Penetrating atherosclerotic ulcers of the aorta

Transcription

1 Penetrating atherosclerotic ulcers of the aorta James A. Harris, MD, Kostaki G. Bis, MD, John L. Glover, MD, Phillip J. Bendick, PhD, Anil Shetty, PhD, and O. William Brown, MD, RDyal Oak, Mich. Purpose: This study investigates the natural history and optimal imaging modality of penetrating atherosclerotic ulcers of the aorta. Methods: We reviewed our experience with 29 penetrating ulcers in 18 patients. Computed tomography (17 patients), magnetic resonance imaging (nine patients), and aortography (five patients) were used for diagnosis and follow-up. Patients were typically elderly (average age 74 years) and had hypertension and coronary artery disease. Ulcers were most common in the distal descending thoracic aorta (31 %) and were characterized by a discrete ulcer crater (100%) and thickened aortic wall (89%). Modes of presentation included chest or back pain in four patients, distal embolization in two patients, and abnormal chest radiography results in one; the remaining were incidental findings. Results: Follow-up was available in ten patients with 17 ulcers from 1 to 7 years. Recurrent pain occurred in two patients, recurrent embolization occurred in one patient, and seven patients remained symptom free. Progression to saccular pseudo aneurysm occurred in five ulcers, and fusiform aneurysm occurred in two ulcers. Two ulcers were associated with an increase in aortic diameter, and nine ulcers did not change. There were no cases of aortic dissection or rupture in the follow-up period. There were no deaths and only one patient underwent resection. Conclusion: The natural history of penetrating atherosclerotic ulcers is one of progressive aortic enlargement, with saccular and fusiform aneurysms the result if follow-up is sufficient. Aortic dissection, aortic rupture, and embolization can also occur but are less common. Contrast-enhanced computed tomography is the primary imaging modality. (J VASe SURG 1994;19:90-9.) In 1986 Stanson et al. described penetrating atherosclerotic ulcers of the aorta as a distinct clinical and pathologic entity in which ulceration penetrates the internal elastic lamina into the media and is associated with a variable amount of hematoma within the aortic wall. 1 It is distinct from, but may be associated with, aortic dissection, saccular or fusiform aneurysms, and intimal ulceration. Penetrating atherosclerotic ulcers typically occur in elderly patients with a history of hypertension, hyperlipidemia, and severe atherosclerotic disease. Although often symptom free, these patients may have chest or back From the Departments of Surgery and Radiology (Drs. Bis and Sherry), William Beaumont Hospital, Royal Oak. Presented at the Forty-first Scientific Meeting of the International Society for Cardiovascular Surgery, North American Chapter, Washington, D.C., Tune 7-8, Reprint requests: John L. Glover, MD, Chief, Department of Surgery, William Beaumont Hospital, 3601 W. 13 Mile Rd., Royal Oak, Ml Copyright 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /94/$ /6/ pain. The ulcers are most often found in the distal descending thoracic aorta but can occur throughout the thoracic and abdominal aorta and have a characteristic appearance on computed tomography (CT) and magnetic resonance imaging (MRI). The clinical significance of these lesions is unclear because little is known about their natural history. Most cases reported in the literature have undergone surgical resection, l and the others have had only short-term follow-up.2,3 This report describes the changes that occurred in 10 patients who had serial imaging ofl7 ulcers during a 1- to 7-year follow-up period. In addition we compared different imaging techniques in another eight patients with 12 ulcers, for whom no follow-up was available. METHODS To document the natural history of penetrating atherosclerotic ulcers, we reviewed our experience with 18 patients with 29 ulcers. The average age was 74 years, with a range of 62 to 87. There were 10 men and eight women. Other medical problems were common in these patients, particularly hypertension

