The effect of dipyridamole was investigated in 360 patients undergoing coronary bypass surgery. They were randomly allocated to receive dipyridamole (100 mg orally q.i.d. for 2 days preoperatively, 5 mg/kg body weight/24 h i.v. peroperatively and 100 mg orally q.i.d. for 1 year postoperatively) or placebo. Withdrawn from the study were 48 patients on dipyridamole and 57 on placebo. Cardiovascular and/or cerebrovascular events or need for anticoagulant treatment were the reasons for withdrawal in 22 (13%) of the dipyridamole, and 34 (18%) of the placebo group. Logistic regression analysis of risk factors influencing graft patency showed significant relation to peroperatively measured coronary blood flow. A positive trend of treatment was observed (p = 0.08). Vein graft blood flow measured during bypass surgery (245 patients) was significantly greater in the dipyridamole group (p less than 0.01). The occlusion rate was lower in vessels with peroperative blood flow greater than 30 ml/min (vein-marginal p less than 0.01, vein-dexter p less than 0.05, vein-diagonal 0.05 less than p less than 0.1). Dipyridamole increases coronary blood flow and graft patency following coronary bypass surgery.