© 2014 Wiley Periodicals, Inc.
641
DISEASES OF THE AORTA ORIGINAL ARTICLE _____________________________________________________________
Low Incidence of Late Pseudoaneurysm and Reoperation After Conventional Repair of Acute Type A Aortic Dissection Ron-Bin Hsu, M.D., and Jeng-Wei Chen, M.D. Department of Surgery, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan, ROC ABSTRACT Background and Aim: Suture line dehiscence and pseudoaneurysm formation is one of the leading causes of late reoperation after surgical repair of acute type A aortic dissection (AAD). A sandwich technique can affect the need of reoperation. We sought to assess the late outcomes (mortality and reoperation) of a modified reinforced sandwich technique in conventional AAD repair. Methods: Retrospective review of 63 consecutive patients undergoing AAD repair between 2003 and 2013. Aortic anastomosis was performed with a modified reinforced sandwich technique using Hemashield strips and two-layer polypropylene continuous and interrupted mattress sutures. Results: Marfan syndrome was diagnosed in five (8%) and bicuspid aortic valve in three patients (5%). Twenty-one patients (33%) had preoperative cardiogenic shock necessitating inotropic support. Replacement of the ascending aorta with aortic valve preservation was performed in 58 (92%) and hemiarch replacement in five patients (8%). Four patients died during initial hospitalization, yielding a hospital mortality of 6%. Median follow-up duration was 73 months (range, 1–124). Kaplan–Meier survival rates were 94 W 3%, 84 W 5%, and 59 W 11% at 1, 5, and 10 years. One patient (1.7%) required proximal reoperation 44 months after AAD repair because of progressive dilatation of the aortic root. No patient had severe aortic regurgitation or pseudoaneurysm after AAD repair. Actuarial freedom from reoperation at 1, 5, and 10 years was 100%, 97%, and 97%. Conclusions: A reinforced sandwich technique was a good technique resulting in a low incidence of late reoperation and pseudoaneurysm formation. doi:
10.1111/jocs.12359 (J Card Surg 2014;29:641–646) Acute type A aortic dissection (AAD) is a surgical emergency having a very poor prognosis.1 The goals of surgical treatment in AAD are to prevent death from aortic rupture, to reestablish blood flow in organs with malperfusion, and to correct severe aortic valvular regurgitation if it is present.1 However, operations do not remove the entire dissected aorta. Most operative survivors have a persistent, dissected residual aorta, often with a patent false lumen. Thus the need for late reoperation remains fairly common after AAD repair.2–14 Several retrospective studies have reported the incidence and mechanisms of late reoperation.2–14 However, the results were usually affected by use of
several different surgical techniques2–14 and involvement of many cardiac surgeons.10 Suture line dehiscence and pseudoaneurysm formation is one of the leading causes of late reoperation after AAD repair.2,3,6–10 Several surgical techniques have been introduced for a more secure aortic anastomosis. However, the long-term efficiency of each sandwich technique is unknown.15–19 Controversy still remains about whether different surgical techniques can affect the need for late reoperation.2–19 A sandwich technique should prevent suture line dehiscence and pseudoaneurysm formation.9 We sought to assess the late outcomes (mortality and reoperation) of a modified reinforced sandwich technique after conventional AAD repair.
Conflict of interest: The authors acknowledge no conflict of interest in the submission.
MATERIALS AND METHODS
Funding: None.
Patients
Address for correspondence: Dr. Ron-Bin Hsu, M.D., National Taiwan University Hospital, No. 7, Chung-Shan S. Rd., Taipei, Taiwan 100, ROC. Fax: þ886-2-2341-0933; e-mail: [email protected]
This was a retrospective, observational cohort study of prospectively collected data. We included all
642
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
J CARD SURG 2014;29:641–646
consecutive patients who have undergone AAD repair by a single surgeon (R.B. Hsu) between September 2003 and May 2013 at the National Taiwan University Hospital. Patients with iatrogenic AAD were excluded. All data were collected by retrospective chart review. The local institutional medical ethics committee approved the study and waived the need for informed consent. Operation All operations were performed on an emergency basis. Standard procedures were used for femoral or subclavian artery cannulation, median sternotomy, and bicaval venous cannulation for total cardiopulmonary bypass. Once adequate core hypothermia was achieved, the ascending aorta was opened and the heart was arrested by continuous retrograde crystalloid cardioplegia. Before September 2004, deep hypothermic circulatory arrest (18 8C) was used for the distal aortic anastomosis with an open technique in patients undergoing retrograde cerebral perfusion. After September 2004, moderate hypothermic circulatory arrest (25 8C) was used for the distal anastomosis with an open technique in patients undergoing selective antegrade cerebral perfusion through an 8 mm tube graft connected to the right subclavian artery. The aortic repair was performed with several techniques. The open distal anastomosis was performed first (arch first). The ascending aorta containing the area of most severe injury and intimal tear, if present, was resected and replaced with a Hemashield tubular prosthetic graft (Hemashield, Boston Scientific, Oakland, NJ, USA). If the intimal tear in the aorta was localized to the ascending aorta, the distal anastomosis was done just proximal to the innominate artery. If the intimal tear originated in the aortic arch, aortic replacement was in most cases extended to include excision of the segment of aortic arch containing the intimal tear (hemiarch replacement). After completion of the distal anastomosis, the arterial cannula was shifted to the prosthetic graft and cardiopulmonary bypass was reconstituted. Replacement of the ascending aorta with aortic valve preservation was performed. The geometry of the sinotubular junction was restored without aortic valve resuspension. Composite graft replacement of the aortic valve, sinuses, and ascending aorta with a modified Bentall technique was performed in patients with severe root dilatation. Separate replacement of the supracoronary ascending aorta and aortic valve was performed in patients with aortic valve calcification. Reinforced sandwich technique Two Hemashield strips, which were harvested from the tube graft and placed outside the adventitia and inside the intima, were used for the sandwich technique. The sandwich anastomosis was reinforced with two layers of 4/0 polypropylene continuous and interrupted mattress sutures (Fig. 1). The first layer was a 4/0 polypropylene continuous running suture.
