Author + information
- Received May 12, 2008
- Revision received June 23, 2008
- Accepted June 24, 2008
- Published online September 30, 2008.
- Chih-Ping Chang, MD⁎,
- Juhn-Cherng Liu, MD†,§,
- Ying-Ming Liou, PhD‡,
- Shih-Sheng Chang, MD⁎ and
- Jan-Yow Chen, MD⁎,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Jan-Yow Chen, Division of Cardiology, Department of Medicine, China Medical University Hospital, 2 Yuh-Der Road, Taichung 404, Taiwan
Objectives The aim of this study was to determine whether false lumen size predicts in-hospital complications for acute type B aortic dissection.
Background The incidence of complications developing in patients with acute type B aortic dissection has been high. However, methods for recognizing high-risk patients have not been well-studied. We used quantitative analysis by computed tomography (CT) to predict the occurrence of in-hospital complications.
Methods Fifty-five consecutive patients with acute type B aortic dissection documented by CT imaging were analyzed. They were divided into groups, with and without in-hospital complications, and compared regarding maximal aortic diameter (MAD), maximal false lumen area (MFLA), minimal true lumen area (MTLA), branch-vessel involvement (BVI), and longitudinal length (LL) of aortic dissection.
Results There were 31 patients with a stable course (group 1) and 24 patients who developed complications (group 2). The MFLA of group 2 was significantly larger than that of group 1 (group 1 vs. group 2 = 577.7 ± 273.2 mm2 vs. 1,899.3 ± 1,642.4 mm2, p < 0.001). The BVI number was also higher in group 2 (group 1 vs. group 2 = 1.0 ± 1.1 vs. 3.3 ± 2.0, p < 0.001). On multivariate analysis, only MFLA and BVI number were independent predictors of in-hospital complications. Patients with initial MFLA ≥922 mm2 or BVI number ≥2 showed a significantly higher incidence of in-hospital complications than the other patients (p < 0.001).
Conclusions A large MFLA and a higher BVI number are powerful predictors of in-hospital complications after acute type B aortic dissection.
Acute type B aortic dissection is an emergent cardiovascular disease with potentially serious complications (1,2). Despite significant advancement in the diagnosis and treatment of aortic dissection, mortality and morbidity rates remain high (3,4–7). The early mortality rate for type B aortic dissection ranges from 10% to 12% (1,8–10). The incidence of early complications is also significant and usually lethal (1,10). Yet, the optimal treatment for patients with acute type B aortic dissection remains unclear (11). Generally, the preferred treatment for patients without complications after Standford type B aortic dissection is medication therapy (1,12,13). Surgical intervention is reserved for cases that are complicated by progression of dissection, rupture or impending rupture, refractory chest pain, or end-organ ischemia caused by compromised aortic branches (11,14). These complications might occur early and frequently, and the outcome is usually fatal (15,16). However, prediction of patients who are at high risk of early lethal complications is still not well understood. The aortic diameter has been reported as a predictor of lethal complications of aortic dissection (17,18). In part, those results might be explained by La Place's law, which describes wall stress as being proportional to the pressure exerted by the fluid contents and its radius and being inversely proportional to wall thickness (17,19,20): therefore, the larger the diameter, the higher the resulting stress on the vessel wall and the greater the possibility of lethal complications.
Nevertheless, aortic dissection presents with a dual-lumen rather than a single-lumen. The anatomy of aortic dissection is usually complex, including irregular lumen shape and length. The impact of anatomic factors in aortic dissection is complicated and difficult to explain by total aortic diameter alone. Therefore, we propose that the occurrence of acute in-hospital complications of aortic dissection might be influenced by multiple anatomic factors, not only the total aortic diameter.
Computed tomography (CT) has been used to evaluate the aorta and to predict the outcome of aortic disorders (1–3,10,17,19). In the present study, we used CT imaging to identify the characteristics and abnormalities of aortic dissection. Thereafter, we tried to identify CT predictors of in-hospital complications of type B aortic dissection. Such predictors might be beneficial when determining strategies for managing type B aortic dissections.
