OBJECTIVE The effect of terminal warm blood cardioplegia was analyzed in 191 patients undergoing either coronary artery bypass grafting (CABG) or prosthetic heart valve replacement between Jan. 1990 and Dec. 1995. METHODS Patients were subdivided into 3 historical cohorts based on the method of myocardial protection: Group A (n = 106), multidose cold crystalloid glucose-potassium cardioplegia, alone; Group B (n = 37), cold crystalloid glucose-potassium cardioplegia plus terminal warm blood cardioplegia, Group C (n = 48), cardioplegia induction with cold crystalloid glucose-potassium cardioplegia, maintenance with multidose cold blood cardioplegia, and terminal warm blood cardioplegia. RESULTS Of patients undergoing CABG, 5.6% of group A, 70.4% of group B, and 86.7% of group C spontaneously resumed sinus rhythm after aortic declamping, as did 9.1% of group A, 60.0% of group B, and 55.6% of group C of patients undergoing prosthetic heart valve replacement. The incidence of spontaneous recovery was significantly better in groups B and C than in group A (p < 0.05). Over 90% of patients without terminal warm blood cardioplegia developed ventricular fibrillation or tachycardia requiring electrical cardioversion (p < 0.05). Postoperatively, patients without terminal warm blood cardioplegia required temporary epicardial pacing more frequently than those with terminal warm blood cardioplegia (p < 0.05). In patients undergoing prosthetic heart valve replacement, groups B and C, the incidence of postoperative atrial fibrillation was significantly lower than in group A. CONCLUSIONS Terminal warm blood cardioplegia thus promoted better postoperative electrophysiological cardiac recovery.