 21), and low (10 to 14). During a median follow-up of about 6.5 years, 31 strokes occurred among 1103 participants with high scores (2.8%) as compared with 90 strokes in 5091 participants with low scores (1.8%). Participants with moderate scores had a risk similar to that of participants with low scores. The crude (ie, unadjusted) hazard ratio of high versus low scores was 1.60 (95% CI 1.07 to 2.41), virtually identical to the cumulative incidence ratio of 1.59 (2.81%/1.77%). The hazard ratio differed according to 2 of 8 studied subgroup characteristics: age and HDL-C level. Among persons aged 60 and below, the crude hazard ratio was 2.67 as opposed to 0.94 in the older participants; in persons with HDL-C levels above the median (47 mmol/L), the age-adjusted hazard ratio was 3.08 versus 1.10 among those with HDL-C at or below the median. Such differences were not observed between subgroups by LDL-C level, gender, hypertensive status, diabetes, prevalent coronary heart disease, and cigarette-years of smoking. The hazard ratios tended to be slightly higher if the analyses were restricted to ischemic stroke.

In this comment I limit myself to 1 methodological issue and some suggestions for future research; aspects of behavioral attributes and cerebrovascular disease have been addressed recently in 2 other editorial comments.^1,2 In the ARIC study, the authors tried to unravel the etiologic contribution of trait anger to stroke occurrence. This may be a difficult task at a time when only limited knowledge is available on the pathophysiological mechanisms involved. Such knowledge is required to determine which other characteristics may be in the causal pathway between trait anger and stroke occurrence. These insights stipulate how one should proceed during data analysis. Characteristics extraneous to the causal path from trait anger to stroke might obscure the true relationship between the two. Such characteristics are potential confounders. Rothman, in his classic book on epidemiology, describes 3 conditions for a characteristic to be a confounder³: (1) it should be related to the disease among the nonexposed; (2) it should be related to the exposure; and (3) the characteristic should not be an intermediary in the causal pathway between exposure and disease. The first 2 conditions may be investigated in one’s own database, but the third one requires a vision on the pertinent pathophysiological processes. If data on the relevant characteristics are recorded, confounding may be resolved in the data analysis, basically by stratified analyses. In case the confounding effects of many characteristics need to be addressed simultaneously, multivariate regression analyses are employed. The resulting effect estimates are "adjusted" for the nuisance effect of the confounders and may be higher or lower than the crude value. Adjustment, however, for characteristics that are intermediate in the causal path from trait anger to stroke would lead to an underestimation of the "true" strength of the relationship. It is of note that the anger-stroke relationship might be not directly causal at all, eg, trait anger could be a proxy for another truly causal factor.

Williams et al adjusted the crude hazard ratio in several steps, first for age alone, then additionally for gender and race/ethnicity, and finally also for educational attainment, waist-to-hip ratio, cigarette-years of smoking, LDL-C, diabetes, hypertensive status, von Willebrand factor, prevalent coronary heart disease, and left ventricular hypertrophy. Thus, the readers get partial insight in the confounding effects of a dozen variables. A more detailed idea of such effects may be obtained in so-called bivariate analyses, ie, with trait anger and 1 potential confounder at a time. One may select those variables for multivariate adjustment that appeared to be of influence on the hazard ratio in the bivariate analyses. Thus, a parsimonious model would arise with sufficient statistical stability. So readers may fully evaluate for themselves which effect estimate they believe to be the most plausible one. For example, should one adjust for hypertensive status? One might speculate that people with a high score of trait anger would have (more frequent episodes of) higher blood pressures, and that such blood pressure elevations would be responsible for an increased stroke incidence. In such a situation it would be inappropriate to adjust for hypertension. Although the ARIC data hardly point into this direction, the literature on this issue not yet conclusive,⁴ so one might argue both ways: adjust or do not adjust.

To try and understand the mechanisms operative in the trait anger-stroke relationship it would be of interest to study whether its strength differs according to stroke type. If the relationship is stronger with cardioembolic stroke, hemostasis might be involved in the causal route; a stronger relationship with cerebral ischemia of arterial origin would plead for atherosclerosis in the path. Another suggestion for future research would be to employ the case-crossover study design to determine whether fits of anger are directly related to stroke occurrence. In such an approach the stroke patient would serve as his or her own control. Mittleman et al applied this design to assess whether episodes of anger could trigger myocardial infarction.⁵

In conclusion, epidemiological studies should go hand in hand with more fundamental ones to solve the difficult puzzles on causation. Insight in causal mechanisms helps to develop efficient preventive and therapeutic measures.

Dr. Algra is indebted to Dr. M.L. Bots for helpful remarks on an earlier version of this comment.

	Appendix

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
Spielberger Trait Anger Scale

(1) I am quick tempered.

(2) I have a fiery temper.

(3) I am a hot-headed person.

(4) I get angry when I am slowed down by others’ mistakes.

(5) I feel annoyed when I am not given recognition for doing good work.

(6) I feel infuriated when I do a good job and get a poor evaluation.

(7) I fly off the handle.

