The potential predictors of ischemic heart disease mortality were studied for 931 male foundry workers in Finland who participated in a health examination in 1973. These workers were followed up to 1993 through registers and by using a questionnaire. In 1973, the systolic and diastolic blood pressures of workers exposed to carbon monoxide (CO) were slightly higher than those of unexposed workers. The prevalence of angina pectoris showed a clear dose-response relation to CO exposure. Electrocardiogram (ECG) findings indicating past myocardial infarction or suggesting coronary artery disease as a function of smoking and/or CO exposure were not evident. In the 1987 follow-up, the rate ratio for ischemic heart disease mortality was estimated as 4.4 for CO-exposed smokers compared with unexposed nonsmokers. Ischemic heart disease mortality in 1973-1993 was analyzed by using the Cox proportional hazards model. The statistically significant predictors were age, pathologic ECG findings in 1973, regular CO exposure, and abundant alcohol drinking. Of the ECG findings, changes in Q or QS and ST-J or ST waves and in ventricular extrasystoles were statistically significant. The risk of mortality from ischemic heart disease was increased by working in iron foundries, by hypertension, and by smoking.