The population studied consisted of 55 consecutive patients (51 male, 4 female) with a mean age of 58.3 ± 14.7 years. All patients underwent cross-sectional imaging by CT in our emergency department and met the criteria for having a Standford type B aortic dissection. Those patients with a stable hospital course were enrolled in group 1, and the patients with acute in-hospital complications comprised group 2. The criteria used to define a complicated type B aortic dissection are as follows: 1) in-hospital death due to aortic dissection; 2) expansion and progression of dissection in the acute phase; 3) impending rupture or actual rupture of the aorta; and 4) end-organ malperfusion. The clinical and radiological characteristics were compared between both groups.
Both nonenhanced and enhanced images were acquired with a spiral CT scanner for all patients. The enhanced CT was performed with a bolus injection of 100 ml of ionic or nonionic contrast material. The CT was performed with a Light Speed 16 scanner (GE Medical System, Milwaukee, Wisconsin) or Twin Flash scanner (Elscint, Haifa, Israel) that generated axial images with contiguous 5-mm–thick sections from the top of the aortic arch to the abdominal aorta. In the spiral CT scanner, a power injector was used and scanning began at 30 s and 120 to 150 s, after starting to inject the contrast material.
The CT images were evaluated by 1 radiologist (J.-C.L.) and by 1 cardiologist (C.-P.C.). All CT images taken during the hospital stay of each patient were analyzed prospectively and independently. The observers were blinded to the presence of complications. The aortic arch was defined as the segments between the brachiocephalic artery and the ligamentum arteriosus. The descending aorta was defined as the segment below the ligamentum arteriosus. Parameters of CT images, including the maximal aortic diameter (MAD), maximal false lumen area (MFLA), minimal true lumen area (MTLA), number of branch-vessel involvement (BVI), and total longitudinal length (LL) of the dissection, were measured off-line. The largest short-axial diameter of the outer contour of the affected aortic segment was measured and defined as the MAD, on the basis of the method that has been reported (Fig. 1A) (17). The MFLA was defined as the area over the segment with maximal false lumen in the descending aorta (Fig. 1B). The MTLA was defined as the area over the segment with minimal true lumen in the descending aorta with dissection. The false lumen area, true lumen area, and the diameter of the aortic dissection were measured by the CT imaging system software and corrected to the appropriate scale. The total LL of the dissection was measured by the total thickness of all affected contiguous sections of aortic dissection. The number of BVI was defined as the number of branch vessels including the carotid artery, celiac trunk, superior mesentery artery, inferior mesentery artery, renal artery, and iliac artery with involvement of dissection and presence of a false lumen.
Continuous data were expressed as mean ± SD. The Student t test was used when the data were normally distributed; otherwise, the nonparametric Mann-Whitney U test was used. The decision of tests used for comparison of continuous data was based on the test for normality by the Shapiro-Wilk test. Categorical data were compared by the conventional chi-square test if the observation numbers in all categories were larger than 5; otherwise, the Fisher exact test was used. The 5 CT parameters between the groups with either a smooth course or with lethal complications were compared. The relative influence of these variables was tested by multivariable stepwise logistic regression analysis. A receiver-operating characteristic curve analysis and Youden index were used to determine the best cutoff value, which maximized the sum of sensitivity and specificity for predicting in-hospital complications. The sensitivity and specificity of a cutoff value for predicting the in-hospital complication were considered to be both important in our study. The criterion of Youden index represents a situation when the loss of false positive equals that of the false negative. A p value <0.05 was considered statistically significant.
Of the 55 patients, 31 had a stable course during the hospital stay (group 1) and 24 had complications (group 2). The clinical features of patients in groups 1 and 2 are summarized in Table 1. Baseline characteristics were similar between groups.