Figure 1. Scheme of a reinforced sandwich technique in open aortic anastomosis: Two Hemashield strips and two layers of 4/0 polypropylene continuous vertical and interrupted horizontal mattress sutures for reinforcement.
Multiple 4/0 polypropylene interrupted nonpledgetted horizontal mattress sutures were passed from the inside to the outside of the sandwich anastomosis. This reinforced aorta was then sutured end-to-end to a Hemashield tubular graft using a 4/0 polypropylene continuous suture. No gelatin-resorcinol-formaldehyde glue or other biologic glue was put inside the false lumen to reconstitute the dissected aortic layers. Tissucol duo fibrin glue (Tissucol duo, Baxter AG, Vienna, Austria) was used as a sealant on the external suture line. Definition Dissection was considered acute when treated within 14 days from the onset of symptoms. Pseudoaneurysm was defined as a localized hematoma or false aneurysm at the proximal or distal suture lines.2,3,6–10 Reoperation included any cardiac or vascular surgical intervention with an indication that could be related to complications occurring in the segments of aorta involved in the initial repair of AAD or to complications stemming from the spontaneous evolution of the dissection at sites not treated in the initial operation. The indications for reoperation included aortic diameter exceeding 60 mm, rapid progression of an aortic aneurysm exceeding 1 cm per year, and presence of pseudoaneurysm or severe aortic regurgitation.2–14 Reoperation was defined as a proximal reoperation if the operation involved the aorta before the replaced graft and distal reoperation if the operation involved the aorta immediately after the replaced graft.
J CARD SURG 2014;29:641–646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
Follow-up Follow-up was conducted by telephone interviews with surviving patients or with patients’ families, or by direct examination at the outpatient clinic from July to September 2013. Information included the patients’ functional status and the results of follow-up echocardiographic and computed tomographic scans. The severity of aortic regurgitation was evaluated with serial echocardiograpy every year. Aortic size was evaluated with serial computed tomographic scans every year. Statistical analysis Categorical variables were reported as the percentage of patients in the subgroup. Continuous variables with normal distribution were presented as mean SD; those with skewed distribution were presented as medians with ranges. Survival and time to event (mortality and reoperation) was estimated by the Kaplan–Meier method. RESULTS Patients In the period 2003 to 2013, 63 consecutive patients underwent emergency surgical repair of AAD. There were 40 males and 23 females with the median age of 61 years (range, 27 to 90). AAD was DeBakey type I in 57 (90%) and type II in six patients (10%). Intramural hematoma of the ascending aorta was present in 10 (17%) of 57 DeBakey type I AADs. Baseline patient characteristics and predisposing factors for aortic dissection are shown in Table 1. The most common underlying medical disorders were hypertension in 45 patients (71%), and diabetes mellitus in eight patients (13%). Marfan syndrome was diagnosed in five patients (8%). Bicuspid aortic valve was present in three patients (5%). Twenty-one patients (33%) had preoperative cardiogenic shock necessitating at least temporary inotropic support. Cardiac tamponade and shock was present in 19 patients (30%) before operation. Clinically significant malperfusion syndrome was noted in two patients (3%). The affected end organs were the myocardium (acute myocardial infarction) and carotid artery (stroke) each in one patient. Femoral artery cannulation was used in the initial eight patients and subclavian artery cannulation was used in 55 patients. The methods of brain protection included aortic cross-clamping without hypothermic circulatory arrest in a Marfan patient with Debakey type II AAD, deep hypothermic circulatory arrest alone in one redo Bentall patient, deep hypothermic circulatory arrest with additional retrograde cerebral perfusion in eight patients, and moderate hypothermic circulatory arrest with additional selective antegrade cerebral perfusion in 53 patients (84%). The median circulatory arrest time was 59 minutes, ranging from 0 to 83 minutes. The median total cardiopulmonary bypass time was 194 minutes, ranging from 128 to 348 minutes.
643
TABLE 1 Clinical Characteristics of 63 Patients Undergoing Acute Type A Aortic Dissection Repair
Variables Male gender Median age in years (range) Comorbidities Diabetes mellitus Hypertension Coronary artery disease Malignancy End-stage renal disease requiring dialysis Marfan syndrome Bicuspid aortic valve Shock before operation Cardiac tamponade with shock Acute myocardial infarction Severe aortic regurgitation Cardiopulmonary resuscitation DeBakey type DeBakey type I DeBakey type II Operative procedure Bentall operation Hemiarch replacement Aortic valve replacement Redo operation Coronary artery bypass grafting Arterial cannulation Femoral artery Right subclavian artery Methods of brain protection Aortic cross-clamp Deep hypothermic circulatory arrest alone Retrograde cerebral perfusion Selective antegrade cerebral perfusion Operative mortality Postoperative complications Reoperation for bleeding Nosocomial infection Transient delirium Permanent stroke Tracheostomy for respiratory failure
Patient Number (n = 63) 40 (63%) 61 (27 to 90) 8 (13%) 45 (71%) 4 (6%) 6 (10%) 1 (2%) 5 (8%) 3 (5%) 21 (33%) 19 (30%) 1 (2%) 1 (2%) 2 (3%) 57 (90%) 6 (10%) 4 5 1 2 1
(6%) (8%) (2%) (3%) (2%)
8 (13%) 55 (87%) 1 (2%) 1 (2%) 8 (13%) 53 (84%) 4 (6%) 10 4 5 5 3
(16%) (6%) (8%) (6%) (5%)
Replacement of the ascending aorta with aortic valve preservation was performed in 58 cases (92%). Hemiarch replacement was performed in five cases (8%). Composite graft replacement of the aortic valve, sinuses, and ascending aorta with a modified Bentall technique was performed in four cases (6%). Separate replacement of the supracoronary ascending aorta and aortic valve was performed in one case (1.6%). Associated surgical procedures included coronary artery bypass in one patient.
