Objectives - Increased lipoprotein (a) [Lp(a)] concentration was reported to be an independent risk factor for coronary heart disease (CHD). Recent epidemiological studies affirmed the value of C-reactive protein (CRP) as the strongest, univariate predictor of the cardiovascular events. We decided to establish cut-off levels providing maximum diagnostic efficiency for CHD. Methods - In this study we measured CRP and Lp(a) concentrations in patients with angiographically demonstrated CHD (group A, n: 120), patients without any angiographically demonstrable lesion (group B, n: 62) and a group of healthy subjects (group C, n: 41). Data were evaluated correcting for lipid and lipoprotein concentrations, diabetes mellitus, hypertension, smoking, age, and body mass index in men and women. ROC curve based cut-off values (comparing group A versus groups B and C) and associated diagnostic performances of the assays were evaluated. Results - Significant increases were noted in serum CRP concentrations in men and women, in groups A vs. B, A vs. C, B vs. C. Lp(a) concentrations were not different among groups in men but were higher in group A vs. B and C in women. Optimal cut-off levels for CRP in women and men were found as 2.1 and 3.0 mg/l with the diagnostic values of 0.792 and 0.770, respectively. For Lp(a) optimal cut-off levels were found as 22.6 and 9.8 mg/dl with the diagnostic values of 0.612 and 0.596 in women and men, respectively. Conclusion - The CRP level is quite efficient for separation of patients from controls. Therefore keeping in mind the lack of specificity, the CRP level may be a useful tool in the diagnosis of coronary heart disease. However, the Lp(a) level is not efficient enough to support the use of Lp(a) measurement for management of coronary heart disease.