Author + information
- Received June 11, 2001
- Revision received October 31, 2001
- Accepted November 16, 2001
- Published online February 20, 2002.
- Yskert von Kodolitsch, MD*,* (, )
- Muhammet A Aydin, MD*,
- Dietmar H Koschyk, MD*,
- Roger Loose, MD§,
- Ilka Schalwat, MD*,
- Matthias Karck, MD‡,
- Jochen Cremer, MD§,
- Axel Haverich, MD‡,
- Jürgen Berger, PhD†,
- Thomas Meinertz, MD* and
- Christoph A Nienaber, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. Yskert von Kodolitsch, Department of Internal Medicine, Division of Cardiology, University Hospital Hamburg-Eppendorf, Martinistrasse 52, 20246, Hamburg, Germany.
Objectives We sought to identify the predictors of aneurysmal formation after surgical correction of aortic coarctation.
Background In 9% of patients, aneurysms develop late after corrective surgery of coarctation of the aorta, with a 36% mortality rate if left untreated. However, the predictors of postsurgical aneurysmal formation are unknown.
Methods Of 25 aortic aneurysms requiring corrective surgery 152 ± 78 months after the initial coarctation repair, 8 were located in the ascending aorta (type A) and 17 at the site of previous repair (local type). Seventy-four patients without progression of the aortic diameter within 189 ± 71 months after coarctation repair were used for categorical data analysis in an attempt to identify the predictors of postsurgical aneurysmal formation.
Results Advanced age at coarctation repair (p = 0.004) and patch graft technique (p < 0.0005) independently predicted local aneurysmal formation. Type A aneurysm was univariately associated with the presence of a bicuspid aortic valve (p = 0.02), advanced age at coarctation repair (p = 0.044) and a high preoperative peak systolic pressure gradient of 74 ± 21 mm Hg (p = 0.041). Conversely, multivariate analysis confirmed only the presence of a bicuspid aortic valve (p = 0.015) as an independent predictor of type A aneurysm. Receiver operating characteristic curve analysis revealed that 72% of patients with a postsurgical aneurysm had an operation at age 13.5 years or more, whereas 69% with no postsurgical aneurysm had an operation at a younger age.
Conclusions Use of the patch graft technique and late correction of coarctation can predict aneurysmal formation at the site of coarctation repair, although patients with a bicuspid aortic valve may be at risk for an aneurysm developing in the ascending aorta, particularly after late repair of aortic coarctation with high preoperative pressure gradients.
Severe coarctation of the thoracic aorta is usually corrected by surgical repair. Despite primary success, however, 9% of patients develop aortic aneurysms late after the operation, with aortic rupture and a lethal outcome reported in some (1–15). To avoid late aneurysmal formation, patients at risk require early identification. A postsurgical aneurysm was observed with subclavian flap angioplasty (17%), patch angioplasty (14%), tube graft repair (6%) or end-to-end anastomosis (3%) (1–15), often associated with persistent systemic hypertension, concomitant cardiovascular malformations or use of Dacron and/or silk sutures (6–9,11), and presented as false, true or dissecting at various sites of the thoracic aorta (3,8,9). With this wide scope of potential risk factors, the prediction of aneurysmal formation after coarctation repair is not feasible to date; thus, this controlled, retrospective analysis of data banks at three cardiovascular centers was conducted to design models for predicting aneurysms both in the ascending and descending aorta.
Between January 1989 and December 1999, 1,665 patients with surgical repair of a dissecting or true aneurysm of the ascending aorta (1,200 patients), the descending aorta (383 patients) or the aortic arch (82 patients) were registered at the Departments of Cardiovascular Surgery at the University Hospital Hamburg-Eppendorf, the Hannover Medical School and the Christian-Albrechts-University, all in Germany. Twenty-five patients (1.5%) with previous surgical repair of coarctation of the aorta and late aneurysmal formation were extracted from this database. The location, extent and diameter of the postsurgical aneurysm were assessed from tomographic imaging or angiography, or both, performed before the operation (16,17)and from surgical records on aneurysm repair. An urgent operation was performed for acute aortic dissection (Patient nos. 1 and 2) (Table 1), rapid progression (Patient nos. 4 and 15), rupture of an aneurysm (Patient nos. 16, 17 and 20) or marked aortic root enlargement with aortic regurgitation (Patient nos. 3 and 5 to 8); elective repair was performed in the remaining patients. Aneurysm in the ascending aorta (termed “type A”) was repaired using the composite graft technique with re-implantation of coronary arteries (Patient nos. 2 and 4 to 7), a supracoronary tube graft (Patient nos. 1, 3 and 8), a biologic aortic valve prosthesis (Patient no. 4) or a bi-leaflet mechanical valve (17). Aneurysm in the descending thoracic aorta (termed “local” type) was repaired by resection of the dilated aortic segment and insertion of a tube graft (Patient nos. 9 to 25).