Survival A total of four patients died during hospitalization or within 30 days after the initial repair procedure, yielding an in-hospital mortality of 6%. There was no intraoperative mortality. The causes of hospital mortality were persistent shock, pneumonia, coronary
644
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
insufficiency, and profound stroke each in one patient. Major postoperative complications included reoperation for bleeding in 10 (16%), nosocomial infection in four (6%), transient delirium in five (8%), permanent stroke in five (6%), and tracheostomy for respiratory failure in three patients (5%). During reoperation, the source of bleeding was identified in three of 10 patients. These include a proximal aortic anastomosis, distal aortic anastomosis, and coronary artery button anastomosis each in one patient. For operative survivors, the median follow-up duration was 73 months (range, 1 to 124). Follow-up was 98% complete (62 of 63 patients). One patient was lost to follow-up one month after the initial AAD repair. The causes of late mortality included malignancy in two patients, pneumonia or sepsis in six patients, intracranial hemorrhage in one patient, and coronary artery disease in one patient. All were not related to aortic dissection. Overall survival estimates after the initial operation were 94 3%, 84 5%, and 59 11% at 1, 5, and 10 years, respectively (Fig. 2). Reoperation Of the 59 operative survivors, one patient (1.7%) discharged after the initial AAD repair required reoperation on the proximal aorta. This 53-year-old male with an aneurysm of the aortic root underwent selected sinus repair (resection of the dilated right and noncoronary sinus, and reimplantation of the right coronary artery button) 44 months after the initial AAD repair.20 None of the patients had severe aortic regurgitation or pseudoaneurysm formation at the aortic graft suture lines. Actuarial freedom from reoperation at 1, 5, and 10 years was 100%, 97%, and 97%, respectively (Fig. 3). Fifty-one patients (86%) had follow-up echocardiography and 50 patients (85%) had follow-up computed tomography. The status of the aortic valve function was moderate regurgitation in 14 patients, mild regurgitation in 22 patients, and no regurgitation in 15 patients. For the distal aorta, there was residual dissection with patent false lumen in 26 patients (53%). The relationship of DeBakey type of AAD and residual dissection
J CARD SURG 2014;29:641–646
Figure 3. Freedom from reoperation after repair of acute type A aortic dissection.
after AAD repair was shown in Table 2. No residual dissection was present in six patients with an initial DeBakey type II AAD. Patients who initially had intramural hematoma experienced a lower rate of residual dissection compared with a patient with patent false lumen. The median aortic diameter of the residual dissection on follow-up computed tomography was 4.7 cm, ranging from 2.9 to 5.9 cm, except in a Marfan patient who developed an acute type B aortic dissection one year after the initial AAD repair. Because of severe scoliosis and respiratory compromise, he declined reoperation despite a large aneurysm (6.7 cm). He is still alive at a follow-up of 46 months. DISCUSSION The reported operative mortality rate of AAD repair is 15% to 30%.1 Late reoperations after AAD repair are relatively common with an average reoperation rate of 5.4% to 18%.2–14 Among them, pseudoaneurysm was responsible for 14% to 59% of late reoperations.2,3,6–10 In this study, we used a reinforced sandwich technique for conventional AAD repair. The rates of late reoperation and pseudoaneurysm formation were comparatively low.2–14 Reoperation
Figure 2. Kaplan–Meier survival after repair of acute type A aortic dissection.
Several mechanisms are responsible for late reoperation after AAD repair: (1) progressive aortic valve regurgitation, (2) progressive dilatation of a preserved aortic root and distal aorta, and (3) pseudoaneurysm formation.2–14 Marfan syndrome, younger age, DeBakey type I dissection, preservation of the aortic root, a patent false aortic lumen, and the use of GRF glue and aortic valve resuspension are major risk factors.2–14 However, controversy remains about the optimal surgical techniques for AAD repair. The available surgical techniques included aortic valve resuspension, aortic stump reconstitution with felt bolsters (sandwich technique) and biologic glues, open aortic anastomosis, and extensive aortic root and arch repair. Several types of sandwich techniques have also been introduced, but
J CARD SURG 2014;29:641–646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
645
TABLE 2 Results of Follow-Up Computed Tomography in 50 Patients Undergoing Acute Type A Aortic Dissection Repair False Lumen Patency at Thoracic Aorta Types of Aortic Dissection DeBakey type I (n ¼ 44) Intramural hematoma at ascending aorta (N ¼ 8) Patent false lumen at ascending aorta (n ¼ 36) DeBakey type II (n ¼ 6)
which method is best for the prevention of pseudoaneurysm formation is unknown.2–19 In AAD repair, extensive repair with total root replacement is indicated in cases of root dilatation, extensive root destruction, in the presence of an intimal tear located in the aortic root or coronary ostia, or in Marfan syndrome. In most cases, reapproximation of the aortic layers and restoration of the sinotubular junction can be effective even in the presence of severe aortic regurgitation. However, a more extensive procedure consisting of both aortic root replacement and arch replacement increases operative mortality.6 In this study, we adopted a less extensive procedure with preservation of the aortic root and aortic arch if the entry of the dissection was located in the ascending aorta. The dissected aorta was resected down to the sinotubular junction immediately above the coronary artery ostia. The geometry of the sinotubular junction was restored with a reinforced sandwich technique. The operative mortality rate was low. And the midterm follow-up result was satisfactory with no late severe aortic regurgitation, no late pseudoaneurysm formation and a low reoperation rate. In this study, we used Hemashield strips to make a sandwich. It probably makes no difference what material (teflon felt or bovine pericardium) is used. In addition, the rate of reoperation for bleeding was 16%. Whether this technique could reduce bleeding at the anastomotic site is unknown. Pseudoaneurysm Suture line dehiscence or anastomostic pseudoaneurysm was mostly located at the distal suture line of the aortic prosthetic graft, but also seen proximally.2,3,6– 10 Several surgical techniques of aortic anastomoses affect the development of anastomotic pseudoaneurysm. Aortic cross-clamp, no felt bolsters, and use of biologic glues are major risk factors. Compared to aortic cross-clamp, the advantage of an open distal anastomosis is to prevent clamp injuries on a fragile aorta.14 In addition, there are concerns about the complications related to the use of biologic glue.2–14 Gelatin– resorcinol–formaldehyde glue was associated with an increased risk of reoperation, and 64% was related to pseudoaneurysm formation.2 Furthermore, there is a possibility of aortic wall necrosis induced by intraoperative use of biologic glues.21,22 Not all glues are the same. Other glues, such as Bioglue, have been safe. In
Patent 27 2 25 0
Thrombosed
(61%) (25%) (69%) (0%)
17 6 11 6
(39%) (75%) (31%) (100%)
this study, we used open distal anastomosis, reinforced sandwich technique, no aortic valve resuspension, and no use of biologic glue for false lumen obliteration. All of the above surgical strategies contributed to the low rate of reoperation in our patients. Study limitations Several limitations of our study should be recognized. First, this was a single center study. The study population was small and the follow-up duration was short. Second, the occurrence of reoperation was infrequent, and hence, we could not properly examine the risk factors of reoperation. Third, the echocardiographic and computed tomographic follow-up was incomplete in 14% and 15% of our patients. The incidence of patients requiring reoperation could be underestimated. However, this study was a single surgeon’s experience using a uniform surgical technique. Thus the effect of surgeon and technique factors was not present.
