Effects of Early Anticoagulation on the Degree of Thrombosis After Repair of Acute DeBakey Type I Aortic Dissection

Transcription

1 Effects of Early Anticoagulation on the Degree of Thrombosis After Repair of Acute DeBakey Type I Aortic Dissection Suk-Won Song, MD, PhD, Kyung-Jong Yoo, MD, PhD, Do-Kyun Kim, MD, Bum-Koo Cho, MD, PhD, Gijong Yi, MD, PhD, and Byung-Chul Chang, MD, PhD Department of Thoracic and Cardiovascular Surgery, Gangnam Severance Hospital, and Department of Thoracic and Cardiovascular Surgery, Yonsei Cardiovascular Hospital, Severance Hospital, Yonsei University Health System, Seoul; Department of Thoracic and Cardiovascular Surgery, Ilsan Hospital, National Health Insurance Corporation, Ilsan; and The Korea Heart Foundation, Seoul, Republic of Korea Background. The degree of false lumen thrombosis after surgical repair of acute DeBakey type I aortic dissection can predict long-term outcomes. However, there are currently no evidence-based recommendations for anticoagulation. We analyzed the effect of early anticoagulation on the residual false lumen and longterm outcomes. Methods. This was a retrospective observational study of 136 patients with acute DeBakey type I aortic dissection who underwent surgical repair between 1997 and We assessed the effect of early anticoagulation on the degree of thrombosis of the false lumen, segmental growth rates, repeat distal procedures, and long-term survival. Results. Among the 136 patients who underwent operations, imaging data in 103 were sufficient for analyzing the degree of thrombosis of the false lumen. Of those, 56 (54%) received anticoagulation therapy immediately postoperatively. The early-anticoagulation group had a higher proportion of completely patent false lumens and lower partial thrombosis than the no-anticoagulation group. Mean segmental aortic growth rate was significantly lower in the early-anticoagulation group than in the no-anticoagulation group ( and mm/year, p ). Overall survival and aorta-related repeat procedure-free survival were significantly better with early anticoagulation than with no anticoagulation (p < 0.05). Conclusions. Early anticoagulation after surgical repair of acute DeBakey type I aortic dissection might have a favorable effect on the onset or extension of thrombosis, aortic growth rate, the need for repeat distal procedures, overall survival, and thrombosis-related complications during long-term follow-up. (Ann Thorac Surg 2011;92: ) 2011 by The Society of Thoracic Surgeons Although the latest data from the International Registry of Acute Aortic Dissection Investigators revealed that in-hospital mortality (23.9%) remains high [1], immediate surgical outcomes for acute aortic dissection have continued to improve at many centers [2 4]. Thus, this is the time to concentrate more on long-term outcomes rather than immediate outcomes after surgical repair of acute DeBakey type I aortic dissection (AIAD). A residual patent false lumen has been reported as a potential risk factor for distal aortic enlargement and poor long-term outcomes [5 7]. However, a recent report by Tsai and colleagues [8] showed that partial thrombosis, which has been thought to be a part of the thrombosed false lumen in patients with acute type B aortic dissection, predicts poor survival. Our previous report [9] also showed that partial thrombosis of the residual false Accepted for publication April 26, Presented at the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, Address correspondence to Dr Chang, Yonsei Cardiovascular Hospital, Severance Hospital, Yonsei University Health System, 134 Shinchondong, Seodaemun-ku, Seoul, Republic of Korea; bcchang@yuhs.ac. lumen after surgical repair of AIAD was an independent risk factor for poor long-term outcomes. We hypothesized that early anticoagulation in the period after repair of AIAD may decrease the incidence of partial thrombosis in the residual false lumen. Although complete thrombosis of the residual false lumen might be a sign of aortic wall healing and remodeling after repair, thrombosis is incomplete in most situations; that is, it is a partial or focal thrombosis and may contribute to malperfusion and distal aortic enlargement during follow-up. Until recently, there have been no recommendations about heparinization or anticoagulation in the period after AIAD repair. In the present study, we assessed the effect of early anticoagulation on the degree of thrombosis of the residual false lumen and long-term outcomes after surgical repair of AIAD. Patients and Methods Between January 1997 and June 2007, 136 consecutive patients (76 men [56%]), who were a mean age of 60 years, underwent surgical repair of AIAD at Gangnam 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1368 SONG ET AL Ann Thorac Surg ANTICOAGULATION AFTER AIAD REPAIR 2011;92: Severance Hospital and Severance Hospital at Yonsei University Health System, Seoul, Korea. Aortic dissection was diagnosed by enhanced computed tomography (CT) or additional findings on echocardiography, and transesophageal echocardiography was always used intraoperatively. We excluded patients with chronic aortic dissections, iatrogenic aortic dissections, or acute aortic dissection limited only to the ascending aorta and proximal aortic