| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Click here to read this article as a CME activity
Ann Thorac Surg 2006;82:1658-1664
© 2006 The Society of Thoracic Surgeons
Cardiac Surgery Division, University of Turin, Turin, Italy
Accepted for publication May 15, 2006.
* Address correspondence to Dr Centofanti, S. Giovanni Battista Hospital, Cardiac Surgery Division, University of Turin, C.so Bramante 88, 10126 Turin, Italy (Email: paolocentofanti{at}tiscali.it).
Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30–Feb 1, 2006.
| Adult cardiac surgery:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
METHODS: From 1980 to 2004, 616 consecutive patients with type A acute aortic dissection underwent surgery in our center. The preoperative univariate risk factors with a probability value less than 0.05 were entered into multivariate analysis. A risk equation was developed: predicted mortality 
.
RESULTS: Early mortality was 25.1% (154 of 616 patients). Five risk factors were identified: age, coma, acute renal failure, shock, and redo operation. The ßi values are age 0.023, shock 0.771, reoperation 0.595, coma 1.162, acute renal failure 0.778; the constant (ß0) is –2.986.
CONCLUSIONS: Our large, single-center experience allowed us to develop a mathematical model to predict 30-day mortality for type A acute aortic dissection. When the expected mortality is 58% or less, surgery is always indicated. When the predicted mortality is greater than 58%, the possibility of survival is similar, according to International Registry of Aortic Dissections data, for surgery and medical treatment. In such cases surgery can no longer be considered mandatory, and a careful evaluation of the individual patient is recommended to choose the more suitable strategy.
Type A acute aortic dissection (AAD) is the most common acute aortic condition requiring urgent surgical therapy. Spontaneous mortality is very high: 50% of the patients die within the first 48 hours (roughly 1% per hour mortality) [1]. The possibility of medical management was not considered possible in this disease, but recent studies showed that patients treated nonsurgically had a better prognosis than expected, with an in-hospital mortality of 58% [2]. Despite improvement in diagnosis and in surgical techniques, early mortality remains high, from 15% to 30%, and has been constant during the decades. Furthermore, many reports pointed out that surgical results depend on preoperative conditions, as these are the main determinant of early outcome [3].
This study was undertaken to analyze a large consecutive series of patients with type A AAD to assess the relationship between preoperative risk factors and early mortality. Our goal was to determine the expected mortality for each patient to evaluate the possibility of alternative treatments.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
The hospital records of all the patients were reviewed to obtain the data used in this study. Mean age was 58.3 ± 12 years, with male predominance (74.3%). A total of 22 variables were obtained, including patient demographics, history, clinical presentation, physical findings, surgical management, and mortality during the first 30 days from surgery (Table 1).
|
|
|
Imaging studies changed during the years of the study. Until the early 1990s, patients underwent a preoperative aortogram either at our institution or at the referring hospital. More recently, diagnosis rested on computerized tomography or echocardiography (transthoracic or transesophageal), or both.
Surgical Management
Surgery was performed within 72 hours of the onset of symptoms, immediately after the patient's arrival in our division. By definition, the ascending aorta was involved in all the patients, directly in 579 patients (94.0%) and indirectly in the remaining 37 (6.0%), in whom the entry was in the descending aorta (retrograde dissection).
A standard median sternotomy was performed. The femoral artery was cannulated before opening the pericardium. Venous drainage was obtained by cannulating the right atrium, and a vent was inserted through the right superior pulmonary vein. Cooling was started immediately after going on cardiopulmonary bypass. Surgical strategies changed during the extensive period of time included in the study reflecting the evolution of the state of the art. Deep hypothermic circulatory arrest was applied in 244 patients (40%) to allow for arch inspection or open distal anastomosis or for total or partial arch replacement. The mean time was 26 ± 11 minutes (range, 2 to 81 minutes). In patients undergoing replacement of the arch or hemiarch, circulatory arrest time ranged from 11 to 81 minutes (mean, 26.3 ± 10.4 minutes). Deep hypothermic circulatory arrest was performed with antegrade or retrograde cerebral perfusion in 68 patients (28%). The aortic valve was not treated in 391 patients (63.5%), whereas in the remaining 225 patients it was suspended (n = 55) or included in a Valsalva prosthesis [5, 6] (n = 5), or replaced (n = 165; isolated in 68 patients or composite valve graft in 97 patients [7, 8]). A coronary artery bypass graft was necessary in 27 patients because of lesions of the coronary ostia or impossibility to wean the patient from cardiopulmonary bypass. Duration of cardiopulmonary bypass ranged from 44 to 291 minutes (mean, 128.5 ± 42.2 minutes; outliers, >296 minutes); duration of aortic cross-clamping ranged from 30 to 185 minutes (mean, 85.4 ± 26.7 minutes; outliers, >196 minutes).