2 JOURNAL OF VASCULAR SURGERY Volume 19, Number 1 Harris et at. 91 (14 patients, 78 %) and coronary artery disease (13 patients, 72%). Both diabetes mellitus and chronic renal insufficiency occurred in five patients (28%). The most common site of penetrating ulcers was the distal descending thoracic aorta, where nine (31 %) were found. The infrarenal aorta was the next most common site, with seven (25%), followed by the mid (five ulcers, 17%) and proximal (four ulcers, 14% ) descending thoracic aorta, the suprarenal aorta (three ulcers, 10%), and the aortic arch (one ulcer, 3%). Twelve patients had a total of 17 concomitant aneurysms, nine of which were associated with the penetrating ulcer. Aneurysms were most commonly found in the thoracic aorta (n = 8); arch (n = 1), proximal descending thoracic aorta (n = 2), mid descending thoracic aorta (n = 1), and distal descending thoracic aorta (n = 4). Other sites included the infrarenal abdominal aorta (n = 5), popliteal artery (n = 2), iliac artery (n = 1) and splenic artery (n = 1). Chest or midscapular back pain was the presenting symptom in four patients with atherosclerotic ulcers in the thoracic aorta. Four patients had ulcers found during work-up of concomitant aneurysms: two infrarenal, one thoracic, and one popliteal. Distal ischemia caused by emboli occurred with no other source for emboli in two patients. One patient was diagnosed with distal right foot ischemia caused by popliteal artery embolus. The other patient was diagnosed with bilateral distal foot ischemia caused by multiple emboli at the trifurcation level. Extensive workup in both patients revealed an infrarenal aortic ulcer with no other source of emboli. One patient was diagnosed with a lower extremity deep venous thrombosis, underwent heparin therapy, and subsequently had an acute drop in his hemoglobin level. Workup of a periaortic density on chest radiography in this symptom-free patient revealed a penetrating thoracic aortic ulcer with mediastinal hemorrhage. The other ulcers were found incidentally in patients with symptoms not attributed to the ulcer. Three of these were in patients being monitored for cancer, one with breast cancer, one with lung cancer, and one with pancreatic cancer. The imaging methods used for diagnosis and follow-up included CT (with and without contrast) in 17 patients, conventional aortography in five patients, and MRI (non-fat saturation spin echo, fat saturation spin echo, and angiography) in nine patients. Initial CT studies were performed with either a Siemens DRH (Siemens Medical Corp., Iselin, N.J.) or Somatom DR3 scanner (Siemens Medical Corp.), and later with a Siemens Somatom Plus (Siemens Medical Corp.). Contrast (120 to 150 cc 60% iodinated contrast) was infused via a power injector with an initial rate of 1.5 to 2.0 cc/sec lasting 45 to 60 sec and subsequently at a rate of 0.08 to 1.5 cc/sec. MRI was performed at 1.5 T with a Siemens SP Magnetom (Siemens Medical Corp.). Tl-weighted spin echo (SE) sequences were performed in eight patients. Electrocardiography gating was used for obtaining an image of the thoracic aorta with triggering to every R wave and a TR/TE of 600 to 800 msec/34 msec. The TE was prolonged to increase spin dephasing to allow for better flow void phenomenon in ulcers that contain sluggish blood flow. Abdominal SE sequences were performed without gating and after intramuscular glucagon injection (1 mg) with a TR/TE of 600 to 700 msec/34 msec. Fat saturation was achieved with SE-chemical shift selective technique (TRITE 600 to 800 msec/te msec) in six patients. Shimming for improving magnetic field homogeneity was performed before fat saturation sequences with a matrix of256 by 192. Finally, images of three patients were obtained with two-dimensional time-of-flight magnetic resonance angiography (TRITE 40 to 55 msec/8 msec, 256/192 matrix, two signal averages, 5 mm slice thickness with 30% overlap). The ulcer width, maximum aortic diameter at the ulcer level, and aortic wall thickness at the ulcer margin were measured. The diagnosis on CT and MRI was made by identifying a well-defined ulcer crater in the aortic wall that demonstrated contrast enhancement on CT, flow void phenomenon on SE and SE-fat saturation MR, and flow-related enhancement on two-dimensional-time-of-flight magnetic resonance angiography. Presence or absence of associated dissection, intramural or surrounding hemorrhage, pleural effusion, and acute dissection were assessed. Displaced intimal calcification could only be seen on CT. RESULTS Clinical follow-up was available in 10 patients with 17 ulcers, ranging from 1 to 7 years. Only one patient was treated surgically. Control of blood pressure was the primary method of treatment for the four patients with chest or back pain from thoracic ulcers. All four had resolution of symptoms, but two had recurrent pain. One patient with four discrete ulcers in the thoracic aorta had recurrent chest pain, presumed to be due to coronary insufficiency. Coronary artery grafting was performed, but the patient

