Objective: Aneurysms and dissections of the thoracic aorta continue to present a surgical challenge and their incidence is increasing in recent years. The mortality rate of surgical treatment is still higher than those of other cardiovascular operations. Neurological injury is the most feared complication resulting from repair of these lesions. This study aims to determine the factors that influence the neurological outcome and mortality after thoracic aortic operations. Methods: During the period from November 1993 through May 1999, 144 patients were operated on for conditions involving the ascending aorta and/or aortic arch. Ninety-five (66.0%) were operated for aortic dissection and 49 (34.0%) were for aortic aneurysms. Sixty-two patients (43.1%) had replacement of ascending aorta with distal open technique; 82 patients (56.9%) had hemiarch or total arch replacement or repair of the distal arch. Results: Twenty-seven (18.7%) early deaths occurred. New stroke occurred in two patients (1.4%) and temporary neurological dysfunction in nine patients (6.3%). Deep hypothermic circulatory arrest with retrograde cerebral perfusion was used in all patients. On multivariate logistic regression analysis, risk factors for mortality were chronic renal failure, preoperative organ malperfusion, rupture, total circulatory arrest time, 60 minutes, postoperative acute renal failure, postoperative low cardiac output, sepsis, and multiple organ failure. Risk factors for neurological morbidity were preoperative chronic renal failure, preoperative hemodynamic instability, postoperative low cardiac output, and pulmonary complications. Conclusions: Hypothermic circulatory arrest with retrograde cerebral perfusion was not an independent predictor of neurological morbidity on multivariate analysis, even if the arrest period was more than 60 minutes. Lengths of circulatory arrest periods and clinical presentations of the patients are important determinants of mortality.