The efficacy of coronary bypass grafting obviously being linked to graft patency, it is compulsory to look for any innovation that could improve the patency rate. Ticlopidine, an antiplatelet drug, was tested against placebo in a double-blind trial: 173 patients (475 grafts) subjected to venous coronary artery bypass grafting were randomly treated with ticlopidine (250 mg twice daily) or placebo from the second postoperative day for 12 months. Graft patency was assessed by digital angiography on days 10 (99.4% of the patients), 180 (98.2%), and 360 (91.7%). The effect of treatment on platelet aggregation and bleeding time were measured concomitantly; a clear-cut effect was demonstrated at each interval. Intention-to-treat graft-by-graft analysis shows that ticlopidine significantly reduced the graft occlusion rate on day 10 (7.1% versus 13.4%, p less than 0.05), day 180 (15.0% versus 24.0%, p less than 0.02), and day 360 (15.9% versus 26.1%, p less than 0.01). Sequential grafts to the left anterior descending coronary artery, with side-to-side anastomosis to diagonal branch(es), are less frequently occluded than individual grafts. On the contrary, grafts to endarterectomized vessels occlude more frequently. Individual patient-by-patient analysis shows that patency of all grafts at each study time, is more frequent in the ticlopidine group. The difference is significant when one considers patients without sequential or endarterectomized grafts. The difference is also present at each study time: day 10 (84.4% versus 66.7%, p less than 0.05), day 180 (74.4% versus 52.3%, p less than 0.05) and day 360 (75.0% versus 52.5%, p less than 0.05). Results are even more impressive if one excludes from analysis the four patients in the ticlopidine group in whom administration of the drug was delayed. This supports previous suggestions that early therapy is necessary. These results show that graft occlusion occurs mainly in the first 6 postoperative months. The incidence of occlusion is significantly reduced by ticlopidine therapy.