BACKGROUND Transesophageal echocardiography (TEE) is a useful means in the diagnosis of acute aortic dissection (AD), owing to its very high sensibility and specificity. In this study, TEE was performed to assess post-surgical evolution. METHODS Between 1982 and 1991, 119 pts. were operated on in our institution for AD (De Bakey I and II type): 87 pts. underwent replacement of the ascending aorta with a composite tubular graft bearing a mechanical valve; 26 had a simple tubular graft and 6 had aortic reconstruction. Sixty-eight of 72 discharged pts. were followed for up to 9.5 years (mean 4.5 +/- 2.6). Nine years after surgery actuarial survival of discharged pts. was 75%. Seven pts. died after a mean period of 3.4 years from surgery: only one died from postoperative complication (dehiscence of proximal anastomosis), none for aortic rupture distal to the graft. TEE was performed in 32 of these pts. and in other two operated on elsewhere, after 4.4 +/- 2.7 years from surgery; before the operation, type I AD was diagnosed in 23 pts. and type II in 11 pts. RESULTS In 10/11 pts. with type II AD the aortic arch and the descending aorta looked normal; in one patient a localized intimal flap was found up to the arch. The descending aorta diameter was somewhat higher than in normal subjects (25.2 +/- 2.8 vs 21.9 +/- 3.7 mm), but in only one case was it beyond 2DS (32 mm). In all type I pts. an intimal flap persisted distal to the graft, along the whole thoracic aorta. Within the false lumen a flow was detected by color-Doppler in 14/23 pts. (61%), and spontaneous echo-contrast was noted in 14 pts. (61%). A thrombus was observed in 7 pts. (30%) and it was generally localized; in only one case it was extensive with total obliteration of the false lumen. In 16 pts. (70%) communications between the two lumina were found. The descending aorta diameter ranged from 25 to 53 mm, and mean value was higher than in normal subjects (34.2 +/- 6.2 vs 21.9 +/- 3.7 mm). CONCLUSIONS In most pts. with type II AD, surgery can be a definitive treatment, as the remaining aorta keeps to normal size and appearance. In type I AD, operation is only palliative, as the dissection persists: the false lumen is often perfused through one or more communications with the true lumen and seldom its obliteration is noted. The persistence of dissection does not necessarily seem to be an ominous finding, as the survival of the study population was high and no patient died from aortic rupture. Nevertheless, long-term prognosis can be affected by aorta dilation that often (but not always) follows the persistence of wall dissection. For its high reliability, easy feasibility and low cost TEE is a very useful method for following up patients operated on for AD and for detecting those who are at higher risk of aortic rupture because of lumen dilation.