BACKGROUND Serum gamma-glutamyltransferase activity (GGT) has been documented as an independent cardiovascular risk factor. However, to-date its value has not been compared with C-reactive protein (CRP) and other indexes in a multimarker prognostic strategy in patients with coronary artery disease. METHODS We prospectively evaluated 474 subjects with angiographically documented CAD. GGT and traditional humoral and clinical parameters were measured at hospital admission. A multivariate model was used to predict all-cause and cardiac mortality. RESULTS GGT showed an independent prognostic value after adjustment for possible confounders, including alcohol consumption, and beyond established risk factors, such as extent of coronary atherosclerotic disease, left ventricular ejection fraction, age, serum glucose, cholesterol subfractions, and C-reactive protein (CRP). At a 3-year follow-up, cardiac mortality was 9% in patients with serum GGT activity >25 U/L vs. 3.5% in those with serum GGT<25 U/L (p=0.028). The association of three independent biomarkers (higher GGT, CRP, fasting glucose) identified a subgroup of 45 patients with the highest risk of cardiac death at 3 years (26.6%, vs. no event or 2.7% in the subsets of 87 and 198 patients with, respectively, no/one risk factor above cut-off value, p<0.0001). CONCLUSIONS GGT is confirmed as independent risk factor in patients with established coronary artery disease. GGT, CRP, fasting glucose show an additive prognostic value, whereas low values of these biomarkers identify a subset of patients with the lowest risk of cardiac death.