arch. The enrolled patients had false lumens that extended from the ascending aorta to the abdominal aorta and were diagnosed with AIAD. The 136 patients underwent emergency operations within 7 days of acute onset of symptoms, and 128 (94.1%) of these were performed within 48 hours of symptom onset. The operation in the remaining 8 patients (5.9%) occurred more than 48 hours after symptom onset due to delay in the diagnosis of AIAD or because of referral from other hospitals. Our study followed the guidelines of the Institutional Review Board of Yonsei University College of Medicine for studies on human beings (Yonsei Institutional Review Board approval number ). We obtained informed consent from all patients or their next of kin for permission to operate and for the use of their clinical records for research purposes. Operative Techniques All operations were performed on an emergency basis within 24 hours after admission. General anesthesia was induced while blood pressure was monitored. We monitored blood pressure and cerebral oxygen saturation using the radial arteries bilaterally, a unilateral femoral arterial catheter, and a regional brain oxygen saturation monitor. A transesophageal echocardiographic probe was inserted and monitored in all patients for diagnostic evaluation and therapeutic confirmation. The surgical procedure was begun through a median sternotomy with a standard cardiopulmonary bypass by cannulation of the right axillary artery, femoral artery, or both. We continuously tried to modify our surgical techniques, such as cannulation techniques, cerebral protection methods, and temperature strategies, as time went on. After 2004, we used simultaneous double arterial cannulation (right axillary artery and femoral artery) and antegrade selective cerebral perfusion under moderate hypothermia. Systemic cooling began immediately after cardiopulmonary bypass was established. When a core temperature of about 28 C was reached (core temperature of cerebral perfusion increased as time went on from 18 to 30 C), unilateral or bilateral antegrade selective cerebral perfusion was initiated by clamping the right brachiocephalic artery, and the ascending aorta was opened. Bilateral antegrade selective cerebral perfusion was performed only when regional brain oxygen saturation decreased more than 10% compared with the baseline value, with insertion of the cerebral perfusion catheter into the left common carotid artery. The aortic arch was explored. If the entry site was in the ascending aorta, an ascending aorta replacement, including hemiarch replacement, was performed using an open distal anastomosis. If the entry site was present in or extended into the aortic arch, total arch replacement was performed. When the entry site could not be identified or was identified in the descending thoracic aorta on transesophageal echocardiography, we simply replaced the ascending aorta. This resulted in graft replacements of the ascending aorta in 63 patients (46%), the hemiarch in 37 (27%), and the total arch in 36 (27%) during the study period. After 2008, although the entry tear was found in the ascending aorta, we tried to replace the total arch, especially in patients who were younger or who had Marfan syndrome. Modified Bentall procedures, which involve aortic root replacement with a composite graft and reimplantation of both coronary arteries, were performed in 12 patients (8.8%) with conspicuous dilatation of the aortic root. Postoperative Anticoagulation Patients with a history of mechanical valve replacement, coronary artery disease, peripheral arterial obstructive disease, aortic root replacement with a mechanical valve at the time of aortic repair, deep vein thrombosis, or any other conditions requiring anticoagulation received postoperative anticoagulation therapy when there was no evidence of mediastinal bleeding or coagulopathy. Target international normalized ratios (INRs) or drug doses were individualized for the standard indications. Follow-Up All data were obtained by a retrospective review of hospital records. Postoperative CT scans were performed serially following our protocols. The immediate CT was performed just before discharge, and the follow-up CT was usually performed at the outpatient clinic 3, 6, and 12 months after hospital discharge and annually thereafter if no major adverse aortic events (MAAE) occurred. Other follow-up data, including overall survival; MAAE, including repeat aortic operations, endovascular aortic repair (EVAR), aortic graft infection, or rupture; and cause of death, were obtained from our outpatient clinic through written or telephone contact with patients or their relatives. The mean follow-up period was years (range, 0.4 to 12.1 years), and follow-up information was recorded for all patients. Degree of Thrombosis in the Residual False Lumen The degree of thrombosis of the false lumen on CT scans was classified as previously reported [8, 9]. The false lumen was classified as being completely patent if flow was present in the absence of thrombus, as having partial thrombosis if both flow and thrombus were present, or as having complete thrombosis if no flow was present. Segmental Aortic Growth Rates The segmental aortic growth rate was assessed in patients who had undergone at least two CT scans postoperatively with at least 6 months between them. If more than two CT scans were performed after hospital discharge, the most recent image was used to determine the aortic growth rate. If an aortic segment had undergone