3 92 Harris et al. JOURNAL OF VASCULAR SURGERY January 1994 Fig. 1. A, Thoracic aortogram shows distinct ulcer crater in distal descending thoracic aorta. B, CT scan shows ulcer has caused focal periaortic wall bulging. C, CT scan after 4 years shows progression to 7.2 cm saccular pseudoaneurysm with intraluminal thrombus. D, CT scan after 6 years, with good contrast enhancement, shows continued enlargement to 7.8 cm. continued to have occasional chest discomfort. Radiologic follow-up of this patient over a I-year period showed no change in three ulcers, but aortic enlargement (3.1 cm to 3.7 cm) at the site of a distal descending thoracic aortic ulcer. The other patient had recurrent sharp midscapular back pain and episodes of epigastric pain. Radiologic follow-up of this patient over a 6-year period showed progressive aortic enlargement (3.0 cm to 4.0 cm) at the site of a mid descending thoracic aortic ulcer and progression to 7.8 cm saccular aneurysm of a distal descending thoracic ulcer (Fig. 1). She refused operative repair. Two patients had no recurrent pain. One patient with a proximal descending thoracic aortic ulcer was symptom free after 7 years. Chest radiography at 2 years showed no change in the dilated proximal aorta. She refused radiologic follow-up at 7 years. The other patient was symptom free after 6 years but had ulcers in the mid and distal descending thoracic aorta that progressed to a complex 5.0 cm saccular/fusiform aneurysm (Fig. 2) and a 5.0 cm fusiform aneurysm, respectively. Of the two patients with distal ischemia caused by emboli, one had recurrent embolism. This patient required transmetatarsal amputation of his right foot as a result of the nonhealing ischemic wound for which he presented. He later was admitted with blue

4 JOURNAL OF VASCULAR SURGERY Volume 19, Number 1 Harris et at. 93 Fig. 2. A, Thoracic aortogram shows bilobed ulcer in middescending thoracic aorta. B, Contrast-enhanced CT shows upper portion of bilobed ulcer. Also seen is adjacent atelectasis and pleural effusion. C, CT scan after 6 years shows progression of upper portion to 5.0 em saccular pseudoaneurysm. toe syndrome from an embolus and required amputation of the great toe. In the patient with mediastinal hemorrhage from a thoracic aortic ulcer, heparin was discontinued, and a caval filter was placed. He remained symptom free after 6 years, but the ulcer progressed to a 7.5 em saccular aneurysm. All other patients remained symptom free. There were no deaths. One patient had resection of an infrarenal ulcer in conjunction with resection of a right iliac artery aneurysm. This patient had an asymptomatic penetrating ulcer on the posterior surface of the infrarenal aorta, which had burrowed through the aortic wall to form focal periaortic wall dilation (Fig. 3). A 4 em segment of aorta was resected, and the saccular dilation is seen in Fig. 3,C. A cross section shows the discrete ulcer margins where the intima and media end abruptly (Fig. 3,D). Microscopically, the intima and elastic fibers of the media ended abruptly, leaving thrombus and connective tissue (adventitia) to form the wall of the lesion (Fig. 3, ). There was no significant extension of intramural hematoma beyond the ulcer edges in this case. The preceding 10 patients underwent radiologic studies during a 1- to 6-year follow-up period (Table I), and longer follow-up was associated with a greater likelihood of progression to aneurysm. Six penetrating atherosclerotic ulcers with a mean follow-up of 4.6 years progressed to aneurysms. Two ulcers in the distal descending thoracic aorta and two in the infrarenal aorta progressed to saccular aneurysms. One bilobed ulcer in the mid descending thoracic

