From 1978 through 1987, 225 patients underwent operations that included replacement of the ascending aorta. One hundred twenty-three patients underwent composite aortic valve and ascending aortic replacement, 30 had aortic valve replacement with separate graft replacement of the ascending aorta, and 72 underwent replacement of the ascending aorta without aortic valve replacement. Thirty-one (13.8%) in-hospital deaths occurred. Univariate testing of preoperative and operative variables followed by logistic regression analyses identified miscellaneous aortic disease, coronary artery bypass grafting, aortic arch replacement, emergency operation, surgical date (1978 to 1983), and age (all p less than 0.05) as factors having independent association with in-hospital mortality. Follow-up of in-hospital survivors (mean interval 46 months, range 8 to 123 months) documented an overall 5-year survival rate of 76%, 83% after primary operation and 37% after reoperation. Univariate analyses followed by multivariate testing indicated that previous operation (p less than 0.0001) and a history of preoperative neurologic symptoms (p = 0.021) were associated with decreased late survival. At follow-up 88% of late survivors were free of symptoms. Seven patients have undergone reoperation 1 day to 69 months postoperatively. Although the in-hospital mortality for operations that include ascending aortic replacement exceeds that for isolated aortic valve replacement, the late death rate and rate of reoperation are low.