3 Ann Thorac Surg SONG ET AL 2011;92: ANTICOAGULATION AFTER AIAD REPAIR 1369 repeat aortic procedures, the patient was excluded from further aortic growth rate assessment. Analyses in each patient were performed at the aortic arch, the proximal (at the level of the pulmonary artery bifurcation) descending thoracic aorta, and the abdominal aorta. The growth rate was calculated as [10, 11] the difference in diameter between the immediate postoperative and most recent measurements divided by the time interval between the two measurements. Statistical Analysis All values are expressed as mean standard deviation. Between-group differences in clinical and morphologic variables were analyzed using 2, Fisher exact test, unpaired t test, or Mann-Whitney U test. The time-related events studied included MAAE and death after hospital discharge. Freedom from these time-related events was estimated by the nonparametric actuarial Kaplan-Meier method. Overall survival and freedom from repeat procedures on the distal aorta for all patients with AIAD are also given. All statistical analyses were performed using SPSS software (SPSS, Inc, Chicago, IL). Results Operative Mortality A total of 22 patients (16.1%) died after the operation. Causes of death included multiorgan failure in 8, postoperative bleeding in 7, pulmonary complications in 3, neurologic complications in 2, and heart failure in 2. Degree of Thrombosis in the Residual False Lumen Imaging data were sufficient in 103 of the 114 survivors (90.4%) determine aortic growth rates. The incidence of residual patent false lumen on initial postoperative CT examination was 74% (76 of 103). Of the 76 patients with a residual patent false lumen, 35 (46%) had a completely patent false lumen in the thoracoabdominal aorta. Partial thrombosis was observed in the remaining 41 patients (54%) in the descending thoracic or abdominal aortas. Only 27 patients (26%) had complete thrombosis in the thoracoabdominal aorta. Anticoagulation therapy was administered to 55 patients (54%) immediately after the operation. Compared with the no-anticoagulation group, the anticoagulation group had a higher proportion of completely patent false lumens (42.9% and 23.4%) and lower partially thrombosed false lumens (32.1% and 51.1%; p 0.045; Fig 1). Segment-Specific Aortic Growth Rate Segmental aortic growth rate was analyzed in 392 CT scans available for the 103 patients with adequate imaging data. The mean overall CT follow-up period was years (median, 2.4 years; range, 1.5 to 11.3 years). Within 2 weeks after repair, the median diameters were aortic arch, 3.5 cm; descending aorta, 3.6 cm; and abdominal aorta, 2.4 cm. The descending thoracic aorta at the level of pulmonary artery bifurcation was the site of fastest aortic enlargement. The mean segmental aortic Fig 1. The anticoagulation group had a higher proportion of complete patent false lumen and lower partial thrombosis than the nonanticoagulation group. growth rate was significantly lower in the earlyanticoagulation group than in the no-anticoagulation group at all segments of the distal aorta ( vs mm/year, p 0.018; Table 1). Most patients in the anticoagulation group showed no or slightly enlarged aortic diameter. In contrast, most patients in noanticoagulation group showed rapid aortic enlargement (Fig 2). Long-Term Survival Fourteen patients died during the follow-up period. The most common cause of death was related to neurologic complications, including 3 patients with intracranial hemorrhages and 3 with cerebral infarctions. The second leading cause of death was aorta-related death (n 4). Two patients (14.3%) in the no-anticoagulation group died of aortic aneurysm ruptures in the distal aorta. Two patients died of graft infections (1 in the anticoagulation group and the other in the no-anticoagulation group). Three patients died of cardiac failure, including 1 patient each of myocardial infarction, mediastinitis, and endocarditis. One patient died of acute respiratory distress syndrome. Actuarial survival was 85.5% at 1 year, 66.7% at 5 years, and 66.7% at 10 years (Fig 3A). Cumulative survival analysis showed the anticoagulation group had the superior prognosis (Fig 3B). Risk Factors for Poor Survival The risk factors for poor survival were an estimated glomerular filtration rate of less than 60 ml/m 2, malperfusion, hemodynamic compromise, renal failure, peripheral arterial obstructive disease, old age, postoperative complication, readmission to the intensive care unit, reintubation, partial thrombosis in the false lumen of the distal aorta, and absence of anticoagulation. Cox proportional hazard analysis revealed that an estimated glomerular filtration rate of less than 60 ml/m 2 (odds ratio [OR], 7.567; 95% confidence interval [CI], to ; p 0.042), need for reintubation (OR, 6.664; 95% CI, to ; p 0.011), partial thrombosis in the false lumen of the distal aorta (OR, 3.208; 95% CI, to

