To develop a more efficient antithrombotic way after coronary artery bypass grafting(CABG),the anticoagulant effects were compared of human tissue factor pathway inhibitor(TFPI) gene transfection and aspirin oral administration(traditional method) on vein grafts.An eukaryotic expression plasmid pCMV-(Kozak) TFPI was prepared.Animal model of carotid artery bypass graft-ing was constructed.In operation,endothelial cells of vein grafts in TFPI group and empty plasmid control group were transfected with pCMV-(Kozak) TFPI and empty plasmid pCMV respectively,while no transfection was conducted in aspirin control group.After operation,aspirin(2 mg·kg-1·d-1) was administered(i.g.) in aspirin control group.Three days later,grafts(n=10) were harvested for RT-PCR,Western blotting and immunohistochemical analyses of exogenous gene expression and for pathological,scanning electron microscopic observation of thrombus.Thirty days later,the patency rates of remnant grafts(n=10) were recorded by vessel Doppler ultrasonography.Human TFPI gene products were detected in gene transferred vein grafts.Three days later,thrombi were found in 7 animals of aspirin control group and in 8 animals of empty plasmid control group,but in only 1 of TFPI group(P<0.01).Thirty days later,5 grafts were occluded in empty plasmid control group,but none of grafts was occluded in the other groups(P<0.05).The endothelial surfaces of grafts in both of the control groups were covered with aggregated erythrocytes and platelets,and it were not seen in TFPI group.It was suggested that the anticoagulant effects on vein grafts of human TFPI gene trans-fection are better than those of aspirin. To develop a more efficient antithrombotic way after coronary artery bypass grafting (CABG), the anticoagulant effects were than of human tissue factor pathway inhibitor (TFPI) gene transfection and aspirin oral administration (traditional method) on vein grafts. Ann eukaryotic expression plasmid pCMV- (Kozak) TFPI was prepared. Animal model of carotid artery bypass graft-ing was constructed.In operation, endothelial cells of vein grafts in TFPI group and empty plasmid control group were transfected with pCMV- (Kozak) TFPI and empty plasmid pCMV respectively, while no transfection was conducted in aspirin control group. After operation, aspirin (2 mg · kg -1 · d -1) was administered (ig) in aspirin control group.Three days later, grafts (n = 10) were harvested for RT -PCR, Western blotting and immunohistochemical analyzes of exogenous gene expression and for pathological, scanning electron microscopic observation of thrombus. Thirty days later, the patency rates of remnant grafts (n = 10) were recorded by vess el Doppler ultrasonography. Human TFPI gene products were detected in gene transferred vein grafts. Three days later, thrombi were found in 7 animals of aspirin control group and in 8 animals of the empty plasmid control group, but only 1 of TFPI group (P < 0.01). Thirty days later, 5 grafts were occluded in the empty plasmid control group, but none of the grafts were occluded in the other groups (P <0.05). The endothelial surfaces of grafts in both of the control groups were covered with aggregated erythrocytes and Platelets, and it were not seen in TFPI group. It was suggested that the anticoagulant effects on vein grafts of human TFPI gene trans-fection are better than those of aspirin.