| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;81:169-177
© 2006 The Society of Thoracic Surgeons
Section of Cardiothoracic Surgery, Yale University School of Medicine, New Haven, Connecticut
Accepted for publication June 8, 2005.
* Address correspondence to Dr Coady, Section of Cardiothoracic Surgery, Yale University School of Medicine, 333 Cedar St, FMB 121, New Haven, CT 06510 (Email: michael.coady{at}yale.edu).
Presented at the Poster Session of the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24–26, 2005.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
METHODS: Our institutional database contains data on 805 patients followed up serially with thoracic aortic aneurysms. Body surface area information was obtained on 410 patients (257 male, 153 female). We calculated a new measure of relative aortic size, the "aortic size index," and examined its ability to predict complications in these patients.
RESULTS: Increasing aortic size index was a significant predictor of increasing rates of rupture (p = 0.0014) as well as the combined endpoint of rupture, death, or dissection (p < 0.0001). Using aortic size index, patients were stratified into three risk groups: less than 2.75 cm/m2 are at low risk (approximately 4% per year), 2.75 to 4.24 cm/m2 are at moderate risk (approximately 8% per year), and those above 4.25 cm/m2 are at high risk (approximately 20% per year).
CONCLUSIONS: This study confirms that (1) thoracic aortic aneurysm is a lethal disease, (2) relative aortic size is more important than absolute aortic size in predicting complications, and (3) a novel measurement of relative aortic size allows for the stratification of patients into three levels of risk, enabling appropriate surgical decision-making.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
The threats of an unoperated aneurysm include rupture, dissection (with subsequent end-organ ischemia), and death. In contrast, the risks of surgery include paraplegia, stroke, bleeding, and death. Despite the armamentarium of modern perioperative and postoperative cardiac surgical care, the operative mortality after elective aortic surgery ranges from 3% to 9% [1–5]. In contrast mortality for emergent repair is much higher, from 16% to 59% [2, 3, 6]. This emphasizes the need for appropriate timing of surgical intervention, before the occurrence of complications.
Our group has previously demonstrated that both the cumulative and yearly risks of aortic dissection and rupture increase with aneurysm size [3, 7]. While our data emphasized the risk present as aneurysms rise above 5 cm in maximal diameter, in multivariate analysis, we also found that women had a higher likelihood of rupture or dissection. Similarly, in univariate analysis, Juvonen and colleagues [8] found a higher percentage of women among patients who ruptured, although this was not apparent in their elegant multivariate model. Although we speculated that these results reflect the fact that a specific size aneurysm represents proportionately greater aortic dilatation in smaller female patients [7], changes in risk based on patient size have never been demonstrated. We therefore undertook the current study to define the impact of body surface area on risk of aortic complications, and to determine whether the added risk of female sex was simply a surrogate for patient size.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
Inclusion criteria for this study were aortic size at least 3.5 cm and age older than 6 years at presentation, absence of congenital aortic malformations (for example, aortic coarctation), and at least one size measurement before operative repair. In all, 1,106 patients met these criteria. Among these patients, 301 had preexisting chronic dissection at presentation and were excluded from the analysis because dissection was an endpoint to this portion of the study.
Since 2003, height and weight have been collected in a prospective fashion; for patients accrued before 2003, height and weight information was obtained from hospital chart review and patient interview. Four hundred ten patients had accurate data from which body surface area could be calculated, and they form the study group for this analysis.
Hospital chart review was conducted on each identified patient, and the data were entered into a computerized database. Data recovered from hospital records and computer files were cross-referenced with hospital discharge abstract data monitored by the Connecticut Hospital Association and the Connecticut State Mortality Records as well as the Social Security Death Index (available at: http://sssdi.rootsweb.com). The patient database is maintained as part of the ongoing studies at the Yale Center for Thoracic Aortic Disease, a major referral center for southern New England. Patients were recruited and followed up between 1985 and 2005.
Statistical Methods
Statistical methods were used to identify and estimate risk factors for the following outcomes: cumulative incidence of major negative events, survival free from these events, and overall long-term survival. Measurements of cumulative incidence tend to overestimate risk in patients followed up for long periods of time while underestimating risk in patients followed up for short time periods; therefore, the survival free from negative events is a better estimate of true risk.