Statistical Analysis
The objective of our study was to realize a model to predict 30-day mortality using a scoring system on the basis of objective risk factors. Stepwise logistic regression was used to realize a model to calculate the mortality score. The variables included in this study were preoperative, perioperative, and postoperative, and are listed in the Appendix. Results are expressed as mean value ± standard deviation unless otherwise indicated. Statistical analysis comparing two independent groups was performed with unpaired two-tailed Student's t test for the means or 
2 test for categorical variables.
Variables were first analyzed with univariate analysis, and only those with a probability value less than 0.05 were included in the stepwise logistic regression. Nonsignificant variables were eliminated one at time, beginning with variables having the highest probability value, and stability of the model was checked after each modification. For continuous variables cut points were determined with fractional polynomials method [9]. All interaction terms (defined as being present in at least 5% of population and having a significant relation with p < 0.05) were eliminated. The weight for each factor was obtained from the ß-coefficient of the linear multivariate regression equation.
The final model is a linear logistic model including variables presented as odds ratios. The odds ratio predicts the magnitude of influence on mortality if the risk factor is present compared with its absence.
A probability value of less than 0.05 is considered to be statistically significant after correction for multiple comparisons.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Univariate variables for 30-day mortality are listed in Table 1. Preoperative univariate predictors of death are age of 60 years or older (p < 0.001), preoperative need for mechanical ventilation (p < 0.001), shock (p < 0.001), coma (p < 0.001), aortic rupture (p = 0.001), acute renal failure (p < 0.001), prior cardiac surgery (p = 0.022), and cardiac tamponade (p < 0.001). This latter variable was not statistically significant in the subgroup of patients with prior cardiac surgery (p = 0.02). Perioperative univariate predictors are myocardial ischemia longer than 80 minutes (p < 0.001), need for coronary artery bypass grafting (p = 0.012), and arch replacement (p = 0.039). Duration of circulatory arrest longer than 20 minutes (p = 0.001) is a significant risk factor only in the absence of cerebral perfusion (p = 0.131). Although surgical treatment has evolved in the past decades, the period of time when surgery was performed is not a risk factor (p = 0.420).
A logistic regression model was used to identify which factors were independently associated with higher 30-day mortality. The model was constructed including preoperative and perioperative (distinct models) risk factors with univariate p values less than 0.05. Variables were included using the likelihood ratio test.
Perioperative independent predictors of 30-day mortality are myocardial ischemia longer than 80 minutes and need for coronary artery bypass grafting (Table 2).
|
|
|
| (1) |
The mean expected mortality in our patients was 24.8% ± 15.2% (6.6% to 88.1%) compared with a real 30-day mortality of 25.1%.
Expected and observed 30-day mortality rates were matched in groups stratified according to the expected risk (Table 4) and according to three different periods of time (Fig 1). Matching was satisfying. Thirty-five patients had an expected mortality of at least 58% (early mortality in the medically treated patients in the International Registry of Aortic Dissections [IRAD] survey). In this subgroup the observed mortality was 65.7%.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
There is no doubt that aggressive surgical treatment of type A AAD has spared a large number of lives compared with medical therapy only; the natural history with medical therapy alone would predict a 90-day mortality rate in the range of 70% to 90% [12]. More recent data, however, reveal a different prognosis, suggesting that optimized medical management may be considered acceptable in certain high-risk groups. In the IRAD survey [2], surgery was not performed in 28% of patients with type A dissection for different reasons, and 42% of these were discharged from the hospital after intensive medical treatment. Chan and coworkers [13] report in-hospital mortality rates of 55.9% in patients who did not receive surgery (34 of 90 patients; 37.8%) for clinical (advanced age and comorbidity) and social (family support and economic problems) conditions, and this mortality decreased to 42.9% after a learning period.
