To determine the association of serum apolipoprotein (apo) A-I and apo B concentrations, and paraoxonase (PON) high-density lipoprotein (HDL) associated enzyme activity with angiographically determined coronary artery disease (CAD) in Iranian diabetic and nondiabetic CAD patients and nondiabetic control subjects, 251 subjects aged 30-70 years, who underwent their first coronary angiography were matched and randomly assigned into three groups: CAD+DM+, CAD+DM-, and CAD-DM- (control). Stenosis of >= 50% in 1 or more coronary arteries was classified as CAD+. CAD- was defined as a maximum stenosis of 10% in any coronary artery. Fasting serum concentrations of cholesterol (TC), triglycerides (TGs), LDL-C, HDL-C, apo A-I, apo B, and PON activity were determined. Apolipoprotein concentrations were measured in a fasting serum sample by immunoturbidometric assay and paraoxonase/arylesterase activities by spectrophotometric assay of p-nitrophenol/phenol production following addition of paraoxon/phenylacetate. Information concerning nonlipid risk factors were collected by questionnaires. No significant difference was observed in HDL-C, LDL-C, apo A-I, and PON/arylesterase activity between the study groups. The values of TC (213±38 vs 196±45, p<0.05), TGs (209±187 vs 151±113, p<0.005), apo B (99±22 vs 96±24, p<0.0001), TC/HDL-C (4.8±1.5 vs 4.0±1.3, p<0.001), and LDL-C/HDL-C (2.9±1.1 vs 2.4±1.1, p<0.05) were higher and apo A-I / apo B (1.7±0.4 vs 2.0±0.6, p<0.01) was lower in CAD+DM+ patients than in control subjects. In CAD+DM- group, only the level of apo B (96±24 vs 85±18, p<0.01), and the ratio of apo A-I/apo B (1.8±0.4 vs 2.0±0.6, p<0.01), were significantly higher than those of control group. On multiple logistic regression analysis, the best markers for discrimination between CAD+ groups and CAD- control subjects were the ratio of apo A-I/apo B in diabetic and apo B in nondiabetic patients. The results suggest that in Iranian diabetic and nondiabetic patients with CAD the concentration of apolipoproteins are better markers than traditional lipid parameters in discriminating between CAD+ and CAD- subjects. Lack of significant difference in PON activity between CAD+ patients and CAD- controls supports the concept of interethnic variability in PON polymorphism and unimodal distribution of its activity in non-Europid populations observed in other studies.