ObjectiveThe idea that anger may predict ischemic heart disease (IHD) is more than 30 years old. Some, but not all, prospective studies have supported this suggestion. Attention has focused on hostility as the critical component of anger for IHD risk. This idea is explored using prospective data from the Caerphilly study.MethodsA sample of 2890 men aged 49 to 65 years living in and around Caerphilly, South Wales, was identified. Anger was assessed using the Framingham scales comprising "anger symptoms," "anger in," "anger out," and "anger discuss." A new "suppressed anger" scale was also constructed. Cardiovascular risk factors assessed included baseline blood pressure, total and high-density lipoprotein cholesterol, fibrinogen, white cell count, psychiatric caseness as assessed by the General Health Questionnaire, social support, smoking habit, alcohol consumption, leisure exercise, body mass index, and calorie intake. Prediction of IHD, measured as the occurrence of a major event over a follow-up period of 9 years, was assessed using multiple logistic regression analysis.ResultsA low anger out score predicted increased risk of a major IHD event (relative odds (RO) = 1.70; 95% confidence interval = 1.26-2.29 for all RO). This association was unchanged on controlling for physiological risk factors (RO = 1.74), psychosocial risk factors (RO = 1.72), and behavioral risk factors (RO = 1.69). Suppressed anger showed associations with incident IHD similar to those of anger out but identified the population at risk more closely.ConclusionsAnger out and suppressed anger were predictive of incident IHD. Neither of these constructs are overtly similar to hostility. These findings suggest there may be mechanisms other than hostility by which anger predicts IHD risk and that a conceptually varied approach to anger is currently appropriate.