This study was undertaken to determine the factors that influence the final outcome after hypothermic circulatory arrest. Between 1985 and 1992 a uniform method of hypothermic circulatory arrest was used in 200 patients as the primary method of cerebral protection during operations on aneurysms of the thoracic aorta. There were 30 hospital deaths (15%). Age greater than 60 years (relative risk 3.7, p < 0.02), emergency operation and hemodynamic compromise (relative risk 22.2, p < 0.000), concomitant procedures (relative risk 2.7, p < 0.04), presentation with new neurologic symptoms (relative risk 5.2, p < 0.04), and postoperative permanent neurologic deficits (relative risk 9.4, p < 0.000) were found to be significant predictors of operative mortality. A total of 183 patients were available for evaluation of neurologic function and outcome. Multivariate analysis of this cohort of patients by multiple logistic regression showed that temporary neurologic dysfunction occurred in 36 cases (19%). Temporary neurologic dysfunction correlated with the duration of hypothermic circulatory arrest (47 ± 16 minutes; odds ratio 1.06/minute; p < 0.001) and age (66 ± 14 years; odds ratio 1.07/year; p < 0.001). Embolic strokes occurred in 22 patients (11%) and were associated with permanent deficits in 13 (7%). Strokes correlated significantly with age (older than 60, 21% versus younger than 60, 1%; p < 0.001) and operations on the arch and descending aortic aneurysms containing clot or atheroma (p < 0.001). This experience shows that the operative mortality is not affected by any parameters related to the use of hypothermic circulatory arrest. The incidence of temporary neurologic dysfunction rises linearly in relation to the age of the patient and the duration of hypothermic circulatory arrest. However, permanent neurologic injury is a result of thromboembolic events and is not related to the method of cerebral protection used. Additional methods to prevent perioperative embolic strokes are needed. Hypothermic circulatory arrest affords adequate cerebral protection if the arrest period is kept less than 60 minutes. We will continue to use this modality until the safety and utility of the alternate methods of cerebral protection are shown to be superior.