REFERENCES 1. Kruger @ T, Conzelmann LO, Bonser RS, et al: Acute aortic dissection type A. Br J Surg 2012;99:1331–1344. 2. Olsson C, Hillebrant CG, Liska J, et al: Mortality and reoperations in survivors operated on for acute type A aortic dissection and implications for catheter-based or hybrid interventions. J Vasc Surg 2013;58:333–339. 3. Malvindi PG, van Putte BP, Sonker U, et al: Reoperation after acute type a aortic dissection repair: A series of 104 patients. Ann Thorac Surg 2013;95:922–927. 4. Bekkers JA, Raap GB, Takkenberg JJ, et al: Acute type A aortic dissection: Long-term results and reoperations. Eur J Cardiothorac Surg 2013;43:389–396. 5. Rylski B, Beyersdorf F, Blanke P, et al: Supracoronary ascending aortic replacement in patients with acute aortic dissection type A: What happens to the aortic root in the long run? J Thorac Cardiovasc Surg 2013;146:285–290. G, Casali G, Santaniello E, et al: Reoperation 6. Concistre after surgical correction of acute type A aortic dissection: Risk factor analysis. Ann Thorac Surg 2012;93:450–455. 7. Kobuch R, Hilker M, Rupprecht L, et al: Late reoperations after repaired acute type A aortic dissection. J Thorac Cardiovasc Surg 2012;144:300–307. 8. Nishi H, Mitsuno M, Tanaka H, et al: Late reoperations after repair of acute type A aortic dissection. J Card Surg 2010;25:208–213.
646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
9. Tanaka H, Okada K, Kawanishi Y, et al: Clinical significance of anastomotic leak in ascending aortic replacement for acute aortic dissection. Interact Cardiovasc Thorac Surg 2009;9:209–212. 10. Zierer A, Voeller RK, Hill KE, et al: Aortic enlargement and late reoperation after repair of acute type A aortic dissection. Ann Thorac Surg 2007;84:479–486. 11. Geirsson A, Bavaria JE, Swarr D, et al: Fate of the residual distal and proximal aorta after acute type a dissection repair using a contemporary surgical reconstruction algorithm. Ann Thorac Surg 2007;84:1955–1964. 12. Gariboldi V, Grisoli D, Kerbaul F, et al: Long-term outcomes after repaired acute type A aortic dissections. Interact Cardiovasc Thorac Surg 2007;6:47–51. 13. Kazui T, Washiyama N, Bashar AH, et al: Surgical outcome of acute type A aortic dissection: Analysis of risk factors. Ann Thorac Surg 2002;74:75–81. 14. Kirsch M, Soustelle C, Houe¨l R, et al: Risk factor analysis for proximal and distal reoperations after surgery for acute type A aortic dissection. J Thorac Cardiovasc Surg 2002; 123:318–325. 15. Kan CD, Yang YJ: Double telescopic anastomosis with interrupted suture technique in acute aortic dissection. Ann Thorac Surg 2011;91:1630–1631.
J CARD SURG 2014;29:641–646
16. Chen LW, Dai XF, Wu XJ: A modified technique of anastomosis between dissected aortic stump and a Dacron tube graft: The suture line inclusion technique. Ann Thorac Surg 2010;89:1685–1687. 17. Ohata T, Miyamoto Y, Mitsuno M, et al: Modified sandwich technique for acute aortic dissection. Asian Cardiovasc Thorac Ann 2007;15:261–263. 18. Rignano A, Keller GC, Carmo M, et al: A new approach for proximal anastomosis in type ‘‘A’’ acute aortic dissection: Prosthesis eversion. Ann Thorac Surg 2003;76:949–951. 19. Fleck TM, Tschernich H, Grabenwoger M, et al: A double patch sandwich technique for surgical repair of acute aortic dissection type A. Ann Thorac Surg 2003;76:499–502. 20. Urbanski PP, Hijazi H, Dinstak W, et al: Valve-sparing aortic root repair in acute type A dissection: How many sinuses have to be repaired for curative surgery? Eur J Cardiothorac Surg 2013;44:439–444. 21. Hata H, Takano H, Matsumiya G, et al: Late complications of gelatin–resorcin–formalin glue in the repair of acute type A aortic dissection. Ann Thorac Surg 2007;83:1621–1626. 22. Kazui T, Washiyama N, Bashar AH, et al: Role of biologic glue repair of proximal aortic dissection in the development of early and midterm redissection of the aortic root. Ann Thorac Surg 2001;72:509–514.
