An in vitro method for the risk stratification of patients with stable arteriosclerosis, especially stable coronary artery disease, is disclosed wherein the concentration of procalcitonin is determined in the circulation of such patients using a highly sensitive PCT assay, and wherein within the range of PCT concentrations in the typical normal range of healthy individuals cutoff values are defined which distinguish groups of individual patients with stable arteriosclerosis in accordance with personal cardiac risk, and patients are allotted to one of said risk groups on the basis of their individual PCT concentrations.