Plasma anticardiolipin antibody (ACA) was measured in 83 patients having coronary artery bypass graft surgery and results were correlated with the incidence of early (1-2 weeks) and late (12 months) graft occlusion, as judged by angiography. There was an association between preoperative ACA level and the incidence of late graft occlusion in relation to both number of patients with an occlusion and number of distal anastomoses occluded. 8 of 15 patients (53.3%) whose maximum ACA level exceeded 4 SD of the mean of controls had a late graft occlusion. When the ACA titre was 2-4, 0-2, or less than 0 SD above the mean of the controls, the occlusion rates were 3 of 13 (23.1%), 3 of 33 (9.1%), and 1 of 15 (6.7%), respectively (p less than 0.03). The incidence of a postoperative rise in ACA level was higher in patients who had had a myocardial infarction in the past than in those with a history of angina only (chi 2 = 4.08, p less than 0.05). This observation supports the notion that one mechanism of ACA production is an immune response to myocardial injury. These results raise the further possibility that the ACA, or related antiphospholipid antibodies, may play a part in progressive coronary vessel disease.