4 1370 SONG ET AL Ann Thorac Surg ANTICOAGULATION AFTER AIAD REPAIR 2011;92: Table 1. Segmental Aortic Growth Rate (mm/year) Aortic Segment (n 103) Yes (n 56 [54%]) Anticoagulation No (n 47 [46%]) p Value Aortic arch DTA Abdominal aorta All segments DTA descending thoracic aorta ; p 0.035), and no anticoagulation (OR, ; 95% CI, to ; p 0.016) were independent risk factors for poor survival (Table 2). Major Adverse Aortic Events We previously [9] defined MAAE as the need for distal aorta-related repeat operation, EVAR, aortic rupture, or graft infection [9]. There were 14 repeat procedures of the distal aorta (6 repeat operations, 6 thoracic EVAR, and 2 EVAR), 2 distal aortic ruptures, and 1 graft infection during the 10-year follow-up period. The two distal aortic ruptures occurred in patients who did not take anticoagulants and who died immediately after arrival at the emergency department of our hospital. Four MAAEs occurred in the anticoagulation group and 13 in the no-anticoagulation group (Table 3). For all hospital survivors, freedom from MAAE was 93.7% at 1 year, 81.5% at 5 years, and 68.3% at 10 years (Fig 4A). The rate of MAAE was significantly higher in the no-anticoagulation group than in the anticoagulation group (p 0.016; Fig 4B). Among 14 repeat procedures, there were 6 thoracic EVAR and 2 EVAR, which were performed with technical success, and all patients were alive at the last follow-up. Late Repeat Procedures on the Distal Aorta Repeat procedures on the distal aorta were performed when a residual aortic diameter of more than 60 mm was found or a patient had symptoms of an enlarged aortic dissecting aneurysm. During the follow-up period, 14 patients (12.3%) underwent 14 late repeat procedures on Fig 2. Three examples of the aortic remodeling process, with or without anticoagulation therapy. Two patients with anticoagulation showed (A) a complete thrombosis of the residual false lumen or (B) did not show an aneurysmal dilatation. (C) Another patient without anticoagulation showed an aortic dilatation with an aortic growth rate of 10 mm/year.

5 Ann Thorac Surg SONG ET AL 2011;92: ANTICOAGULATION AFTER AIAD REPAIR 1371 the distal aorta, comprising the aortic arch and descending thoracic aorta in 2, the descending thoracic aorta in 6, the thoracoabdominal aorta in 2, and the abdominal aorta in 4. Of the 14 patients who underwent late repeat procedures on the distal aorta, 10 (71%) belonged to the no-anticoagulation group and 4 to the anticoagulation group. All repeat procedures were performed electively, and there was no periprocedural death associated with these repeat procedures. Comment Table 2. Risk Factors for Poor Survival Factor p Value OR (95% CI) a p Value egfr 60 ml/min/m ( ) Malperfusion Hemodynamic compromise Renal failure PAOD Age Post-op complications ICU readmission Reintubation ( ) DTA-FL partial ( ) thrombosis No anticoagulation ( ) a Cox proportional hazard analysis. CI confidence interval; egfr estimated glomerular filtration rate; DTA-FL descending thoracic aorta false lumen; ICU intensive care unit; OR odds ratio; PAOD peripheral arterial obstructive disease. Although the immediate surgical outcomes of AIAD repair have recently improved with advances in surgical techniques and perioperative care [2 4], surviving the emergency operation does not guarantee freedom from subsequent MAAE during long-term follow-up. Many previous reports [5, 6, 12] have demonstrated that later aortic events are strongly related to the characteristics of the residual false lumen of the distal aorta after AIAD repair, the extent of initial operation, Marfan syndrome, younger age, and uncontrolled hypertension. The characteristics of the false lumen include the degree of thrombosis, the absolute size, and the relative size compared with the true lumen in the residual thoracoabdominal aorta. Tsai and colleagues [8] reported that the natural course of partial thrombosis of a false lumen in an acute type B aortic dissection has a worse prognosis than that of the completely patent group. In our previous study [9], we hypothesized that the morphologic status of the residual distal aorta after repair of an AIAD would mirror a type B aortic dissection. We concluded that partial thrombosis of the false lumen after repair of AIAD is an independent risk factor for rapid aortic dilatation and aorta-related repeat procedures, as well as overall survival. Song and colleagues reported that the upper descending thoracic aorta is the major site of late aneurysmal dilatation and that an upper descending thoracic aorta false lumen diameter larger than 22 mm showed higher adverse events, such as aneurysm or deaths [13]. Others have reported that a more radical and extensive ap- Table 3. Major Adverse Aortic Events by Anticoagulation Status Group/procedure Aortic Graft Reoperation EVAR Rupture Infection No. No. No. No. Fig 3. (A) Overall survival is shown after surgical repair of acute DeBakey I aortic dissection (AIAD). (B) Cumulative survival rate is shown after surgical repair of AIAD according to the presence (AC, solid line) or absence (NAC, dashed line) of anticoagulation. Anticoagulation... DTA No anticoagulation 1 DTA DTA TAR TAAA AAA AAA abdominal aortic aneurysm; DTA descending thoracic aorta; EVAR endovascular aortic repair; TAAA thoracoabdominal aortic aneurysm; TAR total arch replacement.

