BACKGROUND Little is known about the risks of mortality and morbidity after descending thoracic aortic aneurysm repair using left heart bypass and temporary arterioarterial bypass. METHODS A multicenter, retrospective study was performed on 120 patients who were admitted to one of nine cardiovascular centers between January 1988 and December 1993 and underwent operation for nondissecting thoracic aortic aneurysm. The present series included 10 patients with ruptured aneurysm. Graft replacement was performed in 95 patients, patch repair in 22, and suture of the ruptured aorta in 3. Venoarterial bypass was used in 45 patients, left heart bypass in 56, and temporary arterioarterial bypass in 19 as circulatory support. The mean postoperative follow-up period was 30 +/- 21 months. RESULTS Hospital mortality occurred in 7 patients (5.8%). Univariate analysis revealed that only aneurysmal rupture was related to hospital mortality. Brain or cord injury was observed in 4. Of nine deaths that occurred after discharge, five were related to aneurysm and two were due to vascular event. No significant difference was noticed in probability of survival according to the circulatory supporting method. Only aneurysmal rupture affected probability of survival. Multivariate analysis revealed that aneurysmal rupture was the only independent predictor for vascular death including hospital mortality. CONCLUSIONS The present study confirms that aneurysmal rupture is a significant predictor for mortality and morbidity in aortic operations for nondissecting descending thoracic aneurysm, and that a similarly good outcome would be expected when using left heart bypass, temporary arterioarterial bypass, or venoarterial bypass.