Between January 1, 1986, and December 31, 1991, 4,507 adult patients underwent cardiac surgical procedures requiring cardiopulmonary bypass. Of these patients, 3,983 patients who did not undergo operation for supraventricular tachycardia and who were in normal sinus rhythm preoperatively form the study group for the present study. Postoperatively, all patients were monitored continuously for the development of arrhythmias until the time of hospital discharge. The incidence of atrial arrhythmias requiring treatment for the most commonly performed operative procedures were as follows: coronary artery bypass grafting, 31.9%; coronary artery bypass grafting and mitral valve replacement, 63.6%; coronary artery bypass grafting and aortic valve replacement, 48.8%; and heart transplantation, 11.1%. For all patients considered collectively, the risk factors associated with an increased incidence of postoperative atrial arrhythmias (p < 0.05 by multivariate logistic regression) included increasing patient age, preoperative use of digoxin, history of rheumatic heart disease, chronic obstructive pulmonary disease, and increasing aortic cross-clamp time. Postoperative atrial fibrillation was associated with an increased incidence of postoperative stroke (3.3% versus 1.4%; p < 0.0005), increased length of hospitalization in the intensive care unit (5.7 versus 3.4 days; p = 0.001) and postoperative nursing ward (10.9 versus 7.5 days; p = 0.0001), increased incidence of postoperative ventricular tachycardia or fibrillation (9.2% versus 4.0%; p < 0.0005), and an increased need for placement of a permanent pacemaker (3.7% versus 1.6%; p < 0.0005). These data provide a basis for targeting specific patient subgroups for prospective, randomized trials of therapeutic modalities designed to decrease the incidence of postoperative atrial arrhythmias.