The concentration of plasma LDL subfractions is described in four groups of normocholesterolaemic (total plasma cholesterol < 6.5 mmol/l) male subjects consisting of men with and without coronary artery disease (CAD+/-), as determined by angiography, post-myocardial infarct survivors (PMI) and normal, healthy controls. The CAD(+) and PMI groups were distinguished from the CAD(-) and controls by raised concentrations of plasma triglyceride, very low density lipoprotein (VLDL) cholesterol, small, dense LDL (LDL-III density (d) 1.044-1.060 g/ml) and lower concentrations of high density lipoprotein (HDL) cholesterol and large, buoyant LDL (LDL-I d 1.025-1.034 g/ml). In all groups, a subfraction of intermediate density, LDL-II (d 1.034-1.044 g/ml), was the predominant LDL species but was not related to coronary heart disease risk. Plasma triglyceride showed a positive association with LDL-II (r = 0.51, P < 0.001) below a triglyceride level of 1.5 mmol/l. Above this threshold of 1.5 mmol/l, LDL-II and LDL-I showed significant negative associations with triglyceride (LDL-II r = -0.5, P < 0.001; LDL-I r = -0.45, P < 0.001). Small, dense LDL-III showed a weak positive association with triglyceride that became highly significant above the 1.5 mmol/l threshold (r = 0.54, P < 0.001). While age was positively related to LDL-II within the control subjects (r = 0.3, P < 0.05), there was no difference in the percentage abundance or concentration of LDL-III within control and CAD(-) subjects above and below the age of 40 years. Smoking was associated with a relative deficiency of the LDL-I subfraction (LDL-I to LDL-III ratio in smokers = 0.77, in ex-smokers = 0.95, in non-smokers = 1.89; P < 0.01), as was beta-blocker medication (% LDL-I, users vs. non-users, P < 0.05). Both of these effects could be explained by their primary influence on plasma triglyceride. Analysis of the frequency distributions for the three LDL subfractions revealed the concentration of small, dense LDL-III to be bimodal around a concentration of 100 mg (lipoprotein mass)/100 mi plasma. The calculation of odds ratios based on this figure indicated relative risk estimates of 4.5 (chi(2): P < 0.01) for the presence of coronary artery disease and 6.9 (chi(2): P < 0.001) for myocardial infarction. These likelihood scores are greater than those previously described for LDL subclass pattern B as determined by gradient gel electrophoresis and, surprisingly, were maintained after correction for triglyceride, age, body mass index (BMI), smoking and drug status. The present study offers a quantitative description of LDL subfractions in clinically defined coronary artery disease and after myocardial infarction. It offers a plausible explanation for the underlying relationships between LDL subfractions and triglyceride that are operative in healthy and diseased individuals, and demonstrates the value of small, dense LDL as a predictive index of coronary risk status.
