| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;83:1603-1609
© 2007 The Society of Thoracic Surgeons
Department of Cardiothoracic and Vascular Surgery, University of Texas-Houston Medical School, Memorial Hermann Heart and Vascular Institute, Houston, Texas
Accepted for publication January 15, 2007.
* Address correspondence to Dr Estrera, Department of Cardiothoracic and Vascular Surgery, University of Texas-Houston Medical School, 6410 Fannin St, Suite 450, Houston, TX 77030 (Email: anthony.l.estrera{at}uth.tmc.edu).
Presented at the Fifty-third Annual Meeting of the Southern Thoracic Surgical Association, Tucson, AZ, Nov 8–11, 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Methods: Between January 1991 and March 2006, 63 patients underwent reoperation after previous repair for acute type A aortic dissection. Procedures performed at reoperation included ascending (94%, 59 of 63), total arch (62%, 39 of 63), elephant trunk (56%, 35 of 63), aortic valve replacement (38%, 24 of 63), aortic root (27%, 17 of 63), and coronary artery bypass graft (8%, 5 of 63). Preoperative, operative, and postoperative variables were analyzed retrospectively with regard to early and late mortality.
Results: Thirty-day mortality was 11.1% (7 of 63). No strokes occurred. Incidence of renal failure, respiratory failure, and bleeding was 6% (4 of 63), 23% (15 of 63), and 6% (4 of 63), respectively. Mean time from initial repair to reoperation was 69 months (range, 1 to 258). Procedure performed (root versus ascending/resuspension) at initial repair did not affect the time to reoperation (p > 0.05). Median follow-up was 40 months; and 1-, 5-, and 10-year survival was 82%, 74%, and 62%, respectively. Multivariate predictors of late mortality were prior coronary artery bypass graft surgery (odds ratio = 6.5, p < 0.003), bypass time (odds ratio = 3.6, p < 0.02), and renal dysfunction (odds ratio = 3.7, p < 0.05).
Conclusions: Proximal reoperations for repaired acute type A aortic dissection can be performed with acceptable early and late mortality. The concern for proximal reoperation should not dictate the initial procedure choice during acute type A aortic dissection. Continued clinical and radiographic surveillance of repaired type A aortic dissection is warranted.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
If patients survive the primary repair, late complications requiring reoperations may still occur in as many as 40% of patients [6–8]. These are often related to aneurysmal progression either proximal (aortic root) or distal (transverse arch) to the primary repair, valvular heart insufficiency secondary to degeneration of the native valve, progressive atherosclerotic coronary artery disease, and infection. Concerned with the associated risks of proximal reoperation, some have proposed an aggressive approach of aortic root replacement during emergent repair of acute type A aortic dissection [9]. Because few data exist regarding late reoperations, we report outcomes of proximal reoperation after repaired type A aortic dissection.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Between January 1991 and March 2006, 63 patients were referred for reoperation after previous repair for acute type A aortic dissection. Fifty-nine of 63 patients (94%) were referred from outside our institution. We had performed the primary repair in 4 patients of our entire experience of 280 acute type A dissection repairs during the study period. Mean patient age was 58 years and 78% of patients (49 of 63) were men. Eleven patients (18%) had a history of Marfan’s syndrome. Table 1 lists patient comorbidities. Seventeen patients (27%) had previously undergone two or more median sternotomies with 1 patient having undergone three and 2 patients having undergone four previous sternotomies. Procedures performed at the initial type A aortic dissection repair and before the reoperation are listed on Table 2. Details from the previous procedures were extracted from operative reports and direct communications with the referring physicians. Procedures previously reported were confirmed at the time of reoperation.
|
|
Operative Procedure
Standard anesthetic management was performed with endotracheal intubation as well as arterial and pulmonary arterial line monitoring. Aprotonin was administered. The chest was entered by a redo-sternotomy median sternotomy as previously described [10]. After full intravenous anticoagulation, cannulation was performed. If feasible, direct cannulation of the previous ascending aortic graft was performed This was accomplished in 63% of cases (40 of 63). If not, then femoral or axillary cannulation was utilized. Irrespective of the cannulation technique, cerebral monitoring remained integral to confirming adequate cerebral perfusion [11]. Initial cannulation and cardiopulmonary bypass with circulatory arrest was performed in cases of false aneurysm (infection), and known rupture.
