Surgical treatment of acute aortic dissection has been largely standardized, but some patients develop late complications that require repeat surgery. The 10-year freedom from proximal and distal reoperation is approximately 70-80%, and the risk of proximal reoperation is approximately 10%. Aggressive resection of the aortic root has not been proven to eliminate the need for proximal reoperation, but it appears reasonable for pre-existent root dilatation. The coronary button technique has resulted in greater freedom from death and reoperation and thus appears advisable. The risk of distal reoperations seems generally lower than that of proximal operations (0-4%). Routine extension of replacement into the total arch in the initial operation is associated with increased risk, and it appears not to decrease the probability of reoperation. Aggressive replacement of the total arch may be justified only for connective tissue disease. The choice of thoracic incision is important for aortic reoperations. A median sternotomy provides excellent exposure to the proximal aorta. A posterolateral thoracotomy allows good exposure for distal arch or descending aorta but may be associated with pain-related pulmonary impairment. A bilateral thoracotomy gives easy access for arch and descending aorta but is associated with the highest degree of respiratory impairment. We use a median sternotomy in most instances and choose a bilateral thoracotomy for complex aortic pathology. A posterolateral thoracotomy seems preferable for most distal arch and/or descending aortic replacements. Since femoral arterial cannulation can contribute to an increased incidence of neurological complication, cannulation of the axillary artery appears advisable. We prefer implantation of a Dacron graft and to cannulate the graft. If the aorta is adjacent to the sternum, the patient is placed on cardiopulmonary bypass (CPB) before repeat sternotomy. The patient is cooled to a nasopharyngeal temperature of 28-30 degrees C and at the time of sternotomy CPB is temporarily interrupted. We have not yet had to use transapical venting. Using these approaches we have been able to maintain a procedure-related mortality of 4%, which is not different from primary operations on an aneurysmatic aorta.