- Email: [email protected]
Predictive Factors for Cerebrovascular Accidents After Thoracic Endovascular Aortic Repair
Giovanni Mariscalco, MD, PhD, Gabriele Piffaretti, MD, Matteo Tozzi, MD, Alessandro Bacuzzi, MD, Giampaolo Carrafiello, MD, Andrea Sala, MD, and Patrizio Castelli, MD, FACS Department of Surgical Sciences—Cardiac Surgery; Department of Surgical Sciences—Vascular Surgery; Anaesthesia and Palliative Care; and Department of Radiology—Interventional Radiology, Varese University Hospital, University of Insubria, Varese, Italy
Background. Cerebrovascular accidents are devastating and worrisome complications after thoracic endovascular aortic repair. The aim of this study was to determine cerebrovascular accident predictors after thoracic endovascular aortic repair. Methods. Between January 2001 and June 2008, 76 patients treated with thoracic endovascular aortic repair were prospectively enrolled. The study cohort included 61 men; mean age was 65.4 ⴞ 16.8 years. All patients underwent a specific neurologic assessment on an hourly basis postoperatively to detect neurologic deficits. Cerebrovascular accidents were diagnosed on the basis of physical examination, tomography scan or magnetic resonance imaging, or autopsy. Results. Cerebrovascular accidents occurred in 8 (10.5%) patients, including 4 transient ischemic attack and 4 major strokes. Four cases were observed within the first 24-hours. Multivariable analysis revealed that anatomic
incompleteness of the Willis circle (odds ratio [OR] 17.19, 95% confidence interval [CI] 2.10 to 140.66), as well as the presence of coronary artery disease (OR 6.86, 95 CI% 1.18 to 40.05), were independently associated with postoperative cerebrovascular accident development. Overall hospital mortality was 9.2%, with no significant difference for patients hit by cerebrovascular accidents (25.0% vs 7.3%, p ⴝ 0.102). Conclusions. Preexisting coronary artery disease, reflecting a severe diseased aorta and anomalies of Willis circle are independent cerebrovascular accident predictors after thoracic endovascular aortic repair procedures. A careful evaluation of the arch vessels and cerebral vascularization should be mandatory for patients suitable for thoracic endovascular aortic repair.
C
matous degree as well as the length of the procedure, and coverage of the left subclavian artery [5, 7, 9 –11]. The identification of CVAs predictors may contribute to explain the etiology, to optimize the risk stratification for this devastating complication, and finally improve procedural techniques in order to prevent and decrease the CVA incidence after TEVAR [9]. The aim of this study was to identify the CVAs determinants after TEVAR.
erebrovascular accidents (CVAs), and in particular stroke, have been reported to occur in 2% to 11% of the patients undergoing ascending-arch aortic repair, and in over 5% after descending or thoracoabdominal aortic operations [1– 4]. With the advent of endovascular repair, several studies have documented improved survival rates. However, despite endograft advances and technical improvements, a disturbing number of patients still have prolonged, complicated courses; in particular, a disappointing stroke rate of 3% to 9% has also been reported after thoracic endovascular aortic repair (TEVAR) [5–7]. Atherosclerotic disease of the thoracic aorta was renowned as one of the most important CVA risk factors after aortic open operations [8]. Underlying mechanisms of CVAs after TEVAR remain unclear though previous studies have identified a variety of factors associated with strokes, such as the extent of aortic disease, the atheroAccepted for publication Aug 11, 2009. Address correspondence to Dr Piffaretti, Department of Surgical SciencesVascular Surgery, Varese University Hospital, University of Insubria, Viale Guicciardini 9, Varese, 21100, Italy; e-mail: gabriele.piffaretti@ uninsubria.it.