5 94 Harris et at. JOURNAL OF VASCULAR SURGERY January 1994 Fig. 2. D-F. D, Contrast-enhanced CT scan shows lower portion of bilobed ulcer at presentation. E, CT scan after 10 days shows lower portion of ulcer has begun to incorporate itself into aortic wall. F, CT scan after 6 years shows concentric aortic dilation to 5.0 cm fusiform aneurysm. aorta progressed to a complex saccular and fusiform aneurysm (Fig. 2). The average growth rate of those progressing to saccular aneurysms was 0.31 cm per year or 7.2% per year. Progressive.enlargementofthe ulcer to saccular aneurysm occurred in all cases once focal periaortic adventitial bulging was observed. An additional ulcer in the distal descending thoracic aorta progressed to a fusiform aneurysm. Two ulcers, with a mean 3.5 years of follow-up, were associated with an average 20% increase in aortic diameter at the site of the ulcer (5.7%/yr). As the aortic dilation occurred, the definition of the ulcer craters disappeared. (2,D to F). This same process occurred in ulcers that became fusiform aneurysms, in contrast to those that progressed to saccular aneurysms. Average growth rate of the four ulcers associated with concentric aortic dilation was 0.33 cm/yr or lo%/yr. Nine penetrating ulcers and the aortic diameter at that level were unchanged in the follow-up period, but five of these ulcers were adjacent to fusiform aneurysms, three of which enlarged in the follow-up period. In addition to the preceding 17 ulcers with multiple serial studies, 12 ulcers in eight patients were evaluated to determine the characteristic features of these lesions. Several patients were studied with different imaging modalities for comparison of their efficacy. At diagnosis, all penetrating atherosclerotic ulcers had a discrete ulcer crater, 89% had aortic wall thickening, 67% had intramural hematoma, and 44% had displaced intimal calcifications. Ulcer width and

6 JOURNAL OF VASCULAR SURGERY Volume 19, Number 1 Hams et al. 95 Fig. 3. A, Digital subtraction angiogram shows penetrating ulcer on posterior surface of infrarenal aorta. B, Spin-echo MR (non-fat saturation) shows ulcer just distal to left renal vein. aortic wall thickness on cross-sectional imaging averaged 1.4 cm (range 0.7 to 2.5 cm) and 0.8 cm (range 0.4 to 2.0 cm) at the time of diagnosis. Pleural effusions were noted in three patients with ulcers in the descending thoracic aorta, atelectasis in two, and acute dissection in one, and mediastinal hemorrhage mone. False-negative diagnoses with cr (43%) were due to lack of contrast (24%) or suboptimal aortic enhancement (19%). Knowledge of these ulcers was based on previous or follow-up contrast CT or MRI. Renal insufficiency or contrast allergy were the reasons that contrast was not used. Suboptimal aortic enhancement was due to either technique (slow infusion rate, suboptimal timing of initiation of scanning) or patient-related problems (poor circulation and atherosclerotic disease, poor venous access). There were no false-positive diagnoses with CT. False-negative diagnoses with non-fat saturation MRI were seen in 4.8% as a result of stagnant blood flow within the ulcer. False-negative diagnoses with fat saturation MRI was high (29%) because of poor shimming and subsequent degradation of image quality (metal clips in three, poor positioning of magnetic field in three). One false-positive diagnosis with MRI was due to signal void from focal aortic wall calcification that mimicked flow void phenomenon within the ulcer crater. Magnetic resonance angiography with two-dimensional time-of-flight technique supplemented Table 1. Radiologic follow-up (lo patients) Progressed to aneurysms (4.6 yr) Saccular (5) Fusiform (2)* Increase in aortic diameter (3.5 yr) No change (2.2 yr) Adjacent to enlarging aneurysms (3) Adjacent to stable aneurysms (2) *One bilobed ulcer progressed to a complex saccular/fusiform aneurysm. MRI in three patients. Two of the ulcers were visualized, but one was not seen as a result of saturation of stagnant blood flow within the ulcer. DISCUSSION Experience with penetrating atherosclerotic ulcers is limited, as evidenced by the few reports found in the literature, and the natural history is unknown because nearly all cases reported either have undergone resection or have limited follow-up. We elected to treat the patients in this series without operation for several reasons: their advanced age and poor general health, the high morbidity and mortality rates associated with surgical resection, and the lack of data concerning the clinical course of these lesions. This is the first series of patients with these lesions in which long-term follow-up is available; and No