6 1372 SONG ET AL Ann Thorac Surg ANTICOAGULATION AFTER AIAD REPAIR 2011;92: Fig 4. (A) Aorta-related repeat procedure-free survival rate after surgical repair of acute DeBakey type I aortic dissection. (B) Cumulative aorta-related repeat procedure rate after surgical repair of acute DeBakey type I aortic dissection according to the presence (solid line) or absence (dashed line) of anticoagulation therapy. (AC anticoagulation; MAAE major adverse aortic events; NAC nonanticoagulation.) proach, such as total arch replacement, has similar morbidity and mortality compared with less invasive procedures, such as ascending aorta or hemiarch replacement, and they suggested that total arch replacement could help the long-term remodeling processes of the residual aorta after repair [3, 12, 14, 15]. However, initial operations for AIAD fail to achieve complete resection of all entry tears, particularly in patients with primary entry or reentry located in the descending thoracic or abdominal aorta. In all cases, many reentries remain in the residual thoracoabdominal aorta after central repair of AIAD, and these reentries take the form of a patent false lumen. Residual patent false lumen is already a well-known risk factor for future aortic enlargement [5, 6, 12] and has been reported to be related to unresected intimal tears, leakage from needle holes at the distal anastomosis site, and reentry in the distal aorta after surgical repair of AIAD. Currently, many surgeons advocate insertion of a short segment of the elephant trunk [16], and furthermore, some perform a frozen elephant trunk technique, which is expected to enhance the obliteration of the residual false lumen of the thoracoabdominal aorta [17 20]. Zierer and colleagues [21] reported that uncontrolled hypertension is a risk factor for late aortic enlargement. There is clear consensus among physicians that the use of an -blocker, -blocker, and angiotensin II receptor blocker in patients who undergo surgical repair of AIAD helps the remodeling process in the residual distal aorta. As far as we know, no report has analyzed the correlation between the fate of the residual aorta and anticoagulation after repair of AIAD. In 2008, Lachat and colleagues [22] first raised a question about the use of early anticoagulation in patients with acute type B aortic dissection. They cautiously made a demand on the medical community for performing studies focused on coagulation disorders, thrombosis extension, and anticoagulation in acute aortic dissection, especially of the type B variety. Although there are no current recommendations regarding the use of anticoagulation, many physicians believe anticoagulation after AIAD repair would have a negative effect on remodeling of the residual distal aorta. However, our data revealed that anticoagulation could prevent formation of partial thrombosis of the false lumen in the residual aorta and decrease the false lumen pressure, which results in decreased wall tension, thereby preventing aneurysmal dilatation of the dissected aorta and, eventually, aortic rupture. Anticoagulation may also prevent or treat end-organ thromboembolism from the false lumen through multiple reentries. In the postrepair period of AIAD, false lumen thrombosis onset, location, and extension are unpredictable, and data are lacking on the use of anticoagulation after repair of AIAD. The question of how to manage postrepair anticoagulation is unanswered. Many physicians have had experiences with end-organ thromboembolism, such as renal infarction, or splenic infarction, in patients who have undergone AIAD repair. The proper management in that situation would be to initiate intravenous heparinization with anticoagulation until the INR was at the therapeutic target level. The use of anticoagulation immediately after the repair would not only prevent end-organ thromboembolism but also help in the remodeling process of the residual distal aorta by decreasing false lumen pressure through prevention of thrombosis. As a result of this study, our postoperative anticoagulation protocol has been changed as follows: Intravenous low-molecular-weight heparin and anticoagulation are started on the second or third postoperative day when there is no evidence of mediastinal bleeding. The target activated partial thromboplastin time is between 60 and 90 seconds, and the target INR is about 1.5. If the aorta CT before discharge shows a false lumen with partial thrombosis in the residual thoracoabdominal aorta, the target INR is about 2.0. If there is no partial thrombosis of the false