641
DISEASES OF THE AORTA ORIGINAL ARTICLE _____________________________________________________________
Low Incidence of Late Pseudoaneurysm and Reoperation After Conventional Repair of Acute Type A Aortic Dissection Ron-Bin Hsu, M.D., and Jeng-Wei Chen, M.D. Department of Surgery, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan, ROC ABSTRACT Background and Aim: Suture line dehiscence and pseudoaneurysm formation is one of the leading causes of late reoperation after surgical repair of acute type A aortic dissection (AAD). A sandwich technique can affect the need of reoperation. We sought to assess the late outcomes (mortality and reoperation) of a modified reinforced sandwich technique in conventional AAD repair. Methods: Retrospective review of 63 consecutive patients undergoing AAD repair between 2003 and 2013. Aortic anastomosis was performed with a modified reinforced sandwich technique using Hemashield strips and two-layer polypropylene continuous and interrupted mattress sutures. Results: Marfan syndrome was diagnosed in five (8%) and bicuspid aortic valve in three patients (5%). Twenty-one patients (33%) had preoperative cardiogenic shock necessitating inotropic support. Replacement of the ascending aorta with aortic valve preservation was performed in 58 (92%) and hemiarch replacement in five patients (8%). Four patients died during initial hospitalization, yielding a hospital mortality of 6%. Median follow-up duration was 73 months (range, 1–124). Kaplan–Meier survival rates were 94 W 3%, 84 W 5%, and 59 W 11% at 1, 5, and 10 years. One patient (1.7%) required proximal reoperation 44 months after AAD repair because of progressive dilatation of the aortic root. No patient had severe aortic regurgitation or pseudoaneurysm after AAD repair. Actuarial freedom from reoperation at 1, 5, and 10 years was 100%, 97%, and 97%. Conclusions: A reinforced sandwich technique was a good technique resulting in a low incidence of late reoperation and pseudoaneurysm formation. doi:
10.1111/jocs.12359 (J Card Surg 2014;29:641–646) Acute type A aortic dissection (AAD) is a surgical emergency having a very poor prognosis.1 The goals of surgical treatment in AAD are to prevent death from aortic rupture, to reestablish blood flow in organs with malperfusion, and to correct severe aortic valvular regurgitation if it is present.1 However, operations do not remove the entire dissected aorta. Most operative survivors have a persistent, dissected residual aorta, often with a patent false lumen. Thus the need for late reoperation remains fairly common after AAD repair.2–14 Several retrospective studies have reported the incidence and mechanisms of late reoperation.2–14 However, the results were usually affected by use of
several different surgical techniques2–14 and involvement of many cardiac surgeons.10 Suture line dehiscence and pseudoaneurysm formation is one of the leading causes of late reoperation after AAD repair.2,3,6–10 Several surgical techniques have been introduced for a more secure aortic anastomosis. However, the long-term efficiency of each sandwich technique is unknown.15–19 Controversy still remains about whether different surgical techniques can affect the need for late reoperation.2–19 A sandwich technique should prevent suture line dehiscence and pseudoaneurysm formation.9 We sought to assess the late outcomes (mortality and reoperation) of a modified reinforced sandwich technique after conventional AAD repair.
Conflict of interest: The authors acknowledge no conflict of interest in the submission.
MATERIALS AND METHODS
Funding: None.
Patients
Address for correspondence: Dr. Ron-Bin Hsu, M.D., National Taiwan University Hospital, No. 7, Chung-Shan S. Rd., Taipei, Taiwan 100, ROC. Fax: þ886-2-2341-0933; e-mail: [email protected]
This was a retrospective, observational cohort study of prospectively collected data. We included all
642
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
J CARD SURG 2014;29:641–646
consecutive patients who have undergone AAD repair by a single surgeon (R.B. Hsu) between September 2003 and May 2013 at the National Taiwan University Hospital. Patients with iatrogenic AAD were excluded. All data were collected by retrospective chart review. The local institutional medical ethics committee approved the study and waived the need for informed consent. Operation All operations were performed on an emergency basis. Standard procedures were used for femoral or subclavian artery cannulation, median sternotomy, and bicaval venous cannulation for total cardiopulmonary bypass. Once adequate core hypothermia was achieved, the ascending aorta was opened and the heart was arrested by continuous retrograde crystalloid cardioplegia. Before September 2004, deep hypothermic circulatory arrest (18 8C) was used for the distal aortic anastomosis with an open technique in patients undergoing retrograde cerebral perfusion. After September 2004, moderate hypothermic circulatory arrest (25 8C) was used for the distal anastomosis with an open technique in patients undergoing selective antegrade cerebral perfusion through an 8 mm tube graft connected to the right subclavian artery. The aortic repair was performed with several techniques. The open distal anastomosis was performed first (arch first). The ascending aorta containing the area of most severe injury and intimal tear, if present, was resected and replaced with a Hemashield tubular prosthetic graft (Hemashield, Boston Scientific, Oakland, NJ, USA). If the intimal tear in the aorta was localized to the ascending aorta, the distal anastomosis was done just proximal to the innominate artery. If the intimal tear originated in the aortic arch, aortic replacement was in most cases extended to include excision of the segment of aortic arch containing the intimal tear (hemiarch replacement). After completion of the distal anastomosis, the arterial cannula was shifted to the prosthetic graft and cardiopulmonary bypass was reconstituted. Replacement of the ascending aorta with aortic valve preservation was performed. The geometry of the sinotubular junction was restored without aortic valve resuspension. Composite graft replacement of the aortic valve, sinuses, and ascending aorta with a modified Bentall technique was performed in patients with severe root dilatation. Separate replacement of the supracoronary ascending aorta and aortic valve was performed in patients with aortic valve calcification. Reinforced sandwich technique Two Hemashield strips, which were harvested from the tube graft and placed outside the adventitia and inside the intima, were used for the sandwich technique. The sandwich anastomosis was reinforced with two layers of 4/0 polypropylene continuous and interrupted mattress sutures (Fig. 1). The first layer was a 4/0 polypropylene continuous running suture.