In the attempt to identify the stratifiers of risk for postsurgical aneurysmal formation, we studied a control group of 74 consecutive patients (40 men and 34 women; 11 ± 14 years of age at index operation) who had successful repair of aortic coarctation within the same time interval as the study patients; 38 of these patients had end-to-end anastomosis, 21 received patch grafts and 15 had a tube graft for coarctation repair (Table 2). This group had an uneventful follow-up of 189 ± 71 months, with exclusion of postsurgical complications, aortic wall abnormalities, aortic diameters exceeding 3.5 mm and progression of diameters, as assessed by serial tomographic imaging (e.g., transesophageal echocardiography, computed tomography, magnetic resonance imaging) during repeat follow-up visits (16,17). Patients with inflammatory cardiovascular disease and complex, concomitant cardiovascular malformations were not included (e.g., Turner’s syndrome, Marfan syndrome, transposition of the great arteries, dextrocardia, ventricular septal defect, double-outlet right ventricle). Complete clinical and morphometric documentation and follow-up data were available for all patients (Table 1). Any cardiovascular event requiring hospital admission during follow-up after a repeat intervention for postsurgical aneurysmal repair was considered a “cardiovascular event.”
Sixteen variables were assessed, including: 1) age at the time of surgical correction of coarctation and 2) gender (Table 2). 3) Coarctation of the aorta was considered as ductal, preductal or postductal, with abrupt narrowing of the aorta at, proximal to or distal to the site of the ductus arteriosus, respectively (18). 4) The surgical technique of coarctation repair and 5) the presence of previous surgical repair of coarctation were assessed in each patient. Rest right arm blood pressure (BP) was assessed by sphygmomanometry on three different days before coarctation repair and during follow-up visits. Both systolic and diastolic BPs were given as the average of three preoperative and three postoperative measurements. Systolic BP (variables 6 and 9) and diastolic BP (variables 7 and 10) were elevated, with measurements above the 90th percentile with age below 18 years and systolic BP >140 mm Hg and with age above 18 years and diastolic BP >90 mm Hg (19,20). Peak systolic pressure gradients were recorded from percutaneous retrograde catheterization as the intra-aortic pullback pressure across either the coarctation or the site of coarctation repair. In 10 control subjects and 3 patients (Patient nos. 21, 24 and 25), estimates of preoperative pressure gradients were derived from pulsed Doppler flow studies (variables 8 and 11) (21). 12) The presence of a congenitally bicuspid aortic valve (CBAV) was assessed by echocardiography and from surgical records (Patient nos. 1 to 8). 13) The presence of patent ductus arteriosus was documented on preoperative cardiac catheterization or color flow Doppler mapping. 14) Each patient had a similar clinical and echocardiographic examination for the presence of coarctation (21). 15) Patients were considered to be under postoperative beta-receptor blockade with oral administration of atenolol or metoprolol over at least six months after coarctation repair; 16) left ventricular hypertrophy was assessed at least six months after the operation (22).
Review of published data
Using MEDLINE (key words: aneurysm, aorta, coarctation and surgery), the English and German published data were screened for records on both the prevalence and outcome of aneurysmal formation after surgical correction of coarctation. Reports from series including at least 20 patients with surgical correction of coarctation were included in this analysis, as shown in Table 3.
The relationship between developing aneurysms after coarctation repair and each of the 16 variables was evaluated for aneurysms in the ascending aorta (type A) and descending aorta (local) and for the control group, using analysis of variance for continuous variables and the Freeman-Halton test for nominal and categorical variables. With the global test significant at p < 0.05 (two-sided), the Dunnett test was performed for post hoc multiple comparisons of continuous variables between both type A and local aneurysm groups and the control group; categorical variables were compared using the Freeman-Halton test with the Bonferroni correction. Only variables with statistical significance set at p < 0.05 (two-sided) for these univariate tests were included in a multivariate logistic regression model. Estimates of risk (odds ratio) were calculated based on coefficients from the logistic models. Receiver operating characteristic (ROC) curve analysis was performed to assess the patient’s age at coarctation repair as a discriminator of increased (from low) risk for postsurgical aneurysm. Any cardiovascular event requiring hospital admission was analyzed by the Kaplan-Meier method separately for aneurysms in the ascending and descending aortas, with the log-rank test for comparison of both curves; Patient no. 17 developed an aneurysm in the ascending aorta after repair of a postsurgical aneurysm in the descending aorta and, thus, could not be assigned to either group. Data are presented as the mean value ± SD or frequencies. Statistical analysis was performed using SPSS software (SPSS for Windows, Release 10.0.7, SPSS Inc. 1989 to 1999, Chicago, Illinois).