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;88:1877– 81) © 2009 by The Society of Thoracic Surgeons
Material and Methods Study Population Between January 2001 and June 2008 all consecutive patients undergoing TEVAR of the descending thoracic or thoracoabdominal aorta were enrolled into the study. All clinical and procedural data were prospectively collected and recorded onto a computerized database registry that remained consistent over the study period. Data entry was managed by physicians and anesthesiologists involved in patient care. In addition, demographic characteristics, clinical details, and perioperative variables were supplemented by review charts, operative notes, and diagnostic reports. The study group included 61 males and 15 females; mean age was 65.4 ⫾ 16.8 years (range, 17 to 87 years). 0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.08.020
ADULT CARDIAC
Predictive Factors for Cerebrovascular Accidents After Thoracic Endovascular Aortic Repair
1878
MARISCALCO ET AL CVAS AFTER TEVAR
ADULT CARDIAC
The underlying aortic pathologies included atherosclerotic descending thoracic aortic aneurysms (n ⫽ 35, ruptured n ⫽ 3), chronic type B dissecting aneurysm or complicated acute type B dissection (n ⫽ 16), penetrating ulcers (n ⫽ 14, ruptured n ⫽ 2), and traumatic pseudoaneurysms (n ⫽ 11, chronic n ⫽ 3). The study protocol was in compliance with the local Institutional Review Board and received full approval. All patients gave their consent to participate. The authors had full access to, and take full responsibility for, the integrity of the data. All authors have read and agree to the manuscript as written.
Patient Management All elective patients underwent preoperative evaluation with transthoracic echocardiography, whereas computedtomography angiography of the brain was performed in every patient in order to assess the integrity of the Willis circle and the dominance of the vertebral arteries. Briefly, intervention was performed under general anesthesia with orotracheal intubation; short-term antibiotic prophylaxis with vancomicyn (2 gr twice a day) and endovenous heparinization (70 units/kg) were administered on a routine basis. Cerebrospinal fluid drainage was used selectively (extensive aortic coverage, previous or synchronous abdominal aortic repair). Transcranial-Doppler was performed in every case of arch vessels debranching. In addition, intraoperative transesophageal echocardiography was also used for all dissection cases, but not for the aneurysmatic cases. Postoperative neurologic assessments were performed on an hourly basis by the surgeons to detect neurologic deficits. For every minimal suspect of neurologic deficit, a full neurologic examination was promptly performed by a neurologist. However, CVA was diagnosed on the basis of physical examination, tomographic scan, magnetic resonance imaging, or autopsy. Three different devices were used in our experience: Talent/Valiant (Medtronic Vascular-Santa Rosa, Santa Rosa, CA), Excluder/TAG (W. L. Gore and Associates, Flagstaff, AZ), and TX-1/TX-2 (Cook, Bloomington, IN). Generally, endograft with bare stents were not used for traumatic aneurysms or dissections. Controlled hypotension was used during the deployment of the endograft; the proximal attachment site of the endograft was routinely ballooned except for selective cases.
Ann Thorac Surg 2009;88:1877– 81
using a modified scale (I to IV levels) based on a previously described grade for atheromatous disease that has been shown to correlate with stroke risk [13]. The extent of the disease was classified into three categories as described by Estrera and colleagues [14]: extent 1 included the proximal third of the descending thoracic aorta, extent 2 targeted the inferior third, extent 3 was considered the entire descending thoracic aorta. Primary technical success rate was defined as successful deployment of the endograft without any type of endoleak at the final angiogram. The CVAs included both stroke or transient ischemic attack (TIA): stroke was defined as any new onset of a focal neurologic impairment of sudden onset, and lasting more than 24 hours (or leading to death), whereas TIA was considered a new focal neurologic symptom but lasting less than 24 hours. All focal neurologic deficit was confirmed by computed tomography, together with a full neurologic examination by a neurologist.
Data Analysis Clinical data were prospectively recorded and tabulated with Microsoft Excel (Microsoft Corp, Redmond, WA). Continuous variables were tested for normal distribution by the Kolmogorov-Smirnov test and compared between groups with the unpaired Student t test for normally distributed values; otherwise, the Mann-Whitney U test was employed. In case of dichotomous variables, group differences were examined by 2 or Fisher exact tests as appropriate. A stepwise logistic regression model was developed to identify patients and procedural variables associated with postoperative cerebrovascular accidents. The model was built using variables that demonstrated a p value less than 0.20 in univariable analysis. The strength of the association of variables with CVA was estimated by calculating the odds ratio (OR) and 95% confidence intervals (CI). The calibration of the model was obtained by calculating the Hosmer-Lemeshow goodness-of-fit test. Results are expressed as mean ⫾ SD for continuous variables and frequencies for the categoric ones. A 2-sided p value less than 0.05 was considered statistically significant. Statistical analysis was computed with SPSS, release 16.0 for Windows (SPSS Inc, Chicago, IL).
