To evaluate the myocardial protective effect of glucose-insulin (GI) in a cold cardioplegic solution, we compared the effect of a solution not containing GI (solution I) with that of an identical solution containing GI (solution II) under cardiopulmonary bypass in mongrel dogs. After 120 minutes of ischemic arrest, defibrillation was necessary to restore cardiac activity in 12.5% of the solution I group and 42.8% of the solution II group, respectively. No distinct intergroup differences were noted in recovery of left ventricular function, oxygen and lactate extraction of the myocardium during the post-arrest period, the myocardial ultrastructure, and myocardial ATP levels. Although creatine phosphate was less well preserved, glucose-6-phosphate and lactate were more accumulated in the solution II group than in the solution I group. Thus, significant advantages of substrate enhancement by addition of GI to cold cardioplegic solution were not found in spite of multidose cardioplegia and presence of noncoronary collateral blood flow.