Figure 1. Scheme of a reinforced sandwich technique in open aortic anastomosis: Two Hemashield strips and two layers of 4/0 polypropylene continuous vertical and interrupted horizontal mattress sutures for reinforcement.
Multiple 4/0 polypropylene interrupted nonpledgetted horizontal mattress sutures were passed from the inside to the outside of the sandwich anastomosis. This reinforced aorta was then sutured end-to-end to a Hemashield tubular graft using a 4/0 polypropylene continuous suture. No gelatin-resorcinol-formaldehyde glue or other biologic glue was put inside the false lumen to reconstitute the dissected aortic layers. Tissucol duo fibrin glue (Tissucol duo, Baxter AG, Vienna, Austria) was used as a sealant on the external suture line. Definition Dissection was considered acute when treated within 14 days from the onset of symptoms. Pseudoaneurysm was defined as a localized hematoma or false aneurysm at the proximal or distal suture lines.2,3,6–10 Reoperation included any cardiac or vascular surgical intervention with an indication that could be related to complications occurring in the segments of aorta involved in the initial repair of AAD or to complications stemming from the spontaneous evolution of the dissection at sites not treated in the initial operation. The indications for reoperation included aortic diameter exceeding 60 mm, rapid progression of an aortic aneurysm exceeding 1 cm per year, and presence of pseudoaneurysm or severe aortic regurgitation.2–14 Reoperation was defined as a proximal reoperation if the operation involved the aorta before the replaced graft and distal reoperation if the operation involved the aorta immediately after the replaced graft.
J CARD SURG 2014;29:641–646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
Follow-up Follow-up was conducted by telephone interviews with surviving patients or with patients’ families, or by direct examination at the outpatient clinic from July to September 2013. Information included the patients’ functional status and the results of follow-up echocardiographic and computed tomographic scans. The severity of aortic regurgitation was evaluated with serial echocardiograpy every year. Aortic size was evaluated with serial computed tomographic scans every year. Statistical analysis Categorical variables were reported as the percentage of patients in the subgroup. Continuous variables with normal distribution were presented as mean SD; those with skewed distribution were presented as medians with ranges. Survival and time to event (mortality and reoperation) was estimated by the Kaplan–Meier method. RESULTS Patients In the period 2003 to 2013, 63 consecutive patients underwent emergency surgical repair of AAD. There were 40 males and 23 females with the median age of 61 years (range, 27 to 90). AAD was DeBakey type I in 57 (90%) and type II in six patients (10%). Intramural hematoma of the ascending aorta was present in 10 (17%) of 57 DeBakey type I AADs. Baseline patient characteristics and predisposing factors for aortic dissection are shown in Table 1. The most common underlying medical disorders were hypertension in 45 patients (71%), and diabetes mellitus in eight patients (13%). Marfan syndrome was diagnosed in five patients (8%). Bicuspid aortic valve was present in three patients (5%). Twenty-one patients (33%) had preoperative cardiogenic shock necessitating at least temporary inotropic support. Cardiac tamponade and shock was present in 19 patients (30%) before operation. Clinically significant malperfusion syndrome was noted in two patients (3%). The affected end organs were the myocardium (acute myocardial infarction) and carotid artery (stroke) each in one patient. Femoral artery cannulation was used in the initial eight patients and subclavian artery cannulation was used in 55 patients. The methods of brain protection included aortic cross-clamping without hypothermic circulatory arrest in a Marfan patient with Debakey type II AAD, deep hypothermic circulatory arrest alone in one redo Bentall patient, deep hypothermic circulatory arrest with additional retrograde cerebral perfusion in eight patients, and moderate hypothermic circulatory arrest with additional selective antegrade cerebral perfusion in 53 patients (84%). The median circulatory arrest time was 59 minutes, ranging from 0 to 83 minutes. The median total cardiopulmonary bypass time was 194 minutes, ranging from 128 to 348 minutes.
643
TABLE 1 Clinical Characteristics of 63 Patients Undergoing Acute Type A Aortic Dissection Repair
Variables Male gender Median age in years (range) Comorbidities Diabetes mellitus Hypertension Coronary artery disease Malignancy End-stage renal disease requiring dialysis Marfan syndrome Bicuspid aortic valve Shock before operation Cardiac tamponade with shock Acute myocardial infarction Severe aortic regurgitation Cardiopulmonary resuscitation DeBakey type DeBakey type I DeBakey type II Operative procedure Bentall operation Hemiarch replacement Aortic valve replacement Redo operation Coronary artery bypass grafting Arterial cannulation Femoral artery Right subclavian artery Methods of brain protection Aortic cross-clamp Deep hypothermic circulatory arrest alone Retrograde cerebral perfusion Selective antegrade cerebral perfusion Operative mortality Postoperative complications Reoperation for bleeding Nosocomial infection Transient delirium Permanent stroke Tracheostomy for respiratory failure
Patient Number (n = 63) 40 (63%) 61 (27 to 90) 8 (13%) 45 (71%) 4 (6%) 6 (10%) 1 (2%) 5 (8%) 3 (5%) 21 (33%) 19 (30%) 1 (2%) 1 (2%) 2 (3%) 57 (90%) 6 (10%) 4 5 1 2 1
(6%) (8%) (2%) (3%) (2%)
8 (13%) 55 (87%) 1 (2%) 1 (2%) 8 (13%) 53 (84%) 4 (6%) 10 4 5 5 3
(16%) (6%) (8%) (6%) (5%)
Replacement of the ascending aorta with aortic valve preservation was performed in 58 cases (92%). Hemiarch replacement was performed in five cases (8%). Composite graft replacement of the aortic valve, sinuses, and ascending aorta with a modified Bentall technique was performed in four cases (6%). Separate replacement of the supracoronary ascending aorta and aortic valve was performed in one case (1.6%). Associated surgical procedures included coronary artery bypass in one patient.