Body surface area (BSA) was calculated using the Dubois and Dubois formula [9]:
|
|
It was used both as a continuous variable and stratified into two groups: less than 2.00 m2 and 2.00 m2 or more. In addition, the interaction between BSA and aortic size was evaluated using the aortic size index (ASI) which was calculated as: ASI = aortic diameter (cm) divided by body surface area (m2).
When analyzing smoking history, hypertension, and the presence of cardiac, pulmonary, or renal disease, patients were stratified according to established criteria of risk for complications from vascular disease [10], and the analysis was performed both with the stratified severity levels and with a dichotomous variables indicating the presence of disease of any severity. Of note, all patients diagnosed with hypertension received treatment with antihypertensive medications during follow-up. Results are not shown for the analysis with stratified levels because they did not provide any additional information.
The methods of statistical analysis included 
2 test for comparisons of dichotomous risk factors with negative outcomes (rupture, dissection, death); Mantel-Haenszel 2 test for comparisons taking into consideration disease severity (cardiac disease, pulmonary disease, progressively larger aneurysms, and so forth); and the Wilcoxon test for comparisons of continuous variables with negative outcomes (p < 0.05). Logistic regression analysis of the cumulative incidence was used to evaluate the influence of risk factors for rupture or dissection. Product-limit estimates (Kaplan-Meier) were calculated using the LIFETEST procedure of SAS 9.1 for Windows (SAS Institute, Cary, North Carolina) with the log-rank test for difference between strata. Yearly rates of complications were calculated as the mean yearly rate over the first 5 years. The Cox regression model (using the PHREG procedure) was used to identify the most predictive variables.
After univariate analysis, variables were entered into the models using three selection criteria: forward, forward stepwise, and backward. All variables predictive of the outcome in univariate analysis were entered into regression analyses; certain variables were included in all analyses independent of univiariate prediction because of their significance in previous publications, specifically pulmonary disease, sex, and age [3, 7, 8]. Models were compared using the –2 Log L score and the Wald score for significance; in all cases the most predictive model is presented. Threshold for entry into the model for both logistic regression and Cox regression was p less than 0.20, except that all categories of aortic size index were included in the models in order to assess the impact of this variable on outcomes.
Event Rates
The incidence of acute dissection or rupture (or both) before operative repair was evaluated by both descriptive and multivariable analyses. Rupture and dissection were confirmed by at least one of the following: autopsy, operation, death certificate, or computed tomography or magnetic resonance imaging.
The multivariable analysis specifies a logistic regression model relating event occurrence to each of the following: initial aortic size (both stratified and unstratified), initial ASI (both stratified and unstratified), final aortic size (both stratified and unstratified), final ASI (both stratified and unstratified), aneurysm location, age at presentation, sex, Marfan syndrome, presence of a nonsyndromic family history of aortic or aneurysmal disease, cardiac status, hypertension, pulmonary disease, renal disease, history of smoking, and a history of abdominal aortic aneurysm, coronary artery disease, congestive heart failure, or stroke.
Event-Free Survival and Long-Term Survival
Five-year survival estimates were calculated by product-limit estimates (Kaplen-Meier). For these analyses, patients were entered into the analysis at the time of presentation. For the event-free survival analysis, patients were censored when they were lost to follow-up, underwent surgical correction, or died without rupture or dissection. Only two major complications—rupture, dissection, or both—were considered in analyzing event-free survival. For the long-term survival analysis, patients were censored when they were lost to follow-up or underwent surgical correction. Yearly event rates for subgroups were estimated from the Cox regression analysis using the BASELINE statement of PROC PHREG (SAS 9.12 for Windows) and represent the mean event rate for each year during the first 5 years after diagnosis.
