Description/Samples/Epidemiologic Evidence

HOSTILITY AND ANGER

Hostility is multidimensional, with behavioral (aggression, physical or verbal expression of hostility) cognitive (negative beliefs about others, including cynicism & mistrust), and affective (anger, ranging from irritation to rage) components.

(1) Cook-Medley Hostility Inventory (Cook & Medley, 1954). (See Ostfeld & Eaker, p.243). Used to assess cognitive component of hostility. 50 items. Derived empirically from MMPI. Costa et al. (1986) factor analysis: two highly correlated factors - suspiciousness (the belief that other intend one harm, e.g. paranoia, fear of threat to self) & cynicism (negative view of human nature as untrustworthy and selfish).

o Shekelle et al. (1983). Western Electric Study. 1877 employed middle-aged men. Ho scale was positively associated with 20-yr all-cause mortality.

o Almada SJ, Zonderman AB, Shekelle RB et al. (1991). Western Electric Study. 1871 employed middle-aged men in Chicago. MMPI Cynicism Factor was an independent predictor of 25-yr cardiac mortality.

o Barefoot et al. (1983). Examined relationship between Cook-Medley Hostility Scale (Ho) & 25-yr CHD incidence among 255 men who had taken MMPI in med school. Those with Ho scores above the median had greater CHD incidence & total mortality as those with scores below the median.

o Barefoot JC et al. (1989). Classified Ho on an a priori basis into 6 categories: Cynicism, Hostile Attributions, Hostile Affect, Aggressive Responding, Social Avoidance, and Other. Among 118 lawyers, Ho scores (Cynicism, Hostile Affect, Aggressive Responding) were associated with reduced survival over 28-yr period.

o Siegler, Peterson, Barefoot et al. (1992). Among 4710 men & women, Ho scores measured in college predicted many coronary RF over 20-yr period, including greater caffeine consumption, greater BMI, current smoking, larger lipid ratio, more hours of exercise at follow-up.

o Everson SA et al. (1997). Kuipio Ischemic Heart Disease Study. High levels of hostility, as measured by the Cynical Distrust Scale, an 8-item subscale of the Cook-Medley Hostility Inventory, were associated with increased risk of ACM & incident MI; effects were primarily mediated thru behavioral factors.

o Iribarren et al. (2000). Coronary Artery Risk Development in Young Adults (CARDIA) Study. Subsample of 374 respondents aged 18-30 followed for 5-10 yrs. High Ho scores predispose young adults to coronary artery calcification. Results using cynical distrust subscale were somewhat weaker than those using the global hostility score.

o Kubzansky et al. (1999). Normative Aging Study. Higher hostility (Cook-Medley subscales: Hostile Affect, Hostile Attribution, Aggressive Responding) was associated with higher allostatic load scores.

o Niaura R et al. (2000). Normative Aging Study. Hostility (Cook-Medley) was associated with the metabolic syndrome.

(2) Buss-Durkee Hostility Inventory (Buss & Durkee, 1957). (See Ostfeld & Eaker, p.233).

75 true/false items; global hostility, plus 7 hostility subclasses.

o Medline search, using "Durkee" (kw) & "Longitudinal Studies" (MeSH), yielded no references for the years 1975-2000.

(3) Spielberger State, Trait, Expression of Anger Inventory (Spielberger et al., 1985).

20 items concerning frequency of feeling quick tempered & flying off the handle. Expression of Anger subscales include: Anger-out: assesses degree to which respondent will do something hostile (e.g. slam doors, argue, say something nasty, lose one's temper) to a person who provokes their anger. Anger-in: assesses likelihood respondent will conceal his/her anger toward others. Anger-control: assesses extent of perceived ability to control one's expression of anger.

o Everson SA et al. (1998). 537 initially normotensive men from eastern Finland followed for 4 yrs. Anger expression, assessed by Spielberger's Anger-out & Anger-in scales, was associated with incident hypertension.

o Everson SA et al. (1999). Kuipio Ischemic Heart Disease Study. Pop'n based longitudinal study of 2074 men (mean age=53 yr) followed for 8 yrs. Men who reported high level of expressed anger, as assessed by Spielberger Anger Expression Scales, were at increased risk of incident stroke. Additional analyses showed that this assoc was evident only in men with a hx of ischemic heart disease, among whom high levels of outwardly expressed anger (high anger-out) predicted a more than 6-fold increased risk of stroke. Suppressed anger (anger-in) and controlled anger (anger-control) were not consistently related to stroke risk. Note: Confirmatory factor analysis in the Kuipio cohort revealed that 7 of the original 24 items on the 3 anger scales had inadequate goodness of fit; thus, the anger scales were recalculated, and several items from the original scales were dropped (see Miller TQ et al., 1995, p. 1154, for the revised anger scales used in this study).

o Williams JF, Paton CC et al. (2000). Atherosclerosis Risk in Communities (ARIC) study. 12,986 participants. Among persons with normal BP at baseline, those more prone to anger (Spielberger Trait Anger Scale) were almost 3 times more likely to have an MI or sudden cardiac death than those who were least anger prone. There was a dose-response relation between anger level & CHD risk. (Appendix gives 10-item Spielberger Trait Anger Scale.)

o Mittleman MA et al. (1995). Determinants of Myocardial Infarction Onset Study. Case crossover study design. 1623 patients interviewed 4 days post MI. RR of MI onset was associated with scores on Spielberger State Anger subscale. (Similar results found for Onset Anger Scale).