Survival A total of four patients died during hospitalization or within 30 days after the initial repair procedure, yielding an in-hospital mortality of 6%. There was no intraoperative mortality. The causes of hospital mortality were persistent shock, pneumonia, coronary
644
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
insufficiency, and profound stroke each in one patient. Major postoperative complications included reoperation for bleeding in 10 (16%), nosocomial infection in four (6%), transient delirium in five (8%), permanent stroke in five (6%), and tracheostomy for respiratory failure in three patients (5%). During reoperation, the source of bleeding was identified in three of 10 patients. These include a proximal aortic anastomosis, distal aortic anastomosis, and coronary artery button anastomosis each in one patient. For operative survivors, the median follow-up duration was 73 months (range, 1 to 124). Follow-up was 98% complete (62 of 63 patients). One patient was lost to follow-up one month after the initial AAD repair. The causes of late mortality included malignancy in two patients, pneumonia or sepsis in six patients, intracranial hemorrhage in one patient, and coronary artery disease in one patient. All were not related to aortic dissection. Overall survival estimates after the initial operation were 94 3%, 84 5%, and 59 11% at 1, 5, and 10 years, respectively (Fig. 2). Reoperation Of the 59 operative survivors, one patient (1.7%) discharged after the initial AAD repair required reoperation on the proximal aorta. This 53-year-old male with an aneurysm of the aortic root underwent selected sinus repair (resection of the dilated right and noncoronary sinus, and reimplantation of the right coronary artery button) 44 months after the initial AAD repair.20 None of the patients had severe aortic regurgitation or pseudoaneurysm formation at the aortic graft suture lines. Actuarial freedom from reoperation at 1, 5, and 10 years was 100%, 97%, and 97%, respectively (Fig. 3). Fifty-one patients (86%) had follow-up echocardiography and 50 patients (85%) had follow-up computed tomography. The status of the aortic valve function was moderate regurgitation in 14 patients, mild regurgitation in 22 patients, and no regurgitation in 15 patients. For the distal aorta, there was residual dissection with patent false lumen in 26 patients (53%). The relationship of DeBakey type of AAD and residual dissection
J CARD SURG 2014;29:641–646
Figure 3. Freedom from reoperation after repair of acute type A aortic dissection.
after AAD repair was shown in Table 2. No residual dissection was present in six patients with an initial DeBakey type II AAD. Patients who initially had intramural hematoma experienced a lower rate of residual dissection compared with a patient with patent false lumen. The median aortic diameter of the residual dissection on follow-up computed tomography was 4.7 cm, ranging from 2.9 to 5.9 cm, except in a Marfan patient who developed an acute type B aortic dissection one year after the initial AAD repair. Because of severe scoliosis and respiratory compromise, he declined reoperation despite a large aneurysm (6.7 cm). He is still alive at a follow-up of 46 months. DISCUSSION The reported operative mortality rate of AAD repair is 15% to 30%.1 Late reoperations after AAD repair are relatively common with an average reoperation rate of 5.4% to 18%.2–14 Among them, pseudoaneurysm was responsible for 14% to 59% of late reoperations.2,3,6–10 In this study, we used a reinforced sandwich technique for conventional AAD repair. The rates of late reoperation and pseudoaneurysm formation were comparatively low.2–14 Reoperation
Figure 2. Kaplan–Meier survival after repair of acute type A aortic dissection.
Several mechanisms are responsible for late reoperation after AAD repair: (1) progressive aortic valve regurgitation, (2) progressive dilatation of a preserved aortic root and distal aorta, and (3) pseudoaneurysm formation.2–14 Marfan syndrome, younger age, DeBakey type I dissection, preservation of the aortic root, a patent false aortic lumen, and the use of GRF glue and aortic valve resuspension are major risk factors.2–14 However, controversy remains about the optimal surgical techniques for AAD repair. The available surgical techniques included aortic valve resuspension, aortic stump reconstitution with felt bolsters (sandwich technique) and biologic glues, open aortic anastomosis, and extensive aortic root and arch repair. Several types of sandwich techniques have also been introduced, but
J CARD SURG 2014;29:641–646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
645
TABLE 2 Results of Follow-Up Computed Tomography in 50 Patients Undergoing Acute Type A Aortic Dissection Repair False Lumen Patency at Thoracic Aorta Types of Aortic Dissection DeBakey type I (n ¼ 44) Intramural hematoma at ascending aorta (N ¼ 8) Patent false lumen at ascending aorta (n ¼ 36) DeBakey type II (n ¼ 6)
which method is best for the prevention of pseudoaneurysm formation is unknown.2–19 In AAD repair, extensive repair with total root replacement is indicated in cases of root dilatation, extensive root destruction, in the presence of an intimal tear located in the aortic root or coronary ostia, or in Marfan syndrome. In most cases, reapproximation of the aortic layers and restoration of the sinotubular junction can be effective even in the presence of severe aortic regurgitation. However, a more extensive procedure consisting of both aortic root replacement and arch replacement increases operative mortality.6 In this study, we adopted a less extensive procedure with preservation of the aortic root and aortic arch if the entry of the dissection was located in the ascending aorta. The dissected aorta was resected down to the sinotubular junction immediately above the coronary artery ostia. The geometry of the sinotubular junction was restored with a reinforced sandwich technique. The operative mortality rate was low. And the midterm follow-up result was satisfactory with no late severe aortic regurgitation, no late pseudoaneurysm formation and a low reoperation rate. In this study, we used Hemashield strips to make a sandwich. It probably makes no difference what material (teflon felt or bovine pericardium) is used. In addition, the rate of reoperation for bleeding was 16%. Whether this technique could reduce bleeding at the anastomotic site is unknown. Pseudoaneurysm Suture line dehiscence or anastomostic pseudoaneurysm was mostly located at the distal suture line of the aortic prosthetic graft, but also seen proximally.2,3,6– 10 Several surgical techniques of aortic anastomoses affect the development of anastomotic pseudoaneurysm. Aortic cross-clamp, no felt bolsters, and use of biologic glues are major risk factors. Compared to aortic cross-clamp, the advantage of an open distal anastomosis is to prevent clamp injuries on a fragile aorta.14 In addition, there are concerns about the complications related to the use of biologic glue.2–14 Gelatin– resorcinol–formaldehyde glue was associated with an increased risk of reoperation, and 64% was related to pseudoaneurysm formation.2 Furthermore, there is a possibility of aortic wall necrosis induced by intraoperative use of biologic glues.21,22 Not all glues are the same. Other glues, such as Bioglue, have been safe. In
Patent 27 2 25 0
Thrombosed
(61%) (25%) (69%) (0%)
17 6 11 6
(39%) (75%) (31%) (100%)
this study, we used open distal anastomosis, reinforced sandwich technique, no aortic valve resuspension, and no use of biologic glue for false lumen obliteration. All of the above surgical strategies contributed to the low rate of reoperation in our patients. Study limitations Several limitations of our study should be recognized. First, this was a single center study. The study population was small and the follow-up duration was short. Second, the occurrence of reoperation was infrequent, and hence, we could not properly examine the risk factors of reoperation. Third, the echocardiographic and computed tomographic follow-up was incomplete in 14% and 15% of our patients. The incidence of patients requiring reoperation could be underestimated. However, this study was a single surgeon’s experience using a uniform surgical technique. Thus the effect of surgeon and technique factors was not present.
