BACKGROUND Superior patency rates for internal mammary artery (IMA) grafts compared with vein coronary bypass grafts have been demonstrated by retrospective studies. This difference may have been affected by selection bias of patients and coronary arteries for IMA grafting. METHODS AND RESULTS To estimate the difference between IMA and vein grafts, we analyzed graft patency data of 912 patients who entered a randomized clinical drug trial. In this trial, 494 patients received both IMA and vein grafts (group 1) and 418 only vein grafts (group 2). Occlusion rates of IMA grafts and IMA plus vein grafts in group 1 were compared with those of vein grafts in group 2. Multivariate analysis was used to compare occlusion rates of IMA and vein grafts while other variables related to graft patency were controlled for. In addition, 1-year clinical outcome was assessed by the incidence of myocardial infarction, thrombosis, major bleeding, and death. Occlusion rates of distal anastomoses in group 1 versus group 2 were 5.4% (IMA grafts) versus 12.7% (vein grafts) (P < .0001) and 10.4% (IMA plus vein grafts) versus 12.7% (vein grafts) (P = .14). There was no difference in adjusted risk of occlusion between IMA grafts and vein grafts (P = .089). Type and location of distal anastomosis and lumen diameter of the grafted coronary artery were shown to be predictors of occlusion. Clinical events occurred in 17.8% (group 1) and 16.0% (group 2) of patients (P = .53). CONCLUSIONS The observed difference in 1-year occlusion rates between IMA and vein grafts can be explained by a maldistribution of graft characteristics by selection of coronary arteries for IMA grafting rather than being ascribed to graft material. One-year clinical outcome is not improved by IMA grafting.