7 96 Harris et at. JOURNAL OF VASCULAR SURGERY January 1994 Fig. 3. CoB. C, Resected aortic specimen with saccular dilation posteriorly. D, Cross section of resected specimen shows discrete ulcer margins with thrombus-lined pseudoaneurysm. B, Photomicrograph shows intima (In) and elastic fibers of media (Me) end abruptly, leaving thrombus (Tb) and adventitia (Ad) to form wall of pseudoaneurysm. Atheroma is represented by At. combined with other reports, it provides at least some initial perspective on the natural history of penetrating atherosclerotic ulcers. The literature has emphasized comparing and differentiating aortic dissection and atherosclerotic ulcers. Several reports described aortic dissection initiating at the base of an atherosclerotic ulceration. 4-6 Tisnado et al. 4 described ''ulcerlike outpouchings" on aortography and believed that this represented an early sign of aortic dissection. Gore and Hirse noted atherosclerotic ulcers as a rare cause of chronic aortic dissection and found them more often originating in the distal descending thoracic aorta, where atherosclerotic changes are more common. Roberts 8 also noted the inverse relationship between severe atherosclerosis and aortic dissections. The differences in distribution and radiographic characteristics between dissections and ulcers were described in several reports. 1,2,9-11 The clinical presentation of penetrating ulcer may be similar to that of dissection, but dissection is an uncommon consequence of ulceration. Stanson and associates 1 described intramural hematoma extending from the ulcer in 10 of 14 patients who underwent operation, but only one patient was described as having extensive dissection. Similarly, we had only one patient diagnosed with aortic dissection originating from an atherosclerotic ulcer, and no patient progressed to dissection in the follow-up period. Although some degree of intramural hematoma occurs in most ulcers, propagation of the dissection is probably prevented because of extensive fibrosis of the aortic wall from longstanding atherosclerosis. Embolism from intimal ulcerations in peripheral vessels occurs commonly from the carotid bifurcation. Lord and Berry12 described atherosclerotic ulcerations of the aortic arch and proximal great vessels and attributed retinal embolism to these

8 JOURNAL OF VASCULAR SURGERY Volume 19, Number 1 Harris et at. 97 lesions. No cases of embolization have been attributed to the penetrating atherosclerotic ulcer since it was defined in 1986 until this report. One patient had peripheral embolization on two separate occasions from an infrarenal atherosclerotic ulcer, both resulting in ischemic necrosis and amputation. From available reports, free transmural rupture seems to be uncommon. Nora and Hollier13 reported one case of contained transmural rupture of the suprarenal aorta from a penetrating atherosclerotic ulcer. Kazerooni et al. ll also described a patient with contained extrapleural hemorrhage as a result of a leaking penetrating ulcer in the thoracic aorta. Stanson and associates 1 referred to several patients with contained rupture of the aorta, and note that containment is provided by hematoma and adventitia. The hematomas in their specimens were no more than 2 weeks old. We found hematoma formation and progression to saccular dilation to be a chronic process, and there is a question of appropriate terminology when these lesions extend beyond the wall but are relatively small. "Pseudoaneurysm" refers to contained leakage of blood through all layers of the wall and "diverticulum" seems inappropriate for an artery. Progression of these lesions to transmural rupture was. uncommon; only one patient had hemorrhage from a penetrating ulcer. He remained symptom free after 6 years and had slow progressive enlargement of his aneurysm. Most impressive about the natural history of penetrating atherosclerotic ulcers is the propensity toward aneurysm formation. Cooke et al. 2 described a saccular pseudoaneurysm arising from a penetrating atherosclerotic ulcer in the proximal descending thoracic aorta. Stanson and associates 1 reported seven of 16 patients with what they described as contained rupture, and five of our 10 patients have aneurysms formed from atherosclerotic ulcers. The normal aging process of the aorta involves fragmentation of elastic fibers, increase in collagen fibers at the expense of smooth muscle cells, and an increase in ground substance resulting in weakening of the aortic wall and progressive enlargement of the lumen. Although atherosclerosis is mainly a disease of the intima, it has a secondary effect of weakening the media and therefore contributes to the tendency for aortic dilation. Penetrating atherosclerotic ulcers accelerate this process locally, eventually destroying the media and weakening the adventitia to initiate focal aneurysm