The Lp(a) lipoprotein is structurally related to low-density lipoprotein but is found in lower plasma concentration. It has been associated with coronary disease in several white populations. To test the generalizability of this association, we measured serum Lp(a) by quantitative immunoelectrophoresis in 303 Hawaiian men of Japanese ancestry with a prior myocardial infarction (MI) and in 408 population-based controls. Mean values were 17.1 and 13.7 mg/dL (0.171 and 0.137 g/L), respectively. Increased risk for MI was shown mainly for men in the upper quartile of the Lp(a) lipoprotein distribution (greater than or equal to 20.1 mg/dL [greater than or equal to 0.201 g/L]). Odds ratios at younger than 60, 60 to 69, and 70 years of age or older were 2.5, 1.6, and 1.2 times those for men in the lower three quartiles, respectively. In a multiple logistic model the association with MI remained significant and was not explained by differences in total cholesterol levels, high-density lipoprotein or low-density lipoprotein cholesterol levels, subscapular skin fold, systolic blood pressure, history of smoking, alcohol consumption, or age. We conclude that Lp(a) is an important attribute that should often be considered when coronary heart disease risk is assessed.