(4) Framingham Anger Scales (Haines et al., 1978). (See also Ostfeld & Eaker 1985, p. 249)

4 scales: Anger symptoms (5 items); e.g., "When angry or annoyed, do you get tense or worried?"
Anger in (3 items); e.g., "When angry or annoyed, do you keep it to yourself?"
Anger out (2 items); e.g., "When angry or annoyed, do you take it out on others?"
Anger discuss (2 items); e.g., "When angry or annoyed, do you get it off your chest?"

o Haynes et al. (1980). Framingham Study. Anger in was risk factor for CHD among women but not men.

o Gallagher et al. (1999). Caerphilly Study. 2890 men aged 49-65. Framingham Anger Scales, plus a new "suppressed anger" scale. Anger out and suppressed anger were predictive of incident IHD.

(5) Onset Anger Scale (Mittleman MA et al., 1995)

Respondents are shown a chart consisting of 7 levels of anger & asked to estimate usual frequency of exposure to each level. Time periods assessed in original study include "during previous yr" & "during each of 26 hrs before onset of MI."

o Mittleman MA et al. (1995). Determinants of Myocardial Infarction Onset Study. Case crossover study design. 1623 patients interviewed 4 days post MI. Risk of MI onset was elevated during 2-hr period after outburst of anger, as measured by Onset Anger Scale. (MI onset was also associated with Spielberger State Anger subscale.)

o Moller et al. (1999). Stockholm Heart Epidemiology Program (SHEEP). Case-crossover analysis. Used the Onset Anger Scale to study whether anger episode trigger MI.

Epidemiologic evidence:

o There is compelling evidence linking hostility & anger to various aspects of CHD. A number of negative studies have been reported (e.g. Maruta et al., 1993, using Ho), but the number of positive studies outnumber those with no effects (for reviews, see Matthews (1988), Adler & Matthews (1994), Kubzansky & Kawachi, unpublished manuscript).

TYPE A BEHAVIOR (TAB)

Overt manifestations of TAB include hostility, hyper-aggressiveness, & sense of time urgency.

Interview measures:

(1) Structured Interview (SI; Rosenman & Friedman). Goal of 10-minute interview is to provide a challenging environment through the interviewer's style to elicit type A behaviors. Developed for Western Collaborative Group Study.

o Matthews et al. (1977). Western Collaborative Group Study. Variables that were most predictive of CHD included ratings of potential for hostility, anger directed outward, frequent experience of anger, & irritation when waiting in lines.

o Houston et al. (1987). Western Collaborative Group Study. Components of Type A, derived from the SI, were analyzed in relation to risk of CHD. As expected, a pattern of characteristics in which hostility was salient was found to be predictive of CHD. Another pattern of characteristics reflecting pressured, controlling, socially dominant behavior was also predictive of CHD.

o Hecker MH et al. (1988). Western Collaborative Group Study. 250 cases, 500 matched controls were studied to assess 8.5-yr prospective relationship of TABP, as assessed by SI, to incident CHD. Hostility was only TABP component that was a significant predictor when all 12 TABP components were included in model.

o Dembroski TM et al (1989). MRFIT. 192 cases, 384 matched controls. Tested hypothesis that hostility is associated with increased risk of coronary death & nonfatal MI. "Potential for Hostility", but not global TABP, was associated with CHD incidence.

Self-report inventories:

(2) Jenkins Activity Survey (JAS; Jenkins, Zyzanski, & Rosenman, 1971) (see Ostfeld & Eaker 1985, p.217)

o Glynn RJ et al. (1988). Normative Aging Study. JAS Type A score unrelated to alcohol use.

(3) Framingham Type A Scale (Haynes et al., 1978) (see Ostfeld & Eaker 1985, p. 213)

o Reed T et al. (1991). NHLBI twin study. Measures of hostility (Framingham Type A) & cognitive function were not related to placental type.

(4) Bortner Rating Scale (Bortner, 1969) (see Ostfeld & Eaker 1985, p. 211).

Epidemiologic evidence:

o TAB has been more consistently linked with CHD in cross-sectional & retrospective studies than in cohort studies (reviewed by Allen & Scheidt, 1996).

o Negative emotions (hostility, anger) may underlie link between TAB & CHD (Booth-Kewley & Friedman's 1987 meta-analysis of TAB studies).

o Recent research, most of which has focused on hostility rather than anger, corroborates this view (reviewed by Helmers KF et al. 1994; Scheier MF& Bridges MW 1995; Miller TQ et al., 1996; Kubzansky & Kawachi, unpublished manuscript).