The objective of the present study was to examine the possible associations between low molecular weight (LMW) apolipoprotein(a) (apo(a)) isoforms (F,B,S1,S2) and coronary heart disease (CHD). We conducted a nested case-control (prospective) study of five cohorts of white men: The 1936 cohort (baseline 1976, n = 548) and four cohorts from MONICA I born in 1923 (n = 463), 1933 (n = 491), 1943 (n = 504) and 1953 (n = 448) studied at baseline in 1983. At follow up in 1991, 52 subjects had developed a first myocardial infarction and 22 had been hospitalized with angina pectoris. Plasma samples obtained at baseline were stored frozen until 1993-94, when case samples (n = 74) were analyzed together with samples from matched (disease free) controls (n = 190). In a statistical model (conditional logistic regression) including all age groups, cholesterol (or apo B) level (P < 0.01), systolic blood pressure (P = 0.05) and smoking (P = 0.02) predicted CHD. In the statistical model Lp(a) interacted significantly with age (OR = 5.7; 95% CI: 1.4-23.6; P = 0.016), and high Lp(a) (over 45 mg/dl) was associated with significantly increased risk in subjects under 60 years (OR = 3.82; 95% CI: 1.47-9.96), but not in older men (OR = 0.67; 95% CI: 0.235-1.89). Therefore, we studied the impact of Lp(a)/apo(a) and other variables in subjects who had been under 60 years when they became cases. Among the younger subjects the presence of LMW apo(a) isoforms significantly predicted the development of CHD (OR = 3.83; 95% CI: 1.18-12.4). The increased risk pertained to high Lp(a) (above versus below 45 mg/dl: OR = 3.68; 95% CI: 1.03-13.10), and to Lp(a) concentrations when entered into the model as a continuous variable (P = 0.04). Cholesterol or apo B (P < 0.01), smoking (P = 0.02), systolic blood pressure (P = 0.05) and low alcohol consumption (under nine drinks/week) (P = 0.04) were also significant predictors of CHD. We conclude that LMW apo(a) isoforms are significantly associated with increased risk of CHD in men under 60 years.