In a study of 307 white patients who underwent coronary angiography, the relationship of coronary artery disease (CAD) to plasma levels of lipoprotein Lp(a) and other lipid-lipoprotein variables was examined. Lp(a) resembles low-density lipoprotein (LDL) in several ways, but can be distinguished and quantified by electroimmunoassay. CAD was rated as present or absent and was also represented by a quantitative lesion score derived from estimates of stenosis in four major coronary vessels. Coronary lesion scores significantly correlated with Lp(a), total cholesterol, triglycerides, LDL cholesterol, and high-density lipoprotein (HDL) cholesterol levels by univariate statistical analysis. By multivariate analysis levels of Lp(a) were associated significantly and independently with the presence of CAD (p < .02), and tended to correlate with lesion scores (p .06). Among subgroups Lp(a) level was associated with CAD in women of all ages and in men 55 years old or younger. An apparent threshold for coronary risk occurred at Lp(a) lipoprotein mass concentrations of 30 to 40 mg/dl, corresponding to Lp(a) cholesterol concentrations of approximately 10 to 13 mg/dl. Plasma Lp(a) in white patients appears to be a major coronary risk factor with an importance approaching that of the level of LDL or HDL cholesterol.