In-hospital complications of patients with acute type B aortic dissection
Nine patients had renal artery involvement with deteriorated renal function. Eight had limb ischemia due to compromised femoral arteries caused by the aortic dissection, and emergency surgical intervention was required. Two patients suffered from mesenteric ischemia after aortic dissection. Seven patients had symptomatic chest pain and progression of aortic dissection, as documented by follow-up CT during admission. Seven patients suffered from rupture or impending rupture of their aortic dissection, with symptoms such as persistent chest pain and unstable vital signs (Table 2). Cardiovascular surgeons typically recommend surgical interventions for acute and potentially lethal complications. Four patients died during the early stages of hospital stay after receiving emergency surgery for rupture or impending rupture of an aortic dissection.
CT findings and predictors of in-hospital complications
The average MAD was 46.1 ± 14.9 mm (range 26 to 102 mm), and the average total LL of aortic dissections was 29.0 ± 10.9 cm (range 3 to 50 cm). The MFLA was 1,154.4 ± 1,275.9 mm2 (range 86 to 6,740 mm2), and the MTLA was 208.1 ± 201.5 mm2 (range 0 to 901 mm2). The average number of BVI was 2.0 ± 1.9 (range 0 to 6).
Computed tomographic parameters, including MAD, MFLA, total LL, number of BVI, and MTLA, were selected for comparisons between patients of group 1 and group 2. In univariate analysis, there was a significantly higher MFLA, total LL, MAD, and number of BVI in group 2 (p < 0.001, p = 0.004, p = 0.025, and p < 0.001, respectively). In multivariate analysis, only MFLA and number of BVI were independent variables for prediction of complicated type B aortic dissection (p = 0.02 and p = 0.004, respectively) (Table 3).
On the basis of the receiver-operating characteristic curve, the best cutoff value for MFLA to predict in-hospital complications was ≥922 mm2 (sensitivity 79.2%; specificity 96.8%). The best cutoff value for number of BVI was ≥2 (sensitivity 83.3% and specificity 74.2%). In a total of 55 patients, 20 patients with an MFLA ≥922 mm2 showed significantly higher incidences total of in-hospital complications (95% vs. 14.3%, p < 0.001), rupture (30% vs. 2.9%, p = 0.007), progression (35% vs. 0%, p < 0.001), and clinically branch-vessel compromise (50% vs. 14.3%, p = 0.01) than the other 35 patients who had an MFLA <922 mm2 (Fig. 2). Of the total 55 patients, the 28 patients with a number of BVI ≥2 demonstrated a higher total of in-hospital complications (71.4% vs. 14.8%, p < 0.001), progression (25.0% vs. 0%, p = 0.01), and clinically branch-vessel compromise (53.6% vs. 0%, p < 0.001) than the 27 patients with a number of BVI <2 (Fig. 3). The incidence of rupture was not different between patients with a number of BVI ≥2 and those with BVI <2 (10.7% vs. 14.3%, p = 0.71).
The effect of MTLA was analyzed, because of its suspected influence in development of malperfusion syndrome. On univariate analysis, the 15 patients who experienced malperfusion syndrome during hospital stay showed a lower MTLA than the 40 patients without malperfusion syndrome (99.47 ± 64.1 mm2 vs. 248.9 ± 220.2 mm2, p = 0.003). However, on multivariate analysis with parameters including MAD, MFLA, total LL, number of BVI, and MTLA, only BVI was an independent predictor of malperfusion syndrome (p = 0.001). Also, the MTLA did not demonstrate a significant difference between patients with complication of rupture or progression in univariate analysis and those without (p = 0.63 and p = 0.19, respectively).
