The association between number of cups of coffee consumed per day and self-reported heart infarction, angina pectoris, or symptoms of angina pectoris was studied in 11635 men and 11785 women, aged 40-54 years, who attended two cardiovascular screenings. Questionnaire information on heart infarction has previously been found to be reliable. Altogether 240, 201 and 241 men reported heart infarction, angina pectoris (without heart infarction) and symptoms of angina pectoris, respectively, at the last, but not at the first screening. The corresponding figures for women were 35, 102 and 395. Univariate analysis showed a positive association between number of cups of coffee and self-reported heart infarction. The same applies to self-reported angina pectoris. After adjustment for major coronary risk factors, the associations became weaker. The relative risks of self-reported heart infarction between users of nine or more cups and users of less than one cup were 1.7 (95% CI: 1.0, 3.0) and 1.5 (0.3, 7.5) in men and women, respectively. The relative risks were smaller than those found in a previous study relating coffee consumption to coronary heart disease mortality. For self-reported angina pectoris the corresponding relative risks were 1.4 (0.8, 2.7) for men and 2.7 (1.1, 6.7) for women. The study is suggestive of a modest relationship between coffee consumption and self-reported occurrence of both heart infarction and angina pectoris.