formation. Although this process is chronic, once periaortic adventitial bulging occurs, sufficient follow-up is likely to show progressive aneurysmal dilation. It seems reasonable to assume that most saccular aneurysms of the aorta believed to be due to atherosclerosis actually originate from penetrating atherosclerotic ulcers, and growth rates of aneurysms in this series were consistent with those of other reports.14 Serial studies of four ulcers showed concentric aortic dilation and loss of definition of the walls of the ulcer crater. Two were associated with an increase in aortic diameter, and two progressed to frank fusiform aneurysms. The mere presence of penetrating atherosclerotic ulcers indicates advanced atherosclerosis, which itself accelerates generalized aortic dilation at the site of the ulcer or at distant sites. This tendency toward aneurysm formation is evidenced by the high incidence of concomitant aneurysms in these patients. Kazerooni et al. 11 noted six of 16 patients with concomitant aneurysms; three focal abdominal aortic aneurysms, one focal aneurysm of the ascending thoracic aorta, one descending thoracic aortic aneurysm, and one that extended from the aortic root to the abdominal aorta. Stanson et al.1 reported 12 ofl6 patients with concomitant but unassociated aneurysms. In this series of 18 patients, there were 17 concomitant aneurysms, eight of which were not associated with the ulcer. The progression of atherosclerotic ulcers to aneurysm has obvious implications for treatment. As demonstrated by this study, this progression is slow, with a low incidence of acute rupture or other life-threatening complications. Consequently, immediate aortic repair is not mandatory, even in the patient with symptoms. Close follow-up is essential for all penetrating ulcers, even those found incidentally, because they have also been shown to progress to aneurysm. Age and general health of the patient, the location of the ulcer, the presence of saccular or concentric dilation, and the rate of growth should all be considered when deciding whether resection is appropriate. One of our patients with two episodes of peripheral arterial embolism did not undergo operation, first because of severe heart disease and later because he refused. Normally we consider embolism an indication for surgery. In cases with intramural blood and pleural effusion, we individualize therapy, as is done for patients with type III aortic dissection. Advances in imaging techniques have allowed a more detailed view of the aortic wall, and have better defined the penetrating atherosclerotic ulcer. Ulcers found incidentally in symptom-free patients are certainly more common than symptomatic ulcerations. cr has been established as an accurate modality for imaging penetrating ulcers Diagnosis of dissection and penetrating ulcers with CT,

9 98 Harris et at. JOURNAL OF VASCULAR SURGERY January 1994 however, requires optimal intravenous contrast infusion. Nonetheless, if there is good venous access and there are no contraindications to iodinated contrast, CT provides a rapid and accurate means of assessing the thoracic and abdominal aorta. MR! has also been described in recent literature,10 however only non-fat saturation spin-echo techniques were used. Although fat saturation MR! highlighted the aortic wall as a result of fat suppression, it is technically difficult, requiring additional time for shimming and proper patient positioning. Non-fat saturation spin-echo MR!, however, had the highest accuracy when compared with contrastenhanced CT and fat saturation MR!. Because CT is more readily available and less costly, it is the primary imaging modality. MR! is used as a problem-solving tool when there are equivocal CT studies or when contrast injection is contraindicated. In conclusion the natural history of penetrating atherosclerotic ulcers is one of progressive aortic enlargement, commonly resulting in saccular and fusiform aneurysms if follow-up is sufficient. Aortic dissection, aortic rupture, and embolization can also occur but are less common. CT with good contrast enhancement is the primary imaging modality. REFERENCES 1. Stanson AM, Kazmier FJ, Hollier LH, et al. Penetrating atherosclerotic ulcers of the thoracic aorta: natural history and clinicopathologic correlations. Ann Vase Surg 1986; 1: Cooke JP, Kazmier FJ, Orszulak TA. The penetrating aortic ulcer: pathologic manifestations, diagnosis, and management. Mayo Clin Proc 1988;63: Hussain S, Glover