(8) When I get angry, I say nasty things.

(9) It makes me furious when I am criticized in front of others.

(10) When I get frustrated, I feel like hitting someone.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References Adjustment in the Anger-Stroke... Appendix References

 
1. Matthews KA. Something old and something new. Stroke. 2001; 32: 1269–1270.

2. Lane DA, Carroll D, Lip GY. Cardiovascular and behavioral reactions to stress and cerebrovascular disease. Stroke. 2001; 32: 1718–1720.

3. Rothman KJ. Modern Epidemiology. 1st ed. Boston, Mass: Little, Brown, and Company, Inc; 1986.

4. Porter LS, Stone AA, Schwartz JE. Anger expression and ambulatory blood pressure: a comparison of state and trait measures. Psychosom Med. 1999; 61: 454–463.[Abstract/Free Full Text]

5. 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: Determinants of Myocardial Infarction Onset Study Investigators. Circulation. 1995; 92: 1720–1725.

This article has been cited by other articles:

				K. Shirai, H. Iso, T. Ohira, A. Ikeda, H. Noda, K. Honjo, M. Inoue, S. Tsugane, and for the Japan Public Health Center-Based Study Gro Perceived Level of Life Enjoyment and Risks of Cardiovascular Disease Incidence and Mortality: The Japan Public Health Center-Based Study Circulation, September 15, 2009; 120(11): 956 - 963. [Abstract] [Full Text] [PDF]

				N. Ranjit, A. V. Diez-Roux, B. Sanchez, T. Seeman, S. Shea, S. Shrager, and K. Watson Association of Salivary Cortisol Circadian Pattern With Cynical Hostility: Multi-Ethnic Study of Atherosclerosis Psychosom Med, September 1, 2009; 71(7): 748 - 755. [Abstract] [Full Text] [PDF]

				C. M. Caska, B. E. Hendrickson, M. H. Wong, S. Ali, T. Neylan, and M. A. Whooley Anger Expression and Sleep Quality in Patients With Coronary Heart Disease: Findings From the Heart and Soul Study Psychosom Med, April 1, 2009; 71(3): 280 - 285. [Abstract] [Full Text] [PDF]

				B. M. Wilkowski and M. D. Robinson The Cognitive Basis of Trait Anger and Reactive Aggression: An Integrative Analysis Personality and Social Psychology Review, February 1, 2008; 12(1): 3 - 21. [Abstract] [PDF]

				M. S. Player, D. E. King, A. G. Mainous III, and M. E. Geesey Psychosocial Factors and Progression From Prehypertension to Hypertension or Coronary Heart Disease Ann. Fam. Med, September 1, 2007; 5(5): 403 - 411. [Abstract] [Full Text] [PDF]

				P. G. Surtees, N. W.J. Wainwright, R. L. Luben, N. J. Wareham, S. A. Bingham, and K.-T. Khaw Adaptation to Social Adversity Is Associated With Stroke Incidence: Evidence From the EPIC-Norfolk Prospective Cohort Study Stroke, May 1, 2007; 38(5): 1447 - 1453. [Abstract] [Full Text] [PDF]

				L. L. Haheim, I. Holme, I. Hjermann, and S. Tonstad Risk-factor profile for the incidence of subarachnoid and intracerebral haemorrhage, cerebral infarction, and unspecified stroke during 21 years' follow-up in men Scand J Public Health, December 1, 2006; 34(6): 589 - 597. [Abstract] [PDF]

				A. V. Diez Roux, N. Ranjit, L. Powell, S. Jackson, T. T. Lewis, S. Shea, and C. Wu Psychosocial factors and coronary calcium in adults without clinical cardiovascular disease. Ann Intern Med, June 6, 2006; 144(11): 822 - 831. [Abstract] [Full Text] [PDF]

				S Vale Psychosocial stress and cardiovascular diseases Postgrad. Med. J., July 1, 2005; 81(957): 429 - 435. [Abstract] [Full Text] [PDF]

				M. E. Bleil, J. M. McCaffery, M. F. Muldoon, K. Sutton-Tyrrell, and S. B. Manuck Anger-Related Personality Traits and Carotid Artery Atherosclerosis in Untreated Hypertensive Men Psychosom Med, September 1, 2004; 66(5): 633 - 639. [Abstract] [Full Text] [PDF]

				E. C. Suarez C-Reactive Protein Is Associated With Psychological Risk Factors of Cardiovascular Disease in Apparently Healthy Adults Psychosom Med, September 1, 2004; 66(5): 684 - 691. [Abstract] [Full Text] [PDF]

				H. M Arthur High trait anger increased stroke in people <=60 years and those with high density lipoprotein cholesterol concentrations >47 mmol/l Evid. Based Ment. Health, August 1, 2002; 5(3): 94 - 94. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Williams, J. E. Articles by Algra, A. Search for Related Content PubMed PubMed Citation Articles by Williams, J. E. Articles by Algra, A. Related Collections Behavioral/psychosocial - stroke Acute Cerebral Hemorrhage Acute Cerebral Infarction Acute Stroke Syndromes Epidemiology