Acute type B aortic dissection is a devastating disease with high morbidity and mortality (4–7). In general, initial treatment for acute type B aortic dissection is medical, whereas patients with complications from acute type B aortic dissection need a more aggressive intervention (13,14,21–24). Complications tend to occur early, and the outcome is often serious and fatal (3). However, the early mortality rate is still significant, despite aggressive medical management (8,9,25). Conventional surgical reconstruction carries a significant risk of morbidity and mortality in acute complicated type B aortic dissection (13,26–29). The recent development of endovascular stent-graft treatment provides an alternative choice for patients with lethal complications, even in the acute phase (30–37). Surgical intervention or endovascular treatment is recommended for acute aortic dissection patients with acute lethal complications. Those patients have better outcomes than patients with medical treatment alone (2,3,9,18,25,29,31–34,38,39). Therefore, identification of patients at high risk of developing early or in-hospital complications is important.
Predictors of in-hospital complications
The MAD of a type B aortic dissection has been reported to be predictive of the need for elective surgery and of in-hospital mortality (10,18). It has also been reported that the initial descending thoracic aorta diameter is a risk factor for late descending aortic aneurysmal formation in patients with type A aortic dissection after ascending aorta surgery (40). However, the initial aortic diameter still does not explain well the outcome and subsequent complications (18,41–45). The acute complications of type B aortic dissections were always due to the progression of the dissection, a ruptured aneurysm, or compromised end-organ perfusion (2,18). These results were always associated with a large or progressively enlarged false lumen. A large aortic diameter might be a major contribution by an enlarged false lumen. Therefore, we propose that the characteristics of the false lumen play a more important role than a large aortic diameter in the development of complications in type B aortic dissection patients.
In our study, we used CT imaging to quantitatively measure the aortic diameter and true lumen and false lumen sizes, including the MAD, MTLA, MFLA, total LL, and the number of BVI. On univariate analysis, MAD, MFLA, LL of false lumen, and BVI number were all shown to be significantly different between patients with a stable course and patients with acute complications. However, on multivariate analysis, only MFLA and BVI number were independent predictors of in-hospital complications. A large area of the false lumen has been reported to favor secondary dilation of the aorta after acute type A aortic dissection (42). We demonstrated that, consistent with that report, in patients with acute type B aortic dissection, those with a large initial MFLA have a higher incidence of in-hospital complications. The aortic dissection exhibited dual-lumen morphology, and the vessel wall of the false lumen was considered to be weaker than that of the true lumen. In patients with a large false lumen area, blood pressure has been shown to be higher in the false lumen compared with the true lumen (42). Therefore, the effect of wall stress might more significantly affect the false lumen.
Also, the structure of an aortic dissection is usually complex. Therefore, the influence of a false lumen cannot always be predicted, on the basis of La Place's law, with total MAD alone (17,19,20,43). In our study, some patients with a large MAD exhibited a small false lumen and had benign outcomes. In contrast, some patients did not have a large aortic diameter but had a large false lumen, and their outcomes were usually problematic.
Song et al. (43) has reported that a larger false lumen diameter of the upper descending thoracic aorta on the initial CT portends a late aneurysm and an adverse outcome in patients with aortic dissection involving the descending aorta. They showed that the initial false lumen diameter was the most powerful predictor and was better than the initial aorta diameter in predicting late aneurysmal change. Their study also demonstrated that the initial false lumen diameter but not the aorta diameter correlated with the rate of aorta dilation after aortic dissection. Consistent with that report, our data showed that the role of false lumen size is greater than the role of true lumen size in predicting in-hospital complications in acute type B aortic dissection patients. However, we found that the false lumen diameter was sometimes difficult to define due to irregular border and irregular morphology of the false lumen. Patients with a similar false lumen diameter might have quite different false lumen areas, particularly in a patient with a large false lumen surrounding the true lumen. The false lumen size might be underestimated in some patients by a false lumen diameter method. Therefore, we selected the MFLA, not the false lumen diameter, for estimation of the false lumen size and then predicted subsequent complications.