JL, Bree R, Bendick PJ. Penetrating atherosclerotic ulcers of the thoracic aorta. J V ASC SURG 1989;9: Tisnado J, Cho S, Beachley MC, Vines FS. Ulcerlike projections: a precursor angiographic sign to thoracic aortic dissection. Am J RoentgenoI1980;135: Dinsmore RE, Rourke JA, DeSanctis RD, Harthome JW, Austen G. Angiographic findings in dissecting aortic aneurysm. N Engl J Med 1966;275: Stein HL, Steinberg I. Selective aortography, the definitive technique for diagnosis of dissecting aneurysm of the aorta. Am J RoentgenoI1968;102: Gore I, Hirst AE. Dissecting aneurysm of the aorta. Cardiovasc Clin 1973;5: Roberts We. Aortic dissection: anatomy, consequences and causes. Am Heart J 1981;101: , Welch TJ, Stanson AW, Sheedy PF, Johnson CM, McKusick MA. Radiologic evaluation of penetrating atherosclerotic ulcer. Radiographies 1990;10: Yucel EK, Steinberg FL, Egglin TK, Geller SC, Waltrnan AC, Athanasoulis CA. Penetrating atherosclerotic ulcers: diagnosis with MR imaging. Radiology 1990;177: Kazerooni EA, Bree RL, Williams DM. Penetrating atherosclerotic ulcers of the descending thoracic aorta: evaluation with CT and distinction from aortic dissection. Radiology 1992;183: Lord RSA, Berry NA. Atherosclerotic ulceration of the brachiocephalic artery. Aust N Z J Surg 1974;44: Nora JD, Hollier LH. Contained rupture of the suprarenal aorta. J VASC SURG 1987;5: Hirose Y, Hamada S, Takamiya M, et al. Aortic aneurysm: growth rates measured with CT. Radiology'l992;185: Submitted June 10, 1993; accepted Aug. 25, DISCUSSION Dr. Ralph B. Dilley (La Jolla, Calif.). The authors conclude from their detailed observations, which were very well documented, that the natural history of aortic ulceration is more benign than we had previously believed. It is important to ask whether the patients represent those patients who have done well and whether this is really a very highly selected group. It may well be that aortic ulceration produces a spectrum of symptoms. At one end of the scale is a relatively benign course, which may be anticipated, whereas on the other end, perforation with hemothorax, perforation with hemopericardium, aortic dissection, and embolization to the viscera or lower extremities may all occur, which generally requires urgent surgical intervention or, when it occurs in an emergency room, results in death. In these latter examples, it's extremely unlikely that patients would be included in a natural history study. How many other patients with aortic ulcerations or problems associated with them underwent operation in your hospital or were admitted to the hospital during a similar time period and are not included in this study? One of the very important aspects of the study was a demonstration of the high incidence of aneurysmal formation at or adjacent to the site of the ulceration. Although this has been reported in the past, this study demonstrates this point in a prospective fashion: six of 17 ulcers progressed to an aneurysm at the site of ulceration, and at the site of two further ulcers aortic diameter increased. Five

10 JOURNAL OF VASCULAR SURGERY Volume 19, Number 1 Harris et al. 99 ulcers were adjacent to stable or enlarging aneurysms, and in all, including those patients in whom follow-ups were not as long, there were 17 concomitant aneurysms. The implication of these data is that most patients will ultimately need operation for their aneurysmal disease, regardless of whether an ulcer is associated with it, adjacent to it, or asymptomatic. Because there is a high incidence of aortic ulceration with aneurysmal disease, is it possible, at least in the fusiform aneurysms, to tell from your data which came first? You have prospective data in some of your patients that ulceration was accompanied by progression to aneurysm, but in how many patients was the aneurysm there at the time the ulcer was discovered and may then just represent ulceration in an aneurysm? Would you give us some guidelines about who should undergo operation when initially seen with symptomatic aortic ulceration? In two of your patients, embolic disease was present, necessitating amputation in at least one. And in most clinics, this would be an indication for repair of an ulcerating lesion. In addition, I note that pleural effusions were present in two of six patients with ulcerations in the descending thoracic aorta, associated with intramural hematoma. I would be quite reluctant to treat these patients without operation, in view of the possibility of rupture