7 Ann Thorac Surg SONG ET AL 2011;92: ANTICOAGULATION AFTER AIAD REPAIR 1373 lumen on the predischarge CT of the aorta, the target INR is slightly decreased. Anticoagulation is stopped, unless otherwise indicated, and the follow-up aorta CT shows a completely thrombosed false lumen. However, anticoagulation therapy continues if the patient has a history of mechanical valve replacement, coronary artery disease, peripheral arterial obstructive disease, aortic root replacement, deep vein thrombosis, and all other conditions requiring anticoagulation. In our series, we had a high incidence of distal repeat procedures (37.5% cumulative reoperation rate at 10 years), similar to incidences reported by other groups. There were 14 aortic repeat procedures during the follow-up period; of these, 4 occurred in the anticoagulation group, and 10 occurred in the no-anticoagulation group. DeBakey and colleagues [23] reported that rupture of the distal aorta was the most common cause of death in patients with AIAD, accounting for 29.3% of late deaths. In our series, 2 patients died of rupture of the distal aorta during the follow-up period: 1 occurred 1,781 days after ascending aorta replacement, and the other at 1,113 days after hemiarch replacement. CT scans revealed that the maximum diameters of the descending thoracic aortas were 79 and 46 mm, respectively. One patient was in the completely patent group, and the other in the partial thrombosis group. Neither patient received anticoagulation and both died immediately after arrival at the emergency department of our institute. Because only 2 patients presented with ruptures, it would be difficult to extrapolate potential risk factors or anatomic findings of aortic ruptures. However, our impression is that even though the diameter of the distal aorta was less than 5 or 6 cm, partial thrombosis of the distal false lumen might have been an aggravating factor for aortic rupture. Although careful follow-up is necessary to prevent rupture of the distal aorta, the decision to perform repeat procedures on the aorta in a stable, asymptomatic patient is another important issue. For timely repeat procedures, we must take into account not only comorbidities and aortic diameter but also the status of the residual false lumen. Favorable results of endovascular treatment for acute or chronic aortic dissection were recently reported, with high rates of success in cases of false lumen thrombosis [24]. When we consider aorta-related repeat procedures, endovascular treatment can be an effective alternative, especially for elderly or severely compromised patients who are at higher risk. Our study had several limitations: First, this study was performed retrospectively, so the indications for using anticoagulation were too wide to analyze the sole effect of anticoagulation on the degree of thrombosis or long-term outcomes after surgical repair of AIAD. For example, patients with Marfan syndrome who underwent aortic root replacement should use anticoagulation; however, the natural course of the aorta in the setting of Marfan syndrome has a worse prognosis than that of a non- Marfan aorta. Second, the initial and follow-up CT scans after discharge at our hospital and referring hospitals were not standardized. Therefore, we were unable to accurately evaluate the initial aortic diameter and long-term changes in the diameter of the true and false lumens. Third, follow-up CT scans to analyze the cumulative rate of aortic dilatation or time-related change of the thrombosis characteristics in the false lumen were performed at variable intervals after the surgical repair. However, we attempted to perform follow-up CT scans according to our protocol after surgical repair of acute aortic dissection. The immediate postoperative CT scan was performed just before discharge, and follow-up CT scans were usually performed at our outpatient clinic 6 to 12 months after hospital discharge and annually thereafter. Fourth, the follow-up period was relatively short. The mean follow-up was only 4.3 years, and the results were not sufficient to show long-term outcomes after AIAD repair. With a longer follow-up period, we could precisely evaluate changes in the characteristics of the false lumen. Further studies involving larger numbers of patients with longer follow-up are required. In conclusion, the present study revealed that early anticoagulation after surgical repair of AIAD might have a favorable effect on the degree of thrombosis of the residual false lumen, aortic growth rate, aorta-related repeat procedures, and overall survival. We believe that in addition to the usual medical therapy, which includes lowering blood pressure and heart rate, the early use of anticoagulation prevents partial thrombosis in the false lumen of the residual aorta and may provide a more favorable long-term prognosis after surgical intervention for AIAD. We are grateful to Sun-Hee Lim, RN, and Min-Sung Cho, RN for collecting the data used in this article. This study was supported by a faculty research grant of Yonsei University College of Medicine for 2008 ( ). References 1. Rampoldi V, Trimarchi S, Eagle KA, et al. Simple risk models to predict surgical mortality in acute type A aortic dissection: the International Registry of Acute Aortic Dissection score. Ann Thorac Surg 2007;83: Westaby S, Saito S, Katsumata T. Acute type A dissection: conservative methods provide consistently low mortality. Ann Thorac Surg 2002;73: Takahara Y, Sudo Y, Mogi K, Nakayama M, Sakurai M. Total aortic arch grafting for acute type A dissection: analysis of residual false lumen. Ann Thorac Surg 2002;73: Hata M, Shiono M, Sezai A, Iida M, Negishi N, Sezai Y. Type A acute aortic dissection: immediate and mid-term results of emergency replacement with the aid of gelatin resorcin formalin glue. Ann Thorac Surg 2004;78: Yeh CH, Chen MC, Wu YC, Wang YC, Chu JJ, Lin PJ. Risk factors for descending aortic aneurysm formation in medium-term follow-up of patients with type A aortic dissection. Chest 2003;124:

8 1374 SONG ET AL Ann Thorac Surg ANTICOAGULATION AFTER AIAD REPAIR 2011;92: Immer FF, Hagen U, Berdat PA, Eckstein FS, Carrel TP. Risk factor for secondary dilation of the aorta after acute type A aortic dissection. Eur J Cardiothorac Surg 2005;27: Halstead JC, Meier M, Etz C, et al. The fate of the distal aorta after repair of acute type A aortic dissection. J Thorac Cardiovasc Surg 2007;133: Tsai TT, Evangelista A, Nienaber CA, et al. Partial thrombosis of the false lumen in patients with acute type B aortic dissection. N Engl J Med 2007;357: Song SW, Chang BC, Cho BK, et al. Effect of partial thrombosis on distal aorta after repair of acute DeBakey type I aortic dissection. J Thorac Cardiovasc Surg 2010;139: Wolf YG, Thomas WS, Brennan FJ, Goff WG, Sise MJ, Bernstein EF. Computed tomography scanning findings associated with rapid expansion of abdominal aortic aneurysms. J Vasc Surg 1994;20: [Erratum: J Vasc Surg 1995;21:295.] 11. Kazi M, Thyberg J, Religa P, et al. Influence of intraluminal thrombus on structural and cellular composition of abdominal aortic aneurysm wall. J Vasc Surg 2003;38: Hirotani T, Nakamichi T, Munakata M, Takeuchi S. Routine extended graft replacement for an acute type A aortic dissection and patency of the residual false channel. Ann Thorac Surg 2003;76: Song JM, Kim SD, Kim JH, et al. Long-term predictors of descending aorta aneurismal change in patients with aortic dissection. J Am Coll Cardiol 2007;50: Kazui T, Washiyama N, Muhammad BA, et al. Extended total arch replacement for acute type A aortic dissection: experience with seventy patients. J Thorac Cardiovasc Surg 2000;119: Urbanski PP, Siebel A, Zacher M, Hacker RW. Is extended aortic replacement in acute type A dissection justifiable? Ann Thorac Surg 2003;75: Miyamoto S, Hadama T, Anai H, et al. Simplified elephant trunk technique promotes thrombo-occlusion of the false lumen in acute type A aortic dissection. Ann Thorac Cardiovasc Surg 2006;12: Kato M, Kuratani T, Kaneko M, Kyo S, Ohnishi K. The results of total arch graft implantation with open stent-graft placement for type A aortic dissection. J Thorac Cardiovasc Surg 2002;124: Liu ZG, Sun LZ, Chang Q, et al. Should the elephant trunk be skeletonized? Total arch replacement combined with stented elephant trunk implantation for Stanford type A aortic dissection. J Thorac Cardiovasc Surg 2006; 131: Uchida N, Ishihara H, Shibamura H, Kyo Y, Ozawa M. Midterm results of extensive primary repair of the thoracic aorta by means of total arch replacement with open stent graft placement for an acute type A aortic dissection. J Thorac Cardiovasc Surg 2006;131: Shimamura K, Kuratani T, Matsumiya G, et al. Long-term results of the open stent-grafting technique for extended aortic arch disease. J Thorac Cardiovasc Surg 2008;135: Zierer A, Voeller RK, Hill KE, Kouchoukos NT, Damiano RJ Jr, Moon MR. Aortic enlargement and late reoperation after repair of acute type A aortic dissection. Ann Thorac Surg 2007;84: Lachat M, Criado FJ, Veith FJ. The case for anticoagulation in patients with acute type B aortic dissection. J Endovasc Ther 2008;15: DeBakey ME, McCollum CH, Crawford ES, et al. Dissection and dissecting aneurysms of the aorta: 20-year follow-up of 527 patients treated surgically. Surgery 1982;92: Kusagawa H, Shimono T, Ishida M, et al. Changes in false lumen after transluminal stent-graft placement in aortic dissection: six years experience. Circulation 2005;111: DISCUSSION DR GORAV AILAWADI (Charlottesville, VA): I have a couple of questions for you. First, how did you select which patients were going to get anticoagulated, and what was the indication? Second, what was your anticoagulation protocol? Specifically, what was your goal international normalized ratio (INR), when was anticoagulation initiated postoperatively, and how long did the therapy continue? My third question relates to your inclusion of Marfan patients in this cohort. I guess I would caution you that those patients are different and their growth rates we know are going to be different if they have a dissection. And of those Marfan patients, how many of them were stratified in the no anticoagulation versus the anticoagulation group? DR SONG: Thank you for your questions. On the first question, the anticoagulation was begun in patients with preoperative malperfusion flow limitation within the true lumen due to enlarged false lumen thrombosis which was determined by clinical signs of patients, and based on computed tomography (CT), previous mechanical valve replacement, coronary artery disease, peripheral arterial obstructive disease, aortic root