Types of postsurgical aneurysm
Reoperation for aortic aneurysmal formation was performed at 153 ± 71 months (type A, located in the ascending aorta) and 152 ± 78 months (local type, located in the descending aorta) after the index coarctation repair (p = 0.98). Type A aneurysms developed a dissection in two patients (Patient nos. 1 and 2), and six patients had moderate (Patient nos. 5 and 8) or severe aortic valve regurgitation (Patient nos. 1 to 4). With one exception, the diameters of the ascending aorta were at least 70 mm (Patient no. 7) (Fig. 1), and marked (Patient nos. 1, 2, 4 and 5) or moderate (Patient nos. 7 and 8) media degeneration was present in all but one aortic specimen obtained during aneurysm repair (Patient no. 3). Local aneurysms developed proximal to recurrent isthmic stenosis (Patient nos. 9 to 11), at suture lines (Patient nos. 12 to 17), with rupture in two patients (Patient nos. 16 and 17), or at or adjacent to the aortic patch (Patient nos. 18 to 21) or with tube grafts extending into the aortic arch (Patient nos. 22–25). In type A aneurysms, aortic diameters were larger (83 ± 21 mm vs. 56 ± 18 mm in the local type; p = 0.003), CBAVs were more prevalent (Patient nos. 1, 2, 4 to 8, 12, 21 and 24; p = 0.002) (Table 2) and emergency repair was slightly more common, as compared with local aneurysms (37% vs. 23%; p = 0.64).
Postsurgical follow-up of aneurysmal repair was 110 ± 31 months for type A aneurysms and 80 ± 92 months for local aneurysms. The five-year actuarial freedom from cardiovascular events (±SE) was slightly lower in type A than in local aneurysms (75 ± 15% vs. 100%; p = 0.08) (Fig. 2). Cardiovascular events in surgically corrected type A aneurysms included cerebral bleeding after oral anticoagulation (Patient no. 5), acute coronary syndrome (Patient no. 6), recurrent aortic aneurysm (Patient no. 24) and aortic valve incompetence (Patient nos. 2, 4 and 8). Perioperative mortality (30-day), however, was similar in both types of aneurysms (0 in type A aneurysms vs. 1 intraoperative death in local aneurysms; p = 0.48).
Predictors of postsurgical aneurysm
Although preoperative and postoperative BP, anatomic type of coarctation, use of Dacron material and concomitant cardiovascular abnormalities were not associated with postsurgical aneurysms, univariate analysis identified advanced age at the index coarctation repair (p = 0.044), a bicuspid aortic valve (p = 0.02) and high preoperative peak systolic pressure gradients (74 ± 21 mm Hg in patients with type A aneurysms vs. 52 ± 27 mm Hg in control subjects; p = 0.041) as risk factors of type A aneurysms (Table 2). Multivariate analysis, however, corroborated only a bicuspid aortic valve (p = 0.015) as an independent predictor of type A aneurysmal formation (Table 4). Similarly, the use of patch grafts (p < 0.0005) and advanced age at the index coarctation repair (p = 0.004) emerged as independent predictors of a local aneurysm (Table 4).
Receiver operating characteristic curve analysis revealed 13.5 years of age at coarctation repair as the threshold for increased risk of postsurgical aneurysmal formation (p < 0.0005) (Fig. 3). Seventy-two percent of patients with a postsurgical aneurysm were surgically repaired beyond the critical age of 13.5 years, whereas 69% of patients with uneventful cases were repaired under the age of 13.5 years.
Formation of an aortic aneurysm after corrective surgery for coarctation was previously reported in 9% of patients, irrespective of the technique of surgical correction (1–15)(Table 3). Our retrospective analysis identified 17 cases of postsurgical aneurysm located in the descending aorta at the site of previous coarctation repair among a total of 383 surgically corrected descending aortic aneurysms (4.4%) and 8 aneurysms in the ascending aorta after coarctation repair among 1,200 surgically corrected aneurysms in the ascending aorta (0.7%). These aneurysms were subjected to corrective surgery 152 ± 78 months after the initial coarctation repair. The outcome of postsurgical aneurysms was adverse: death occurred in 36% of patients if left untreated and in 9% even after surgical resection, whereas only 4% of patients had an early death after surgical repair of a postoperative aneurysm. The results from this analysis are in agreement with previous reports (1–15)(Table 3).