The specific factors tested for survival differences included initial and final aortic size (both stratified and unstratified), ASI (both stratified and unstratified), aneurysm location, age at presentation, sex, Marfan syndrome, presence of a nonsyndromic family history of aortic or aneurysmal disease, cardiac status, hypertension status, pulmonary disease, renal disease, history of smoking, and history of abdominal aortic aneurysm, coronary artery disease, congestive heart failure, or stroke.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
|
Cumulative Event Rates
Aortic rupture
Increasing initial aortic size was associated with an increased risk of aortic rupture (p = 0.0320). A stronger association was seen with increasing ASI (p = 0.0022). While there was a steady increase in risk with increasing aortic size, analysis with ASI showed a dramatic step-up in risk above 4.25 cm/m2; initial ASI of 4.25 to 4.99 cm/m2 was associated with a sevenfold increase in the incidence of rupture (odds ratio [OR] 7.9577, p < 0.001). Other univariate predictors of rupture included aneurysm location in the descending or thoracoabdominal aorta (OR 3.281, p = 0.004) and a history of abdominal aortic aneurysm (OR 3.484, p = 0.009). There was a trend toward larger aortic size at rupture among patients with descending or thoracoabdominal aneurysms (6.8 cm versus 5.8 cm, p = 0.2452). Nineteen patients ruptured at ASI less than 4.25 cm/m2. This comprised 76% of all ruptures. Predictors of rupture in this subgroup did not differ significantly from those who ruptured at higher ASI, nor did they differ significantly from patients with small ASI who did not rupture.
Multivariate analysis confirmed the increased risk of rupture associated with increasing ASI (p = 0.0032); ASI 4.25 to 4.99 cm/m2 was associated with an 13-fold increase in the cumulative risk of rupture (OR 13.765, 95% confidence interval [CI]: 3.048 to 62.171). Other significant predictors of rupture in this multivariate model included ASI of 5.00 cm/m2 or greater (OR 7.577, 95%CI: 1.167 to 48.932) and aneurysms located in the descending or thoracoabdominal aorta (OR 2.581, 95%CI: 1.012 to 6.584) When ASI was removed from the model and replaced with aortic size, lower BSA emerged as a significant predictor of rupture (p = 0.0202).
Aortic dissection
Increasing ASI also predicted a higher incidence of dissection (p = 0.0215), but the association was less pronounced than with rupture. Aortic dissection was more strongly associated with aortic size independent of BSA (p = 0.0094 versus 0.0215), and BSA less than 2.00 m2 did not predict a lower incidence of dissection (OR 0.6781, 95%CI: 0.3197 to 1.4385). Aneurysms in the descending or thoracoabdominal aorta were more likely to dissect (OR 2.4453, 95%CI: 1.1298 to 5.2924), while male sex was protective against dissection (OR 0.4063, 95%CI: 0.1974 to 0.8363).
When rupture or dissection are considered together, size index remains highly predictive of negative events. Other significant predictors of rupture or dissection include: aneurysm location in the descending or thoracoabdominal aorta (OR 3.105, 95%CI: 1.6498 to 5.8498) and history of AAA (OR 2.3745, 95%CI: 1.0536 to 5.3515). Both BSA greater than 2.00 m2 (OR 0.5249, 95%CI: 0.2781 to 0.9909) and male sex (OR 0.5029, 95%CI: 0.2800 to 0.9032) were protective against the combined endpoint of rupture or dissection.
Logistic regression confirmed the importance of female sex in predicting dissection (p = 0.0085). Despite the inclusion of BSA in the analysis, female sex was most predictive of dissection before operative repair (OR 2.329, 95%CI: 1.122 to 4.838). Similar results were obtained when rupture and dissection were considered together (data not shown).
Mortality Before Operative Repair
Increasing ASI was highly predictive of the combined endpoint of rupture, dissection or mortality before operative repair (p = 0.0008). Other univariate predictors of the combined endpoint were similar to each endpoint considered alone with a history of AAA and aneurysms located in the descending or thoracoabdominal aorta carrying significantly higher risk (OR 2.5643 and 2.5368, p = 0.013 and 0.002). Results of the multivariate analysis are given in Table 2.
|
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
We previously identified female sex as a significant predictor of negative events, in particular the combined endpoint of rupture or dissection [7]. We hypothesized that this might be due in part to differences in mean body size between males and females, with a given aortic dimension representing proportionally greater diameter in smaller women. Here we demonstrate that lower BSA is associated with a higher incidence of negative events including rupture, dissection, and death.
