Animal studies have demonstrated that synchronised coronary sinus retroperfusion with arterial blood can provide effective perfusion of ischaemic myocardium. Preliminary clinical studies have shown that the technique can also be used with safety in human beings, and in the present study its effectiveness was assessed in three patients undergoing repeated coronary artery occlusions during percutaneous transluminal coronary angioplasty. Arterial blood was removed via an 8F catheter positioned in the femoral artery and delivered by a retroperfusion pumping system to a 7F retroperfusion balloon catheter positioned in the anterior cardiac vein. Ischaemia-related indices were monitored both before and during coronary sinus retroperfusion. These indices included high fidelity left ventricular pressure recordings and pressure derived indices (including velocities of isovolumic contraction and relaxation), as well as electrocardiographic changes and symptoms. Analysis of these variables showed that the ischaemic changes induced during coronary artery occlusion were not prevented by this type of coronary sinus retroperfusion. There was no major complication in any of the patients. It may be that adaptation of the technique or the use of alternative end points will establish a benefit, but further modifications of the delivery system are necessary for effective clinical use.