Therefore when the expected mortality is equal or higher than 58%, the possibilities for survival are similar, according to the IRAD surgery and others' data, either for surgery or for medical treatment. In such cases surgery is no longer mandatory, and a careful evaluation of the individual patient indicates more suitable strategies.
The possible alternative strategies are intensive medical therapy, with or without invasive procedures (percutaneous fenestration and stenting) or delay of the surgery to decrease operative mortality risk. Deeb and coworkers [14] demonstrated that the likelihood of death in patients with an acute type A dissection complicated by malperfusion is 33 times greater if they undergo immediate surgical repair as opposed to the delay sequence. This experience shows that there is a 5% risk of aortic rupture and death during an average delay period of 21 days (range, 2 to 67 days), versus an 89% risk of death in type A AAD patients with malperfusion who undergo immediate repair.
Trimarchi and colleagues [3] report that unstable patients from the IRAD registry demonstrated a surgical mortality significantly higher than stable patients (31.4% versus 16.7%), regardless of the type of surgical procedure. The mortality in patients with surgical treatment delayed beyond 24 hours was very similar to that observed for stable patients (17.1% versus 16.7%).
Previous reports have indicated that preoperative myocardial ischemia, preoperative resuscitation, peripheral ischemia, visceral ischemia, hemopericardium, renal failure, older age, preoperative hemodynamic instability, and coma are risk factors for hospital death [15]. Our present study indicates that age, coma, acute renal failure, shock, and redo operation are independently associated with an increased risk of early mortality. Furthermore, myocardial ischemia longer than 80 minutes and need for coronary artery bypass grafting were perioperative independent predictors of surgical mortality.
Myocardial ischemia longer than 80 minutes is a risk factor that may be reduced by surgical expertise, appropriate strategies, and development of vascular graft and surgical material of improved quality. The need for a coronary artery bypass graft is an independent risk factor for mortality, but as coronary angiography was never performed in such patients, we cannot exclude previous coronary atherosclerotic lesions. Anyway, patients who needed coronary artery bypass grafting, either because of lesions at the origin as a result of dissection or because of impossibility to be weaned from cardiopulmonary bypass, can be considered at high risk for early mortality.
Results of this study show that preoperative conditions are often the primary cause for postsurgical death. Careful preoperative evaluation of patients allows the surgeon to predict the risk and to establish the strategy for surgical or medical intervention.
Age is independently associated with an increased risk of early mortality. With the general lengthening of the human lifespan, cardiac surgeons have to deal with an increasing number of elderly patients. The decision to reject a patient only because of advanced age could be unethical if not based on a well-documented clinical experience. Mehta and coworkers [16] report that in-hospital surgical mortality rate was 45.5% among patients age 80 to 84 years, and 50% for patients 85 years of age or older. Although these data should not represent a per se criterion to exclude older patients from surgical treatment, they may help the physician in choosing the most appropriate strategy for these patients.
The other preoperative independent risk factors are coma, acute renal failure, shock, and redo operation; the first three are the variables that, in the IRAD registry, are used to define a patient who is "unstable" [3]. Previous cardiac surgery increases early mortality (p = 0.016). This group of patients develops cardiac tamponade less often than others; we suppose that the adherences contain the expansion of the dissecting aneurysm and can reduce the incidence of aortic rupture. This may suggest, in some selected cases, a less aggressive surgical approach, and the possibility of following the clinical evolution with medical management alone.
Our extensive single-center experience, extended for more than 25 years, allowed us to develop a mathematical model to predict 30-day mortality in patients with type A AAD. The interest of this model is related to the recent finding that medical treatment in this group of patients can be more successful than conventionally expected. Even if, during the time frame of this study, surgery was never denied whenever there was a possibility to succeed, the more recent reports seem to favor a different strategy for patients with expected mortality of at least 58%, as deduced by the results of the IRAD survey. When a patient reaches this value, surgery, in our opinion, is not more strictly mandatory, but has to be performed after a careful evaluation of the clinical and anatomic situation. If a patient shows a complication that cannot be treated without surgery, such as cardiac tamponade, then medical treatment, independently from the expected risk, very likely has no possibility of success and surgery is the only option. In the presence of other major complications, such as organ hypoperfusion, surgery is however not always effective and medical treatment (associated, if necessary, with fenestration) has a similar (or even higher) probability of success than surgery.