To more accurately assess the risk of negative events, we developed a new measurement, the aortic size index (ASI) which takes into account both aortic diameter and BSA. Throughout all methods of analysis, ASI was a better predictor of negative events than maximal aortic diameter. This study confirms that TAA is an intrinsically lethal disease and that aortic diameter, corrected for body surface area, is an important predictor of rupture, dissection and death.
In particular, we found that using ASI, patients could be stratified into three categories of risk (Fig 3). Those with ASI less than 2.75 cm/m2 are at low risk for negative events, with a yearly incidence of approximately 4%, those with ASI between 2.75 and 4.25 cm/m2 are at moderate risk with yearly incidence of approximately 8%, whereas those with ASI above 4.25 cm/m2 have yearly rates of rupture, dissection, or death as high as 20% to 25%. These data are summarized in Table 5. These differences in yearly rates were reflected in the Cox proportional hazards regression where the OR for rupture is more than 11 times higher in patients with aortic size index greater than 4.25 cm/m2.
|
Certain limitations of these data can be enumerated. By analyzing the subset of patients with accurate BSA measurements, we have selected a population of patients more likely to have undergone operative repair. Patients presenting to the emergency room moribund after aortic rupture were unlikely to have accurate height and weight information available in the chart. Over the past year we have been collecting this data prospectively, but the patient population as a whole remains skewed toward those reaching operative repair. Thus, the survival in patients with BSA data available was significantly better than in those without such information (p < 0.0001). Second, definition of rupture, dissection and aneurysm-related death was strict, requiring in-hospital documentation by imaging study, operative findings or postmortem examination. The true rupture rate is likely higher. Third, patients we followed up were operated on electively when they reached size criteria, underwent periods of rapid growth, or became symptomatic, eliminating them from further analysis. These two factors: elective repair in patients at high risk and strict definition of rupture mean that the yearly rupture rates reported here likely underestimate the true rates, which will lie between the estimated rate and the mortality rate, as unoperated rupture is a lethal event.
Although size was a significant predictor of poor outcome, it is interesting to note the small but significant risk of rupture among patients with smaller aneurysms. Even the "safest" group had yearly rates of approximately 2% to 3%. As more patients have aneurysms in this smaller size range, as many as 25% of ruptures occur where ASI is less than the median of 2.50 cm/m2 and below what most would consider appropriate operative intervention criteria. Possible explanations for rupture at small size include differing immunologic responses to aortic injury or variation in the morphology of the aneurysm.
Despite the inclusion of BSA in the analysis, multivariate models of endpoints that included dissection still included a protective effect of male sex. This finding may indicate that differences other than size account for some of the increased risk in women. Possible contributing factors include changes in the activity of inflammatory mediators in the presence of higher estrogen levels [11] and increased proximal thoracic aortic stiffness in elderly women [12].
Consonant with our previous report, we were unable to demonstrate a significant effect of pulmonary disease on rupture and dissection rates [7]. This contrasts with the findings of several other investigators [8, 13, 14] and may represent differing definitions of pulmonary disease. As has been found previously [7, 8], hypertension was not predictive of negative events, likely because a diagnosis of hypertension usually implies treatment; accurate serial measurements of blood pressure in these patients will be required to delineate the role of hypertension in aortic disease.
This study permits the following conclusions. Thoracic aortic aneurysm is a lethal disease with yearly event rates in the highest risk patients above 20%. Absolute aneurysm size is predictive of negative events, but relative aortic size as measured here with ASI better predicts the occurrence of rupture and death before operative repair. We recommend elective operative repair before the patient enters the zone of moderate risk with ASI greater than 2.75 cm/m2.
|
The Society of Thoracic Surgeons: Forty-Second Annual Meeting—New Location |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
Please continue to visit www.sts.org for important program information as it develops!