and death. Consequently, I believe we need some guidelines on how to decide the appropriate therapy. About the imaging techniques, when you evaluate these patients, do you use 1 cm or V2 cm cuts? In our experience with 1 cm cuts, which are standard in our clinic, we have missed one of these aortic ulcerations, and we're going back and using 1/2 cm cuts so we are able to diagnose them. This is a well-documented and well-presented study that suggests that certain patients with aortic ulceration may have a benign prognosis. However, I am not convinced that this is the case with all patients with asymptomatic aortic ulceration, and that although this natural history data is important new information, it may represent only one end of a spectrum of this complex and interesting entity. Dr. Kenneth J. Cherry (Rochester, Minn.). Since the report of Stanson and his colleagues in 1986, which concerned penetrating ulcers of the thoracic aorta only, we have had the opportunity to monitor approximately 30 more patients with penetrating ulcers of both the thoracic and abdominal aortas. Dr. Stanson, who is a vascular radiologist, has monitored these patients closely and gets serial CTs on them at 3, 6, and 12 months. His observations l and ours place us somewhat in agreement with Dr. Harris and colleagues, which is a distinct change from our previous position. We have had one patient with two connecting penetrating ulcers and a consequent thoracoabdominal aneurysm who required repair at presentation. Two other patients have required operation at presentation, one for atheroembolic ischemia and one for persistent pain. But in those who had been monitored by him, only two other patients have enlarged aneurysmal proportions requiring operation. Dr. Stanson finds that the best mode of monitoring these patients is noncontrast CT. He believes he can get a much better image of the hematoma and has found in those patients that the hematoma resolves at 4 months. The aortic wall remodels and does indeed enlarge as you have pointed out, but in most patients that he's monitored, that enlargement has seemed to stabilize at what we might term ectatic but not aneurysmal levels. Like Dr. Dilley, we find that standard CT is not precise enough, and we have been using Imatron CT (Imatron, Inc., San Francisco, Calif.) with 6 mm slices, and I understand from Dr. Stanson that the new generation Imatron will even be able to give more precision than that, and they're looking forward to that. We also agree that it is a spectrum of disease, with pain. We have had one patient with an ulcer with paraplegia, and we have had patients with atheroembolization. Dr. James A. Harris. We have not had other patients undergo operation for penetrating atherosclerotic ulcers. This may be due to the lack of awareness of this entity by cardiothoracic surgeons, pathologists, and radiologists. The true incidence of those lesions is probably much higher than is reported. We have had the opportunity to obtain images of these patients very early, while their ulcer was still confined to the aortic wall. We have therefore documented that the ulcer is the initial lesion in the progression to aneurysm, rather than the aneurysm occurring first and the ulcer lying within the aneurysm. Ulcerations in existing aneurysms are assumed to be in intraluminal thrombus rather than the aortic wall itself. Patients with asymptomatic penetrating ulcers should, at a minimum, be monitored by cross-sectional imaging. Progression to aortic dissection can be treated with blood pressure control, as would a type III aortic dissection. Embolization and hemorrhage is considered an indication for surgery. The presence of saccular pseudoaneurysm is not an absolute indication for resection because this is not an acute event. These patients should be individualized according to their general health, location of lesion, rate of growth, and presence of saccular or fusiform dilation. The occurrence of chest or back pain from penetrating ulcers in the thoracic aorta is treated without operation initially, unless there is recurrent and persistent symptoms. Several of our patients either had prohibitive general health or refused operative repair. We have typically used 8 to 10 mm cuts with CT. On follow-up studies we have concentrated on the ulcer site and decreased the cuts to 5 mm. Even though some small ulcers will be missed with 10 mrn cuts, the clinical significance of these lesions is really not known at this point anyway.