Blood plasma was drawn from 12 healthy subjects and 12 patients with coronary heart disease (CHD). The lipoproteins were fractionated by serial ultracentrifugation. The methyl and methylene regions of proton nuclear magnetic resonance (1 H NMR) spectra of the lipoproteins very low density lipoprotein (VLDL), low density lipoprotein (LDL) and high density lipoprotein (HDL) were analysed by principal component analysis (PCA). Grouping patterns in the score plots and the profiles of the principal components revealed several characteristics of the spectra. LDL subparticle size among the CHD group was consistently skewed against smaller, denser subparticles. This feature was independent of the concentration of LDL cholesterol. Analysis of the LDL spectra by soft independent modelling of class analogy (SIMCA) showed that none of the samples from the CHD group could be assigned as healthy subjects (p < 0.05). We also found that the samples from the healthy subjects were associated with a higher concentration of HDL cholesterol and larger VLDL subparticles. The approach presented, in which PCA is used in combination with NMR spectroscopy, might be implemented in clinical studies to give information about lipoprotein subparticle distribution and lipid content.