The documented efficacy of cardioplegic myocardial preservation has led to increased application of coronary artery bypass grafting, with an increasing number of grafts being performed under prolonged continuous cardiac arrest. Coincident with these developments has been an increased incidence of postoperative conduction block and supraventricular tachyarrhythmias. The frequent return of atrial electromechanical activity during the aortic cross-clamp period implies an inadequate degree of protection afforded the atria and supraventricular conduction system by current techniques of myocardial preservation and suggests that postoperative conduction abnormalities may be due to ischemic injury of atrial structures. In eight patients undergoing coronary artery bypass grafting, a marked disparity was noted in the ability to cool the atrial septum and ventricular septum (mean temperatures 25.9° ± 1.3° C versus 15.7 ± 1.3° C, respectively, p <0.05). Moreover, atrial septal temperature returned to the temperature of the systemic perfusate (28° C) within 3 minutes following cessation of cardioplegic infusion while ventricular septal temperature remained at 15° to 20° C. The effect of intermittent topical iced saline slush on atrial septal temperature was transient. Following the period of cardioplegic arrest (average, 29 minutes), all eight patients demonstrated first-degree atrioventricular (AV) block, and prolonged AV conduction persisted for at least 2 hours after release of the aortic cross-clamp. In 10 dogs, similar differences in atrial septal and ventricular septal cooling and rewarming were noted during cardioplegic arrest. The temperature gradients between he atrial septum and the ventricular septum were due in part to differential delivery of the cardioplegic solution, the atrial septum receiving 1.42 ± 0.14 ml/gm compared to 3.01 ± 0.22 ml/gm in the ventricular septum (p <0.05). Regional atrial and ventricular myocardial blood flow was determined before, during, and after cardioplegic arrest. The atrial and ventricular septae showed similarly diminished blood supply during the cross-clamp period (0.044 ± 0.009 and 0.040 ± 0.004 ml/min/gm, respectively) and reactive hyperemia on reperfusion. Detailed AV node, His bundle, and Purkinje system electrophysiological studies before and after cardioplegic arrest supported the concept that inadequate atrial hypothermia during cardioplegic arrest led to ischemic injury of the supraventricular conduction system. Following the period of cardioplegic arrest, all 10 animals demonstrated conduction delay, the site of the delay being primarily within the AV node. Factors implicated in conduction system dysfunction following cardioplegic arrest are discussed, and emphasis is placed on inadequate atrial and AV nodal hypothermia and subsequent ischemic injury as the major causes of postoperative conduction disturbances.