To evaluate whether a high level of lipoprotein(a) (Lp(a)) is a risk for the development of coronary heart disease (CHD), 94 Japanese patients and 64 age-matched Japanese controls, diagnosed after coronary angiography (CAG), were analyzed with special reference to the relations between the degree of atherosclerosis, Lp(a) levels and the apolipoprotein(a) (apo(a)) genotypes. the degree of atherosclerosis was evaluated based on CAG findings in the following three ways: the number of diseased vessels, the Gensini score, and the presence or absence of vascular ulcers and/or irregular outlines of coronary stenotic lesions. Apo(a) protein sizes and the pentanucleotide (TTTTA) repeat polymorphism in the 5' control region of the apo(a) gene were analyzed. Multivariate predictors for the number of diseased vessels were, in decreased order of significance, plasma Lp(a) levels, history of smoking, hypertension, diabetes mellitus, and body mass index (BMI). Independent factors associated with the Gensini score were Lp(a) levels, BMI, hypertension, and diabetes mellitus. A negative association of Lp(a) levels with apo(a) protein sizes, and higher Lp(a) levels in those homozygous for an allele with 8 8 (TTTTA)-repeats, was found in both the controls and patients. In decreasing order of significance, apo(a) protein sizes, the degree of atherosclerosis, the genotype of the pentanucleotide repeat, and gender were independent predictors of Lp(a) levels in stepwise regression models. Apo(a) protein sizes were a significant predictor, and the genotype homozygous for the 8 (TTTTA)-repeats was a possible predictor, for the degree of atherosclerosis in CHD. These findings support the notion that a high Lp(a) level is a risk for the development of atherosclerosis in CHD.