Low density lipoprotein (LDL) from 36 young post-infarction patients was separated by isopycnic density gradient ultracentrifugation to determine the relationships of plasma levels and chemical composition of different LDL subfractions to the global severity and rate of progression of coronary atherosclerosis assessed by angiography. There were marked elevations of the cholesterol and triglyceride concentrations in the very low density lipoprotein (VLDL) fraction, whereas the high density lipoprotein (HDL) cholesterol level was reduced in the patients compared with 70 healthy population-based controls. Plasma total LDL cholesterol and triglyceride concentrations were similar. The distribution of apolipoprotein B along the LDL density range, viz. the LDL particle distribution, was displaced towards the dense LDL region among the patients compared with 14 healthy normolipidaemic controls. A preponderance of dense LDL particles was associated with elevated plasma VLDL triglyceride concentration. The patients had significantly higher plasma concentrations of lipid and protein in dense LDL (d greater than 1.040 kg/l), while no group differences were found in the light LDL (d less than 1.040 kg/l). However, there were no percentage compositional differences in the light or dense LDL between patients and controls. Among all constituents of lipoprotein fractions and subfractions determined, only the plasma level of triglycerides in both light and dense LDL correlated significantly with the angiographic estimates of global severity and rate of progression of coronary atherosclerosis, respectively. On a percentage composition basis, both light and dense LDL tended to be richer in triglycerides in the subjects with a more severe coronary artery disease. Neither VLDL or HDL, nor LDL cholesterol were associated with the angiographic scores, the plasma LDL triglyceride concentration or the triglyceride enrichment of LDL. Although there is ample experimental evidence that triglyceride-enriched LDL predisposes to atherosclerosis, the LDL associations with coronary lesion severity and progression observed in the present study might not reflect a causal mechanism, but merely mirror the atherogenicity of disturbances affecting the metabolism of triglyceride-rich lipoproteins. Prospective studies of larger groups of unselected patients are needed to corroborate these findings.