Recently retrograde coronary sinus perfusion technique has been frequently used in patients with severe coronary artery disease. However many untoward effects, such as tissue edema and hemorrhage, have also noticed. To evaluate the efficacy of retrograde cardioplegia, 24 mongrel pentobarbital anesthetized dogs were studied. To create hypoperfused area, distal portion of the left anterior descending coronary artery (LAD) was occluded. After cardioplegic arrest under cardiopulmonary bypass, dogs were assigned following 3 experimental groups (8 dogs each). Group I; Glucose-Insulin-Potassium (GIK) solution (K: 20 mEq/l, 20 ml/kg) was given antegradely into the aorta. Group II; GIK was given retrogradely through the coronary sinus. Group III; GIK was given retrogradely with pulsatile device (synchronized retroperfusion pump system). After 30 minutes, same amount of GIK was given again. Then LAD occlusion was released. Sixty minutes after onset of arrest, the aorta was declamped and cardiopulmonary bypass was stopped. The left ventricular contractility (Emax) measured with conductance catheter at the end of experiment was significantly (p < 0.05) better in groups II (13.1 +/- 2.6, mean +/- SD) and III (13.1 +/- 2.9) than in group I (9.6 +/- 2.7). The left ventricular wall-motion measured with ultrasound crystals in hypoperfused area compared to before cardiopulmonary bypass was also significantly better in groups II (76.2 +/- 17.2%) and III (87.9 +/- 16.9%). Regional myocardial temperature suggested that more rapid and homogeneous cooling including right ventricle was achieved in group III than in groups I and II. Retrograde perfusion is more effective method in the setting of coronary stenosis compared to ordinal antegrade technique.(ABSTRACT TRUNCATED AT 250 WORDS)