A 10-lead electroencephalogram (EEG) monitored cerebral function, a near-infrared spectroscopy unit continuously monitored frontal cerebral oxygenation, and power M-mode transcranial Doppler ultrasonography monitored middle cerebral artery blood flow. Once the EEG was isoelectric, which generally coincided with a nasopharyngeal temperature of 15°C to 20°C, cardiopulmonary bypass was discontinued and circulation was arrested. Retrograde cerebral perfusion was begun through the superior vena cava cannula and guided using power M-mode transcranial Doppler ultrasonography [12].
After completion of the distal arch reconstruction, retrograde cerebral perfusion was discontinued, and a cannula was placed into the commercially available side-arm aortic graft. Procedures performed at reoperation are listed in Table 3. In the cases of infected aortic graft, homograft or prosthetic graft was used with omental wrapping performed in both cases.
|
The median retrograde cerebral perfusion time was 38 minutes (range, 4 to 89); median cardiopulmonary bypass time of 126 minutes (range, 101 to 236); and median aortic cross-clamp time was 99 minutes (range, 22 to 218).
Outcome Variables
Preoperative factors analyzed included age, sex, emergency status, diabetes mellitus, hypertension, and renal insufficiency, Marfan’s syndrome, chronic obstructive pulmonary disease, coronary artery disease, congestive heart failure, peripheral arterial disease, cerebrovascular disease, and smoking. Operative factors analyzed were hypothermic circulatory arrest, retrograde cerebral perfusion, aortic cross-clamp time, and cardiopulmonary bypass time. Emergency status was defined as presenting with rupture or leak leading to hypotension or shock. Chronic obstructive pulmonary disease was defined by a history of chronic bronchitis and emphysema, or, while on bronchodilators, less than 60% of predicted forced expired volume in one second (FEV1). A serum creatinine level of greater than 2.0 mg/dL or the need for dialysis defined renal dysfunction. Respiratory failure was defined as prolonged ventilation (more than 3 days) or need for tracheostomy.
Data Analysis
Data were collected prospectively and analysis was retrospective. Survival was ascertained by direct patient contact (telephone or letter) and by searching the Social Security Death Index. Data were collected from chart reviews by a trained nurse abstractor and were entered into a dedicated database. Follow-up was complete in all patients, and patients were followed until death or until follow-up reached the study end date (March 31, 2006). Data were maintained in the database with encrypted patient identifiers and stored on a secure server in a dedicated cold room. The SAS version 9.1.3 service pack 4 (SAS Institute, Cary, North Carolina) was used for analysis. Surviving patients were right-censored from the denominator when their maximum follow-up time was reached or on the date they were lost to follow-up. Univariate dichotomous data were analyzed by contingency table methods. Univariate continuous data were analyzed by logistic regression and also were divided into quartiles and analyzed by contingency table methods. Long-term survival was assessed by univariate and multiple Cox proportional hazards regression. The null hypothesis was rejected at p less than 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Description of Previous Operations
Ten patients (16%) had a previous composite root replacement (Bentall), and 14 patients (22%) had an isolated aortic valve replacement (see Table 2). Of the 10 patients who had a previous aortic root replacement, 4 of 10 (40%) required re-replacement of the root for infection (2 cases) or for false aneurysm (2 cases). Of those patients who had undergone tubular portion of ascending with isolated aortic valve replacement, 50% (7 of 14) required composite aortic root replacement. Of the initial cohort, 39 patients had previous aortic valve resuspension. Of this group, 26 patients maintained integrity of the previous valve resuspension and did not require replacement. Four patients (6%) required the entire proximal aorta replaced, namely, total transverse aortic arch (elephant trunk) with a composite aortic root replacement.
Whether profound hypothermic circulatory arrest with open distal anastomosis or simple ascending aortic cross-clamping was used at the initial operation could be ascertained in 56 of 63 cases. Profound hypothermic circulatory arrest was performed in 30% of cases (17 of 56) and clamping of the dissected aorta in 70% (39 of 56). Neither the use of profound hypothermic circulatory arrest nor simple aortic cross-clamping was associated with earlier time to proximal reoperation or the need of a distal aortic procedure (p > 0.05).