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentThe Society of Thoracic...References |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. M. Tadros, M. D. Klein, and O. M. Shapira Ascending Aortic Dilatation Associated With Bicuspid Aortic Valve: Pathophysiology, Molecular Biology, and Clinical Implications Circulation, February 17, 2009; 119(6): 880 - 890. [Full Text] [PDF]
|
|
|
|
 |
|
 A. T. Yetman and T. Graham The dilated aorta in patients with congenital cardiac defects. J. Am. Coll. Cardiol., February 10, 2009; 53(6): 461 - 467. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Vahanian, B. Iung, L. Piérard, R. Dion, and J. Pepper CHAPTER 21 Valvular Heart Disease ESC Textbook of Cardiovascular Medicine, January 1, 2009; 2(1): med-9780199566990-chapter - med-9780199566990-chapter. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Ruddy, J. A. Jones, F. G. Spinale, and J. S. Ikonomidis Regional heterogeneity within the aorta: Relevance to aneurysm disease. J. Thorac. Cardiovasc. Surg., November 1, 2008; 136(5): 1123 - 1130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J.J. Aalberts, T. W. Waterbolk, J. P. van Tintelen, H. L. Hillege, P. W. Boonstra, and M. P. van den Berg Prophylactic aortic root surgery in patients with Marfan syndrome: 10 years' experience with a protocol based on body surface area Eur. J. Cardiothorac. Surg., September 1, 2008; 34(3): 589 - 594. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. J. Patel and G. M. Deeb Ascending and Arch Aorta: Pathology, Natural History, and Treatment Circulation, July 8, 2008; 118(2): 188 - 195. [Full Text] [PDF]
|
|
|
|
|
|
M. G. Keane and R. E. Pyeritz Medical Management of Marfan Syndrome Circulation, May 27, 2008; 117(21): 2802 - 2813. [Full Text] [PDF]
|
|
|
|
|
|
S. S. Fazel, H. R. Mallidi, R. S. Lee, M. P. Sheehan, D. Liang, D. Fleischman, R. Herfkens, R. S. Mitchell, and D. C. Miller The aortopathy of bicuspid aortic valve disease has distinctive patterns and usually involves the transverse aortic arch. J. Thorac. Cardiovasc. Surg., April 1, 2008; 135(4): 901 - 907.e2. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. G. Svensson, N. T. Kouchoukos, D. C. Miller, J. E. Bavaria, J. S. Coselli, M. A. Curi, H. Eggebrecht, J. A. Elefteriades, R. Erbel, T. G. Gleason, et al. Expert Consensus Document on the Treatment of Descending Thoracic Aortic Disease Using Endovascular Stent-Grafts Ann. Thorac. Surg., January 1, 2008; 85(1_Supplement): S1 - S41. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Matura, V. B. Ho, D. R. Rosing, and C. A. Bondy Aortic Dilatation and Dissection in Turner Syndrome Circulation, October 9, 2007; 116(15): 1663 - 1670. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. E. Achneck, J. A. Rizzo, M. Tranquilli, and J. A. Elefteriades Safety of Thoracic Aortic Surgery in the Present Era Ann. Thorac. Surg., October 1, 2007; 84(4): 1180 - 1185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. R. Davies, R. K. Kaple, D. Mandapati, A. Gallo, D. M. Botta Jr, J. A. Elefteriades, and M. A. Coady Natural History of Ascending Aortic Aneurysms in the Setting of an Unreplaced Bicuspid Aortic Valve Ann. Thorac. Surg., April 1, 2007; 83(4): 1338 - 1344. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Ikonomidis, J. A. Jones, J. R. Barbour, R. E. Stroud, L. L. Clark, B. S. Kaplan, A. Zeeshan, J. E. Bavaria, J. H. Gorman III, F. G. Spinale, et al. Expression of matrix metalloproteinases and endogenous inhibitors within ascending aortic aneurysms of patients with bicuspid or tricuspid aortic valves J. Thorac. Cardiovasc. Surg., April 1, 2007; 133(4): 1028 - 1036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Authors/Task Force Members, A. Vahanian, H. Baumgartner, J. Bax, E. Butchart, R. Dion, G. Filippatos, F. Flachskampf, R. Hall, B. Iung, et al. Guidelines on the management of valvular heart disease: The Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology Eur. Heart J., January 26, 2007; (2007) ehl428v1. [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 |