Results Clinical Data
Definitions Emergent intervention was performed within the first 24 hours from admission. The elective procedures performed on the same admission day were not considered emergent. Coronary artery disease was defined as documented coronary lesions treated with prior coronary revascularization (percutaneous or surgical intervention) or medical therapy on the basis of a previous coronary angiography. Aortic arch map was defined accordingly to the classification proposed by Mitchell and colleagues [12]. The severity of the atheromatous disease was assessed by preoperative computed tomography scans of the aortic arch, and scored
Mena follow-up was 23.9 months (range, 3 to 84). Among the 76 TEVAR cases (elective, n ⫽ 46), CVAs occurred in 8 (10.5%) patients, including 4 transient ischemic attack and 4 major strokes. Clinical onset was detected within the first 24 hours in 4 subjects (stroke/TIA, n ⫽ 2/2), 3 within the first 48 hours (stroke/TIA, n ⫽ 1/2), and the remaining one stroke on the 19th postoperative day. Five CVA cases occurred in the left hemisphere (stroke/TIA, n ⫽ 2/3; 1 TIA was bilateral), while 4 patients had a bilateral event (stroke/TIA, n ⫽ 3/1). The posterior circulation was involved in 5 cases (stroke/TIA, n ⫽ 2/3), of which 4 subjects had the coverage of the left subclavian
Ann Thorac Surg 2009;88:1877– 81
MARISCALCO ET AL CVAS AFTER TEVAR
Table 1. Demographic Data and Comorbidities
Age Mean Range Male sex (n, %) Comorbidities: Hypertension (n, %) Diabetes (n, %) Dyslipidemia (n, %) COPD (n, %) Chronic renal failure (n, %) Preoperative CVA (n, %) Cardiac history: Chronic atrial fibrillation (n, %) Coronary artery disease (n, %) Prior cardiac surgery (n, %) EuroSCORE logistic (n) Emergent cases (n, %) a
Table 3. Aneurysm and Procedure Details
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
67.8 ⫾ 8.3 55–79 7 (87.5)
65.1 ⫾ 17.6 17–87 54 (79.4)
0.999
8 (100.0) 0 (0.0) 1 (12.5) 4 (50.0) 3 (37.5) 2 (25.0)
55 (80.9) 5 (7.4) 17 (25.0) 30 (44.1) 12 (17.6) 9 (13.2)
0.337 0.999 0.672 0.999 0.188 0.326
1 (12.5)
11 (16.2)
0.630
5 (62.5)
16 (23.5)
0.033
10 (14.7) 20.9 ⫾ 19.6 8 (15.1%)
0.587 0.397 0.097
p Value 0.773
0 (0.0) 16.3 ⫾ 15.0 0 (0%)
For continuous variables: mean and SD; for categoric variables: n (%).
COPD ⫽ chronic obstructive pulmonary disease; CVA ⫽ cerebrovascular accident; EuroSCORE ⫽ European system for cardiac operative risk evaluation.
artery. None of the strokes were hemorrhagic. Seven CVAs were classified as embolic and the remaining one as low-flow ischemic. In detail, this latter patient developed a basilar artery thrombosis after the unintentional transient coverage of the left common carotid artery during TEVAR of an extent 3 aneurysm. This operator error was suddenly rescued by distally steering the endograft using the occluding balloon. The autopsy and the second look of the preoperative computed tomographic Table 2. Operative and Postoperative Data Variablea
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
Operative data: Operation time 165.6 ⫾ 124.2 135.9 ⫾ 89.9 (minutes) Bleeding (mL) 370.0 ⫾ 299.67 328.6 ⫾ 407.2 Contrast medium 151.3 ⫾ 57.4 114.3 ⫾ 48.3 (mL) Postoperative data: Transfusion (n, %) 3 (37.5) 21 (30.9) RBC unit (n) 1.1 ⫾ 2.0 0.8 ⫾ 1.0 ICU stay (n, %) 2 (25.0) 23 (33.8) Hospitalization (days) 28.0 ⫾ 34.3 9.8 ⫾ 10.2 Hospital mortality (n, %) 4 (50.0) 8 (11.8) a
p 0.454 0.523 0.060
0.702 0.951 0.999 0.083 0.019
For continuous variables: mean and SD; for categoric variables: n (%).