replacement, deep vein thrombosis, and all other conditions requiring anticoagulation. Regarding the second question, the anticoagulation protocol as a result of this study is now intravenous low-molecularweight heparin, with anticoagulation started on the second or third postoperative day when there is no evidence of mediastinal bleeding. The target activated partial thromboplastin time was between 60 and 90 seconds, and target INR was around 1.5. If the aorta CT before discharge showed a false lumen with partial thrombosis in the residual thoracoabdominal aorta, the target INR was around 2.0. If there was no partial thrombosis of the false lumen on the predischarge CT of the aorta, the target INR was slightly decreased. Anticoagulation was stopped unless otherwise indicated, and the follow-up aorta CT showed a completely thrombosed false lumen. Regarding the third question on the patients with Marfan syndrome, we had 11 Marfan patients in our series. Most of them underwent Bentall procedures with a mechanical aortic valve. So those patients always took anticoagulation therapy. However, the result during the follow-up period revealed that the Marfan aorta was not inferior to non-marfan aorta in terms of aortic growth, aorta-related repeat procedures, and overall survival. DR R. SCOTT MCCLURE (London, Ontario, Canada): My first question is again with regards to the anticoagulation protocol. I still didn t quite get it. I know you said there were differences, but did you have certain INR parameters? Was it lifelong anticoagulation? I just want a little bit more clarity about the protocol for anticoagulation that was utilized. And my second question was with regards to preoperative variables. I understand that your main interest was the lumen size and survival and that is what you were focused on, but did you look at other preoperative variables between the group that had anticoagulation and did the group that did not? And do you think maybe any such differences, if there were any, might have had any bearing on the outcomes in the end?

9 Ann Thorac Surg SONG ET AL 2011;92: ANTICOAGULATION AFTER AIAD REPAIR 1375 DR SONG: I did not fully understand your question. DR ERIC ROSELLI (Cleveland, OH): The question was regarding your definition of anticoagulation, was it Coumadin or platelet therapy or both? Also, how do you anticoagulate the patients now? DR SONG: I have already said that this is a retrospective study, and our anticoagulation group had very wide indications. So antiplatelets, such as aspirin or clopidogrel, or anticoagulation treatment, such as warfarin therapy, were all included in the treatment groups. So my conclusion is not definitive, but I said that, based on this study, I start anticoagulation therapy as early as possible, on the second or third morning. DR ROSELLI: Do you use warfarin therapy? DR SONG: Yes, with warfarin with intravenous heparinization. DR ROSELLI: I think this is a really thought-provoking study and I congratulate you, and your other work has been excellent, also. But I think just as a comment and then a question, it seems that there can be quite a leap to think that anticoagulation is actually what is making a difference in the patients outcomes. When we treat a dissection and the false lumen is completely thrombosed, it usually gets resorbed and the whole thing heals. That brings up one of my concerns with the study by Tsai also: namely, how they define partial thrombosis. In Cleveland when we evaluate patients acutely and in follow-up for aortic dissections, all of those patients are evaluated with a study that is very similar to the studies that we do when we follow-up a stent graft. That is, they get a three-phase study: a noncontrast study, a study where the contrast bolus is timed for their arterial phase, and a study where the contrast bolus is delayed by 5 minutes for what we call a delayed venous phase. Oftentimes we are not surprised to see what looks like a thrombosed false lumen in the arterial phase study has a lot of contrast in it on the delayed venous phase study, sometimes even where the false lumen looks brighter than the true lumen. And so my question to you is based on the fact that this definition of partial thrombosis is a difficult one to make. What imaging technique do you use to follow these patients? And certainly I would suggest that if any future studies are done prospectively, that they be done with that kind of imaging technique: with both an arterial and delayed venous phase of contrast. DR SONG: Thank you for your excellent comments and questions. I completely agree with you that the definition of partial thrombosis of the false lumen was not clear, and our analysis would be more statistically important if this study was performed with a three-phase study as your comments. However, I analyzed the false lumen with CT scans only in the arterial phase, and defined the partial thrombosis as flow and thrombose were present at the same time in the arterial phase. In a future study I would like to perform a three-phase CT study like yours. Thank you.