The total LL in our study showed a significant difference between the patients with a stable hospital course and those with complicated hospital courses on univariate analysis. However, on multivariate analysis, the total LL has not been demonstrated as an independent predictor. The length of the false lumen might somewhat reflect the presence or absence of involvement of the mid and distal abdomen aorta, leading to branch vessel compromise. In our study, patients with a longer false lumen always exhibited a larger BVI number (r = 0.62, p < 0.001). We suggest that a higher BVI number might reflect a sign of a high false lumen tension and shearing forces in the initial stages of type B aortic dissection and result in subsequent branch vessel compromise. To our knowledge, the present report is the first study demonstrating a role for MFLA and BVI number in predicting in-hospital complications of acute type B aortic dissection.
In the present study, MTLA has not been demonstrated as a significant predictor of total in-hospital complications after acute type B aortic dissection. However, patients with malperfusion syndrome showed significant lower MTLA than patients without malperfusion syndrome on univariate analysis. However, on multivariate analysis, only BVI number was an independent predictor for malperfusion syndrome. There was a significant negative correlation between MTLA and BVI number (r = −0.50, p < 0.001). Patients with a higher BVI number always demonstrated a lower MTLA. Patients with a higher BVI number were suspected to have a higher false lumen pressure, which then limited the true lumen area.
In patients with aortic dissection involving the aortic arch, measurement of the false lumen area is limited in the descending aorta because the transectional lumen area could not be well delineated in the aortic arch. Therefore, the identified false lumen areas might not have been maximal in those patients. Another limitation of the study is that the study population is small. Our results should be confirmed in a larger-scale study.
The present study demonstrated a crucial role for false lumen size factors in predicting acute in-hospital complications after acute type B aortic dissection. It suggests that the effect of anatomic factors on aortic dissection is complicated and difficult to explain by total aortic diameter alone. The occurrence of acute in-hospital complications might be predominantly influenced by false lumen size factors including MLFA and BVI number. These 2 anatomic factors were both independent predictors of in-hospital complications. Measurement and estimation of the initial MFLA and BVI number by CT imaging are beneficial for determining strategies for managing type B aortic dissection.
On the basis of our data, patients with a large false lumen area (MFLA ≥922 mm2) and higher BVI number (≥2) are at high risk for developing in-hospital complications including rupture, progression, and malperfusion syndrome. Early and aggressive intervention including surgery or stent-graft treatment might be considered for them.
Computed tomography imaging is a useful tool to identify the patient at high risk of developing in-hospital complications. A large MFLA and high BVI number might predict the development of acute in-hospital complications in patients with acute type B aortic dissection. More aggressive medical management and early intervention is suggested for those patients in the early phase of type B aortic dissection.
The authors are grateful to Dr. Hong-Dar Isaac Wu of China Medical University Biostatics Center for his contribution to the statistical analysis of our study.
- Abbreviations and Acronyms
- branch-vessel involvement
- computed tomography
- longitudinal length
- maximal aortic diameter
- maximal false lumen area
- minimal true lumen area
- Received May 12, 2008.
- Revision received June 23, 2008.
- Accepted June 24, 2008.
- American College of Cardiology Foundation
- Nienaber C.A.,
- Eagle K.A.
- Roseborough G.,
- Burke J.,
- Sperry J.,
- Perler B.,
- Parra J.,
- Williams G.M.
- Erbel R.,
- Alfonso F.,
- Boileau C.,
- et al.
- Chirillo F.,
- Marchiori M.C.,
- Andriolo L.,
- et al.
- Doroghazi R.M.,
- Slater E.E.,
- DeSanctis R.W.,
- Buckley M.J.,
- Austen W.G.,
- Rosenthal S.
- Eggebrecht H.,
- Nienaber C.A.,
- Neuhauser M.,
- et al.
- Bortone A.S.,
- Schena S.,
- Mannatrizio G.,
- et al.
- Pape L.A.,
- Tai T.T.,
- Isselbacher E.M.,
- et al.,
- International Registry of Acute Aortic Dissection (IRAD) Investigators
- Song J.M.,
- Kim S.D.,
- Kim J.H.,
- et al.
- Akutsu K.,
- Nejima J.,
- Kiuchi K.,
- et al.