CVA ⫽ cerebrovascular accident; red blood cells.
ICU ⫽ intensive care unit;
RBC ⫽
Variablea TAA characteristics: diameter (mm) length (mm) TAA etiology: Atherosclerotic (n, %) Chronic dissection (n, %) Traumatic (n, %) Penetrating ulcer (n, %) Atheroma degree: Normal (n, %) ⬍3 mm (n, %) 3–5 mm (n, %) ⬎5 mm (n, %) EG characteristics: Diameter (mm) Coverage length (mm) Number of EG Free-flow (n, %) EG landing zone: LCA (n, %) LSA (n, %) prox DTA (n, %) dist DTA (n, %) LSA coverage (n, %) Debranching (n, %) a
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
6.1 ⫾ 1.5 17.4 ⫾ 9.3
5.5 ⫾ 1.9 13.4 ⫾ 10.6
5 (62.5) 1 (12.5) 0 (0.0) 2 (14.3)
30 (44.1) 15 (22.1) 11 (16.2) 12 (17.6)
0 (0.0) 4 (50.0) 3 (37.5) 1 (12.5)
40 (58.8) 21 (30.9) 7 (10.3) 0 (0.0)
37.0 ⫾ 3.0 24.1 ⫾ 10.6 1.6 ⫾ 0.5 3 (37.5)
35.6 ⫾ 4.3 18.3 ⫾ 17.9 1.3 ⫾ 0.6 15 (22.1)
1 (12.5) 17 (22.4) 4 (50.0) 0 (0.0) 4 (50.0) 1 (12.5)
9 (13.2) 14 (20.6) 35 (51.5) 10 (14.7) 26 (38.2) 11 (16.2)
p 0.183 0.161 0.505
⬍0.001
0.492 0.037 0.035 0.385 0.552
0.705 0.999
For continuous variables: mean and SD; for categoric variables: n (%).
CVA ⫽ cerebrovascular accident; DTA ⫽ descending thoracic aorta; EG ⫽ endograft; LCA ⫽ left carotid artery; LSA ⫽ left subclavian artery; TAA ⫽ thoracic aortic aneurysm.
imaging revealed that the thrombosis occurred on a preexisting, and preoperatively underestimated, stenosis of the same vessel. In the group of patients with Willis incompleteness (n ⫽ 6), no one had ispilateral CVA. Univariable analysis concerning CVAs is represented in Tables 1 and 2. Briefly, no differences were observed in age distribution (67.8 ⫾ 8.3 vs 65.1 ⫾ 17.6 years, p ⫽ 0.773) and gender predominance (male, 87.5 % vs 79.4%, p ⫽ 0.999) between patients affected by CVAs and patients without them. In addition, no differences were noted in terms of comorbidities as well as in the logistic European system for cardiac operative risk evaluation (16.3 ⫾ 15.0 vs 20.9 ⫾ 19.6, p ⫽ 0.397). However, patients with CVAs had a threefold incidence of coronary artery disease compared with patients without them (62.5% vs 23.5%, p ⫽ 0.033). Interestingly, patients with CVAs revealed a higher level anatomic incompleteness of the Willis circle (37.5% vs 4.4%, p ⫽ 0.014). The overall in-hospital mortality was 9.2%, with no significant difference for patients hit by CVAs (25.0% vs 7.3%, p ⫽ 0.102). These patients had also a longer length of hospital stay, although it was not statistically significant (p ⫽ 0.083). Requirement of transfusion and intensive care unit stay were similar in the two groups (p ⫽ 0.702 and p ⫽ 0.999, respectively).