Type A aneurysms
Among a total of 25 aneurysms after coarctation repair, a significant proportion (n = 8) developed in the ascending aorta (32%; type A), with extensive aortic widening of 88 ± 31 mm; 6 of them revealed moderate or severe aortic valve regurgitation (75%), 2 were associated with dissection (25%) and 3 required emergent repair (37%). Although there were no early or late fatalities 110 ± 31 months after aneurysmal resection, the rate of postsurgical five-year actuarial freedom from cardiovascular events was only 75 ± 15%.
Type A aneurysms were independently associated with the presence of a CBAV on multivariate analysis (Table 4). The prevalence of a bicuspid aortic valve in our control group was 40%, similar to other studies, although seven bicuspid aortic valves in eight type A aneurysms clearly exceeded a frequency of 21% to 52% reported in coarctation unassociated with Marfan syndrome or Turner’s syndrome (7,11,23). Bicuspid aortic valves were often associated with aneurysms, aortic valve leakage or stenosis after coarctation repair (7,11,24). A stenotic aortic valve, however, was not observed in our study. Marked media degeneration was observed in six of seven patients with a type A aneurysm (86%), exceeding a frequency of 50% reported in local aneurysms after coarctation repair (4). Because formation of postsurgical aneurysms was related to advanced age at coarctation repair, chronic systemic hypertension preoperatively may have propelled media degeneration long before widening of the aortic diameter. This notion was supported by a high preoperative peak systolic pressure gradient of 74 ± 21 mm Hg associated with a risk of type A aneurysmal formation. Thus, long-standing, uncorrected severe coarctation was likely to increase the inherent risk of a bicuspid aortic valve to develop aneurysm and dissection of the aortic root.
Seventeen of 25 postsurgical aneurysms developed at the site of previous coarctation repair (68%). The present analysis provided statistical confirmation that use of a patch graft was independently associated with late local aneurysmal formation (1–15). Moreover, the current analysis is the first to reveal that advanced age at coarctation surgery (above 13.5 years according to ROC curve analysis) was an independent predictor of late aneurysmal formation. However, different anatomic features of aneurysms indicated various mechanisms of development. Although false local aneurysms arose from suture lines (6 of 9 patients) or at isthmic restenosis (3 of 9 patients), true local aneurysms usually formed opposite to a patch graft (Table 1). In contrast to previous series, however, Dacron was rarely employed (3,9), and silk sutures (1,4,7)or transverse aortic arch hypoplasia (11,12)was not encountered in our study.
Aneurysmal formation is common after coarctation repair and is not exclusively related to patch graft repair. Most notably, postsurgical aneurysms in the ascending aorta (type A) are completely unrelated to surgical techniques. Proximal aneurysms have escaped attention in previous reports, but they represent one-third of aortic aneurysms after coarctation repair. With the high prevalence of severe aortic regurgitation, dissection and rupture, this condition can be life-threatening and continues to carry a high morbidity, even after successful repair. Patients with a bicuspid aortic valve, advanced age at coarctation repair and high preoperative systolic peak pressure gradients are most likely to develop proximal aneurysms. Therefore, regular follow-up imaging should be offered for timing of elective aortic root surgery. Local aneurysms develop at the site of previous coarctation repair and represent two-thirds of postsurgical aortic aneurysms. According to our prediction model, aneurysmal formation at the site of coarctation repair may be prevented by avoiding the patch graft technique and performing repair of coarctation before the age of 13.5 years.
- blood pressure
- congenitally bicuspid aortic valve
- receiver operating characteristic
- Received June 11, 2001.
- Revision received October 31, 2001.
- Accepted November 16, 2001.
- American College of Cardiology
- Koller M.,
- Rothlin M.,
- Senning A.
- Bromberg B.I.,
- Beekman R.H.,
- Rocchini A.P.,
- et al.
- Bogaert J.,
- Gewillig M.,
- Rademakers F.,
- et al.
- Parks W.J.,
- Ngo T.D.,
- Plauth W.H.,
- et al.
- Task Force on Blood Pressure Control in Children
- Ingelfinger J.R.
- Carvalho J.S.,
- Redington A.N.,
- Shinebourne E.A.,
- et al.