REFERENCES 1. Kruger @ T, Conzelmann LO, Bonser RS, et al: Acute aortic dissection type A. Br J Surg 2012;99:1331–1344. 2. Olsson C, Hillebrant CG, Liska J, et al: Mortality and reoperations in survivors operated on for acute type A aortic dissection and implications for catheter-based or hybrid interventions. J Vasc Surg 2013;58:333–339. 3. Malvindi PG, van Putte BP, Sonker U, et al: Reoperation after acute type a aortic dissection repair: A series of 104 patients. Ann Thorac Surg 2013;95:922–927. 4. Bekkers JA, Raap GB, Takkenberg JJ, et al: Acute type A aortic dissection: Long-term results and reoperations. Eur J Cardiothorac Surg 2013;43:389–396. 5. Rylski B, Beyersdorf F, Blanke P, et al: Supracoronary ascending aortic replacement in patients with acute aortic dissection type A: What happens to the aortic root in the long run? J Thorac Cardiovasc Surg 2013;146:285–290. G, Casali G, Santaniello E, et al: Reoperation 6. Concistre after surgical correction of acute type A aortic dissection: Risk factor analysis. Ann Thorac Surg 2012;93:450–455. 7. Kobuch R, Hilker M, Rupprecht L, et al: Late reoperations after repaired acute type A aortic dissection. J Thorac Cardiovasc Surg 2012;144:300–307. 8. Nishi H, Mitsuno M, Tanaka H, et al: Late reoperations after repair of acute type A aortic dissection. J Card Surg 2010;25:208–213.
646
HSU AND CHEN REOPERATION AFTER TYPE A AORTIC DISSECTION REPAIR
9. Tanaka H, Okada K, Kawanishi Y, et al: Clinical significance of anastomotic leak in ascending aortic replacement for acute aortic dissection. Interact Cardiovasc Thorac Surg 2009;9:209–212. 10. Zierer A, Voeller RK, Hill KE, et al: Aortic enlargement and late reoperation after repair of acute type A aortic dissection. Ann Thorac Surg 2007;84:479–486. 11. Geirsson A, Bavaria JE, Swarr D, et al: Fate of the residual distal and proximal aorta after acute type a dissection repair using a contemporary surgical reconstruction algorithm. Ann Thorac Surg 2007;84:1955–1964. 12. Gariboldi V, Grisoli D, Kerbaul F, et al: Long-term outcomes after repaired acute type A aortic dissections. Interact Cardiovasc Thorac Surg 2007;6:47–51. 13. Kazui T, Washiyama N, Bashar AH, et al: Surgical outcome of acute type A aortic dissection: Analysis of risk factors. Ann Thorac Surg 2002;74:75–81. 14. Kirsch M, Soustelle C, Houe¨l R, et al: Risk factor analysis for proximal and distal reoperations after surgery for acute type A aortic dissection. J Thorac Cardiovasc Surg 2002; 123:318–325. 15. Kan CD, Yang YJ: Double telescopic anastomosis with interrupted suture technique in acute aortic dissection. Ann Thorac Surg 2011;91:1630–1631.
J CARD SURG 2014;29:641–646
16. Chen LW, Dai XF, Wu XJ: A modified technique of anastomosis between dissected aortic stump and a Dacron tube graft: The suture line inclusion technique. Ann Thorac Surg 2010;89:1685–1687. 17. Ohata T, Miyamoto Y, Mitsuno M, et al: Modified sandwich technique for acute aortic dissection. Asian Cardiovasc Thorac Ann 2007;15:261–263. 18. Rignano A, Keller GC, Carmo M, et al: A new approach for proximal anastomosis in type ‘‘A’’ acute aortic dissection: Prosthesis eversion. Ann Thorac Surg 2003;76:949–951. 19. Fleck TM, Tschernich H, Grabenwoger M, et al: A double patch sandwich technique for surgical repair of acute aortic dissection type A. Ann Thorac Surg 2003;76:499–502. 20. Urbanski PP, Hijazi H, Dinstak W, et al: Valve-sparing aortic root repair in acute type A dissection: How many sinuses have to be repaired for curative surgery? Eur J Cardiothorac Surg 2013;44:439–444. 21. Hata H, Takano H, Matsumiya G, et al: Late complications of gelatin–resorcin–formalin glue in the repair of acute type A aortic dissection. Ann Thorac Surg 2007;83:1621–1626. 22. Kazui T, Washiyama N, Bashar AH, et al: Role of biologic glue repair of proximal aortic dissection in the development of early and midterm redissection of the aortic root. Ann Thorac Surg 2001;72:509–514.