ADULT CARDIAC
Variablea
1879
1880
MARISCALCO ET AL CVAS AFTER TEVAR
Morphologic and Procedural Data
ADULT CARDIAC
No differences between groups were detected in terms of thoracic aortic disease etiology (P ⫽ 0.505), aortic diseased diameter and extension (p ⫽ 0.183 and p ⫽ 0.161, respectively), as wells as the endograft landing zone (p ⫽ 0.552; Table 3). In contrast, the atheroma degree involving the aortic arch, the length of the covered aorta, and the number of endografts were all significantly different between the two groups (p ⬍ 0.001, p ⫽ 0.037, and p ⫽ 0.035, respectively). Of note, aortic arch involvement (p ⫽ 0.148), endograft with a proximal free-flow (p ⫽ 0.385), coverage of the left subclavian artery (p ⫽ 0.705), or debranching (p ⫽ .999), did not achieve statistical significance (Table 3). The operation time was also similar between the two groups (165.6 ⫾ 124.2 vs 135.9 ⫾ 89.9 minutes, p ⫽ 0.454). Finally, the interaction among coronary artery disease, anomalies in the Willis circle, and atheroma degree were tested and found significant at the univariable level. Patients with a documented coronary artery disease revealed a higher atheroma degree (41.7% vs 15.0%, p ⫽ 0.011) as well as subjects with a documented anatomic incompleteness of the Willis circle (50.0% vs 7.1%, p ⫽ 0.014).
Multivariable Analysis Multivariable analysis showed that anatomic incompleteness of the Willis circle (OR 17.19, 95% CI 2.10 to 140.66; p ⫽ 0.008), as well as the presence of coronary artery disease (OR 6.86, 95 CI% 1.18 to 40.05; p ⫽0.032), were independently associated with postoperative CVAs. The interaction coronary artery disease–atheroma degree was not significant at multivariable level (OR 1.67; 95% CI 0.36 to 7.69; p ⫽ 0.513). With all patients considered, the variables of the model correctly predicted 92% of the observed cases. The Hosmer-Lemeshow goodness-of-fit test was not significant for lack of fit (p ⫽ 0.674), indicating that there was not statistically significant departure from a perfect fit. In addition, CVA occurrence was also an independent predictor of hospital mortality (OR 14.71, 95% CI 2.05 to 105.83).
Comment Neurologic complications and stroke are still considered the most devastating and frequently fatal complications after open thoracic aortic surgery [1– 4]. More recently, data coming from multicenter trials or single-center experiences [5, 9, 15] reported that stroke was the most common neurologic complication also after TEVAR, somewhat higher than spinal cord ischemia, leading to a relevant mortality rate. The clinical analysis, the pattern of brain infarction, and the timing of CVAs, led most of the authors to conclude that cerebral embolization was the primary mechanism of strokes complicating TEVAR, in relation to the difficult advancement and manipulation of the devices in the aortic arch curvature [7, 10, 11, 16]. The stroke rate reported in the present study confirmed the results of the extensive experience published by Demers and colleagues [17]. These authors reported that their high 7% stroke rate was the major drawback over 103 TEVAR cases and suggested a possible causative mechanism to the primitive, first-generation de-
Ann Thorac Surg 2009;88:1877– 81
vices, which required manipulation of large and stiff introducers and delivery sheaths in diseased aortic arch curvatures. In their recent overview, Sullivan and Sundt III [5] reported a stroke incidence of 2.2%; in our experience stroke was the dominant neurologic complication with a similar incidence rate of 5.5%, but in contrast with other available studies the total CVAs accounted for an overall 10.5% rate [5, 6, 9, 10, 16]. This finding underlined that this type of complication still occurs with a disturbingly high frequency also with more recent devices, therefore reflecting the potential combination of different factors [9, 10, 18 –20]. The most important finding of the present study was the presence of coronary artery disease as an independent predictor of CVAs. Our data confirmed the existing experience coming from cardiac operations, and stressed that the combination of a severely diseased aorta with a prolonged arch endovascular handling should be considered a relevant risk condition and should make the operators aware for perioperative stroke. As a matter of fact, coronary artery disease reflected the atheroma degree as confirmed at the univariable level calculating the interaction pattern between these two variables. We suggest that coronary artery disease should be considered a more complex marker that incorporated the role of an underlying high-loaded atheromatous thoracic aorta, as was confirmed at the univariable level calculating the interaction