Risk Factors
The only univariate risk factor for early death was later return for reoperation (p < 0.008). Table 4
lists the univariate analysis of selected preoperative and operative factors for 30-day mortality. Other factors analyzed but not significantly associated with early death were emergency status, chronic obstructive pulmonary disease, coronary artery disease, stroke, diabetes mellitus, hypertension, smoking, aortic cross-clamp time, circulatory arrest time, time from previous repair, transfusion requirements, and postoperative complications of bleeding, renal insufficiency, respiratory failure, myocardial infarction.
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In this study, early mortality was 11%, which is consistent with previous reports involving proximal aortic reoperations, which ranged from 5% to 28% [17, 19–21]. As such, a simplified approach of ascending aortic graft replacement with aortic valve resuspension in the setting of acute type A aortic dissection when technically feasible is justified. Acknowledging that most of these patients often present in serious condition at best, if not in extremis, it behooves the operating surgeon to perform the most expeditious procedure to ultimately achieve patient survival.
In this study, the only risk factor for early death was the time from the initial type A dissection repair to the late proximal reoperation. Patients who presented beyond 10 years from the primary type A dissection repair were at greater risk of early death (p < 0.008). Although increasing age approached significance (p = 0.06), increasing time period between presentations may have simply been a reflection of age. That no other factors were significant was unexpected.
We attempted to analyze other operative factors performed at the initial repair to determine any correlation with the requirement of reoperation. Many differing surgical approaches have been recommended during repairs of acute type A aortic dissection to prevent either early death or late reoperations. Such approaches have included profound hypothermic circulatory arrest for open distal anastomosis (as opposed to ascending aortic cross-clamping), composite root replacement (as opposed to isolated ascending aortic replacement), and complete transverse arch replacement (as opposed to complete or proximal transverse arch replacement). Although the significance of these approaches cannot be determined in a retrospective manner, it was noteworthy that these approaches performed at the initial type A dissection repair were also not immune from reoperation. Thirty percent of the patients in this series had undergone previous profound hypothermic circulatory arrest with open distal anastomosis. Despite open distal anastomosis, there was no difference in the incidence late distal aortic reconstructions (descending thoracic aortic aneurysms or thoracoabdominal aortic aneurysms) when compared with those who had undergone ascending aortic cross-clamping. Moreover, of the 10 patients who had undergone a composite aortic root replacement at the initial repair, 4 (40%) required re-replacement of the composite aortic root owing to either infection or false aneurysm. Finally, the two prior transverse arch replacements required multibranched arch reconstruction for great vessel island patch enlargement. It remains apparent that regardless of the approach undertaken, failures related to technique or more likely natural progression of the disease will undoubtedly occur in the setting of the acutely dissected aorta.
The need for proximal reoperation is likely related to both patient factors as well as operative technique. Based on reports by others as well as our experience from this series, we believe that some of the operative factors that contribute to the need for late reoperations include incomplete excision of the tear, failure to obliterate the false lumen, and proximal redissection [16]. Proximal redissection refers to pressurization of the aortic root from a less than "water-tight" proximal anastomosis in the setting of a supracoronary ascending graft replacement. If the aortic root is nondilated and is to be preserved, then it is of paramount importance to prevent proximal redissection [22]. Proximal redissection may increase the risk of late aortic insufficiency, late dilatation of the sinuses of Valsalva, and late ostial coronary stenosis.
In cases of reoperation for the aortic root, coronary artery reattachment remains a significant concern. Many options are available, but we prefer direct reattachment if a coronary button can be created easily. In reoperations, however, either the inherent disease process (dissection) or extensive adhesions make direct reattachment difficult. For this reason, we prefer the modified Cabrol reattachment using a polyester interposition graft oftentimes for the left main coronary artery and sometimes the right coronary artery [10, 13]. The use of an interposition graft for reattachment allows for the prevention of tension and ultimate bleeding. Our data, however, do not support the use of one technique over another; it was our preference.
