From 1987 to 1994, 116 patients received replacement of the ascending and/or aortic arch using selective cerebral perfusion. They were 82 male and 34 female, with average age of 64 years. There were 63 dissecting and 53 true aneurysms. Extent of replacement was: ascending aorta in 13, aortic root in 2, aortic arch in 93, and aortic root and complete arch in 8. Aortic arch replacements were composed of: 29 partial proximal aortic arch replacements, 44 complete aortic arch replacements, and 20 partial distal aortic arch replacements. Nineteen (16.4%) hospital deaths occurred. Univariate testing of pre-, intra-, and post-operative variables followed by stepwise logistic regression analyses identified elderly, ischemic heart disease, postoperative neurologic complication, cardiac dysfunction, renal failure, and massive bleeding as factors having independent association with hospital mortality. Neurologic complication was found in 10 patients (8.6%), and risk factor for this complication was preoperative peripheral vascular disease. Follow-up of hospital survivors documented an overall cumulative 5-year survive rate of 69%. There was no significant difference between dissection and true aneurysms in 5-year survive ratios, which were 63% and 82%, respectively. During follow-up periods, 18 patients died. Half of these cases were vascular deaths, caused by rupture, sudden death and secondary operation. Univariate analyses followed by stepwise Cox testing indicated that chronic obstructive pulmonary disease and a history of postoperative massive bleeding were associated with decreased later survival. Our experience suggests that selective cerebral perfusion is a safe technique for the repair of ascending aorta and/or aortic arch problems. High-risk subgroups of patients with these aortic problems can be identified by risk factors. Aggressive and careful management is necessary for such subgroups to improve early and late survival rates.