Hyperhomocysteinemia is now regarded as an established risk factor for coronary artery disease and is present frequently in the general population. However, the diagnostic value of this risk factor relative to others has only occasionally been investigated. We compared the diagnostic value of classic risk factors and of homocysteine in a retrospective case-control study in 191 cases with angiographically established coronary artery disease and 231 healthy controls. Life style habits were assessed by a detailed questionnaire. Laboratory parameters including lipoproteins and blood lipids, homocysteine, folate, and vitamin B12 were measured and their diagnostic value compared with each other by use of receiver-operator characteristic analysis. Comparison of the receiver-operator characteristic curves revealed that homocysteine significantly discriminated between cases and control subjects. High-density-lipoprotein cholesterol, triglycerides and non-esterified fatty acids also had an area under the curve significantly different from 0.5 (the area under the curve representing no discrimination). Homocysteine was weakly related to folate, vitamin B12, age and serum creatinine concentration. We conclude that hyperhomocysteinemia is at least as important as conventional risk factors for coronary artery disease and that receiver operator characteristic analysis of homocysteine is suitable to determine patients at the highest risk for coronary artery disease. Clinical trials testing the effect of homocysteine lowering by vitamin supplementation in the prevention of coronary artery disease are needed.