Similar to our previous study on reoperations for the ascending and transverse arch, risk factors for long-term mortality included a history of renal insufficiency and previous history coronary artery bypass grafting, and cardiopulmonary bypass time. Factors not significant in this analysis but previously significant were aortic valve replacement and chronic obstructive pulmonary disease. The differences were likely due to the smaller cohort size represented in this study.
This study should be viewed with certain limitations. It remains a retrospective analysis and as such has inherent limitations and biases. Although we attempted to derive correlations between what was performed at the initial procedure, for example, use of profound hypothermic circulatory arrest versus cross-clamping, because most of these cases were referred by outside physicians, the actual number of patients undergoing the original type A aortic dissection repair by the referring surgeon remains unknown and any definite conclusions regarding prior technique cannot be made. Moreover, more detailed information regarding the initial type A dissection repair such as echocardiographic data, aneurysm size, and status of the tear and dissection flap would have been informative. Although at reoperation we could surmise what may had happened, these were still only descriptive observations.
In summary, proximal reoperations for repaired acute type A aortic dissection can be performed with acceptable early and late mortality. For this reason, the concern for proximal reoperation should not dictate the initial procedure choice during acute type A aortic dissection. Continued clinical and radiographic surveillance of repaired type A aortic dissection is warranted.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
To summarize, the authors reported 63 patients who received reoperations after previous type A repairs between January 1991 and March 2006 for a variety of indications. Eighteen percent of these patients had Marfan syndrome. The mean time from the initial repair to reoperation was about 6 years. The procedure performed at the time of the initial repair did not affect time to reoperation. Thirty-day mortality was 11% with good 1-, 5-, and 10-year survival. Multivariate predictors of late mortality were prior coronary bypass, overall bypass time, and renal dysfunction. Overall, an excellent and well-presented series.
The extent of initial surgical repair for type A aortic dissection, even in the year 2006, remains controversial. The main objective, of course, is preservation of life, and this is done, in simplistic terms, by interrupting the continuity of the dissection in the ascending aorta to avoid proximal extension and the often fatal complications of acute aortic insufficiency, coronary artery dissection with massive myocardial infarction, and ascending aortic rupture with cardiac tamponade.
But what is the best operation to perform in this initial setting? As Dr Estrera points out in his manuscript, the answer to this question depends on many factors, including the acuity of the patient’s situation, the anatomy of the dissection, and the experience and comfort level of the surgeon. Hence, the literature contains all permutations and combinations of repair strategies, ranging from the simplest of repairs to aggressive surgical approaches involving circulatory arrest, aortic arch replacement distally and aortic root replacement proximally at the time of initial operation.
Doctor Estrera’s study suggests that the choice of initial operation has no effect on subsequent requirement for reoperation, supporting the notion that the most expeditious procedure should be performed at the time of initial surgery to maximize patient survival, and in this regard I have three questions.
Of note in your study is that 18% of your patients had Marfan syndrome, and your data analysis shows a very strong odds ratio for Marfan syndrome as a predictor of reoperation. Should this therefore be taken as evidence that the Marfan syndrome population represents a specific patient subset in which an aggressive initial approach is warranted?
Secondly, in your manuscript you indicate that of the 63 patients referred for operation after previous type A repair, 59 of them, or 94%, were referred from outside of your institution. In other words, in this series of reoperations, your institution did the primary repair in only 4 of these patients. How many type A dissection repairs do you perform at your institution per year and what happened to them? Do these data indirectly provide evidence that repair of type A dissections in experienced centers by experienced surgeons such as yourselves will reduce the incidence of primary reoperation?
And finally, I noticed in your series that no patient received a valve-sparing aortic root replacement either at the initial operation or subsequent operation. What is your experience with this procedure at the University of Texas Houston Medical School, and in light of these data, is there a place for this procedure in this patient population?
Finally, I wish to thank the Southern Thoracic Surgical Association for the privilege of being a discussant on this paper.
DR ESTRERA: Thank you, John. I appreciate your comments and your questions. Regarding your first question with Marfan’s, I completely agree with your observation and your analysis of our study. Although it was not significant, it was approaching significance regarding Marfan’s as being a risk factor for early mortality. Hence, this still supports what we do regarding type A dissection repair, in other words, we try to keep it simple. We admit, however, that in patients with Marfan’s, it may be justified to be more aggressive.