In our study, 35 patients had an expected mortality of 58% or higher. In these patients, the possibility to survive, according to the IRAD survey, is comparable for surgery and medical treatment. A qualitative analysis of these patients show that 19 of these had a cardiac tamponade, and surgery was inevitable so far as life is concerned (expected mortality was 70.5% and observed mortality was 63.1%), but for the other 16 patients medical treatment could be considered a better option (expected mortality was 69.1% and observed mortality was 68.7%).
Our study has an inherent bias because the examined series of patients derives from a retrospective review covering a long time interval and based mainly on data as related in the available hospital files. Moreover, the patients included in this study have been operated on by different surgeons. Although each surgeon might have a personal approach (the decision to perform a certain repair or to use a certain technique is made intraoperatively and depends on the surgeon's preference), the techniques used were relatively uniform. As a consequence, this study reflects the progressive change in surgical philosophy and the improvements in technical materials, as well as in surgical skill. Nevertheless, early mortality has remained constant for a long period and only in the more recent period (Fig 3) become lower than expected. For this reason we think that a more reasoned introduction of aggressive medical treatment can modify, in selected cases, the outcome of patients who suffer from a disease for which the results of surgery, in spite of many technical improvements, still exhibit a trend for improvement slower than for other fields.
|
|
Appendix |
|---|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
DR CENTOFANTI: Yes, I think that is a good point, because we operated on all the patients because this is a retrospective analysis and we do not have these data because we operate on these patients. But in the future I hope that better results with medical management will increase this percentage of patients that survive with medical therapy. In the Taiwan group where the patients aren't operated on because of economic reasons or less important reasons, the survival with medical treatment is more than 60%.
DR MARC MOON (St. Louis, MO): I think it is difficult to use the IRAD data as your historical control, especially since they had intramural hematomas and other etiologies that we cannot truly explain. Do you have any idea what the mortality rate for medical therapy of acute type A dissections is at your own institution? Certainly you have had a few over that 20-year period.
DR CENTOFANTI: The mortality of the IRAD registry is similar for surgery to our groups. I think it is impossible to compare our data with the IRAD group because we do not have a group of medically treated patients.
DR MOON: I would suggest that also in the IRAD database centers, there were probably a number of patients who died who were getting medical therapy and declined surgical intervention, before they could consent to be in the study.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. S. Krahenbuhl, F. F. Immer, M. Stalder, L. Englberger, F. S. Eckstein, J. Schmidli, and T. P. Carrel Technical advances improved outcome in patients undergoing surgery of the ascending aorta and/or aortic arch: ten years experience Eur. J. Cardiothorac. Surg., September 1, 2008; 34(3): 595 - 599. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Shrestha, N. Khaladj, A. Haverich, and C. Hagl Is Treatment of Acute Type A Aortic Dissection in Septuagenarians Justifiable? Asian Cardiovasc Thorac Ann, February 1, 2008; 16(1): 33 - 36. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Khaladj, A. Haverich, and C. Hagl Should a Patient With Acute Aortic Dissection Type A Go to the Intensive Care Unit or Operating Room? Ann. Thorac. Surg., September 1, 2007; 84(3): 1069 - 1069. [Full Text] [PDF]
|
|
|
|
|
|
P. P. Lima-Canadas, L. F. Lopez-Almodovar, J. L. Vallejo, and I. Diaz de Tuesta Can the Risk Associated With Surgery in Type A Aortic Dissection Be Predicted Accurately? Ann. Thorac. Surg., September 1, 2007; 84(3): 1070 - 1070. [Full Text] [PDF]
|
|
|
|
|
|
P. Centofanti, R. Flocco, F. Ceresa, M. Attisani, M. La Torre, L. Weltert, and A. M. Calafiore Reply Ann. Thorac. Surg., September 1, 2007; 84(3): 1069 - 1070. [Full Text] [PDF]
|
|
|
|
|
|
P. Centofanti, R. Flocco, F. Ceresa, M. Attisani, M. La Torre, L. Weltert, and A. M. Calafiore Reply Ann. Thorac. Surg., September 1, 2007; 84(3): 1070 - 1071. [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 |