Regarding your second question, our experience with type A dissection repair, we do about 30 to 35 type A dissection repairs a year. A lot of that switched when we moved from Methodist over to UT Memorial Hermann Hospital, and hence the majority of our experience has been over the last 6 or 7 years where we have accumulated almost 280 cases. I will admit that 4 patients have required reoperation. There are 2 other patients who I am following who have moderate aortic insufficiency who will likely need reoperation in the future. Despite this, we still try to keep a simplistic, expeditious approach to type A dissection. We try to just replace the ascending under circulatory arrest, open distal anastomosis, reconstruct the aortic root, and resuspend the valve. In our results, we replace the arch in only about 8% to 9% and replace the root or valve in about 8% to 9%. Those are our results.
Again, I want to emphasize that although we do a lot of this, this is an operation that everyone in this audience should be able to do, and the reality is you have got to be comfortable with what you do, and the fact that your patient may need a reoperation shouldn’t dictate what your first operation is going to be.
Regarding your third question, valve-sparing root replacement, we do like that operation, although we do the David reimplantation, and I really give credit to Duke Cameron because he taught us the sort of six-stitch technique, which has really made the procedure more easy to perform. We don’t do it in the reoperations because you have to dissect the aortic root a significant amount to get to that subannular plane. So we just haven’t done it for the reoperations. In the setting of acute dissection, it might be warranted in the patient with Marfan’s, but again, this is a more complicated operation in a difficult situation, and I believe keeping it simple might be the best approach.
DR JOSEPH S. COSELLI (Houston, TX): Tony, that was an excellent presentation, fantastic results. I was particularly struck by your extremely low incidence of stroke. Many authors have implicated either directly or indirectly that it is partly a reflection of how deeply and thoroughly one looks to stroke and the number of postoperative magnetic resonance images of the head, et cetera, which leads me to ask the question about your position on antegrade cerebral perfusion in these cases. It appears you use primarily retrograde cerebral perfusion, which is somewhat different than many of the others that perform similar operations.
And then finally, we too are very aggressive with root replacement in patients with Marfan syndrome, but there is a group of patients with Lowe’s syndrome for whom we face the same conundrum and would be interested in your policy on that particular cohort of patients.
DR ESTRERA: Thank you, Dr Coselli. To answer your second question first, I agree with you completely. Any patient who has a connective tissue disorders, the genetic defects—and as you know, we are part of the SCORR grant at UT Houston in conjunction with your group at Baylor—we are aggressive at replacing as much of the aorta as necessary in these cases.
To address your first question regarding retrograde cerebral perfusion, we have always utilized retrograde cerebral perfusion with pretty good results. We published actually results on how to use retrograde using our cerebral monitoring device at this meeting 2 or 3 years ago, and continue to do it. There was a study, an unpublished survey of patients, that we analyzed antegrade versus retrograde, and the stark reality is that we did not see a difference in neurologic outcome. Hence, we have sort of just reverted to keeping it simple again and using retrograde for all these cases.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  I-M. Chen and C.-C. Shih Extending hybrid approach to residual Stanford type A dissecting aortic aneurysm Interactive CardioVascular and Thoracic Surgery, October 1, 2008; 7(5): 794 - 796. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Estrera, C. C. Miller III, T.-Y. Lee, P. Shah, and H. J. Safi Ascending and Transverse Aortic Arch Repair: The Impact of Retrograde Cerebral Perfusion Circulation, September 30, 2008; 118(14_suppl_1): S160 - S166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Jakob, K. Tsagakis, P. Tossios, P. Massoudy, M. Thielmann, T. Buck, H. Eggebrecht, and M. Kamler Combining Classic Surgery With Descending Stent Grafting for Acute DeBakey Type I Dissection. Ann. Thorac. Surg., July 1, 2008; 86(1): 95 - 101. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-J. Schafers and T. Kunihara Towards safer reoperations: special aspects in aortic dissection Eur. J. Cardiothorac. Surg., April 1, 2008; 33(4): 700 - 702. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