pattern with the arch atheroma degree. This confirmed the interesting data from Gutsche and colleagues [9], who used preoperative computed tomography angiography and intraoperative transesophageal echocardiography to classify the severity of atherosclerotic disease in the aortic arch and proximal descending aorta, and noted that a computed tomographic angiography atheroma grade of IV (⬎ 5 mm) was a significant predictor for perioperative stroke. Similarly, Buth and colleagues [7] suggested that prolonged instrumentation of the aortic arch in patients with severe atheromatous disease may disrupt vulnerable atheroma, causing embolic phenomena also in a later postoperative period. Another important result of our study was the incomplete integrity of the Willis circle as an independent predictor of postoperative CVAs. Few studies reported on the potential correlation between anatomic incompleteness of the Willis circle and postoperative CVAs. Urbanski and colleagues [21] suggested that an anatomic incompleteness of the Willis circle did not correlate with insufficient cerebral protection during unilateral cerebral perfusion for cardiac operations. Our findings deserve few comments: most of the patients with Willis incomplete integrity had also the highest grade of arch atheroma, one patient with Willis incomplete integrity had an unintentional coverage of the left common carotid artery, and a further one had a posterior stroke after a diseased left subclavian artery was debranched. Hence, we believe the anatomic incompleteness of the Willis should be questioned as a key factor for postoperative CVAs. The landing zone in TEAVR is another interesting issue because of its potential relationship with CVAs. The study of Feezor and colleagues [22] showed that proximal thoracic aortic repair involving landing zones “0 –2” was associated
with increased risk of stroke, particularly for posterior circulation-based strokes. In addition, the management of the left subclavian artery still remains a debated issue; Buth colleagues [7] reported a 26% rate of left subclavian artery overstenting without revascularization in the European collaborators on stent/graft techniques for aortic aneurysm repair series with no significant correlation with postoperative stroke. This finding was in contrast with the Talent registry; the left subclavian artery occlusion without previous revascularization (59.3%) was reported as the only significant CVA predictor [11]. In our analysis debranching of the aortic arch vessels was not associated to an increased risk of CVAs, and according to several authors, in our practice the intentional coverage of the left subclavian artery was anticipated by revascularization only in selected cases [23]. Although half of our strokes had the left subclavian artery covered, two of them had bilateral strokes, and one of the patients had preventive left subclavian artery revascularization. This finding was similar to the results of the Gore TAG trial, in which four of the five patients who developed CVAs had left carotid-left subclavian artery bypass, and out of 28 patients having carotid-left subclavian artery intervention 4 had strokes (14%) compared with 1% of patients who did not require this adjunctive operation [5, 6]. Again, we believe that postoperative CVAs should be considered as multifactorial in origin; therefore, debranching and left subclavian artery revascularization should be selectively performed. This study is limited in its interpretation because data came from a single center only, and further it is limited in statistical terms from the different number of observations between groups. Although data are consonant with the ones reported by other authors [9, 22], a type II error must be considered in view of the limited number of patients in the CVA groups. A further limitation is the absence of a systematic postoperative imaging, in order to detect additional subclinical brain accidents. In conclusion, our study revealed that coronary artery disease, potentially reflecting a severely diseased aorta, and incompleteness of the Willis circle, are independent predictors of CVAs after TEVAR. A careful evaluation of the arch vessels and cerebral vascularization should be performed for patients who are scheduled to receive TEVAR, in order to optimize the risk of intervention.
References 1. Goldstein LJ, Davies RR, Rizzo JA, et al. Stroke in surgery of the thoracic aorta: incidence, impact, etiology, and prevention. J Thorac Cardiovasc Surg 2001;122:935– 45. 2. Okita Y, Ando M, Minatoya K, Kitamura S, Takamoto S, Nakajima N. Predictive factors for mortality and cerebral complications in arteriosclerotic aneurysm of the aortic arch. Ann Thorac Surg 1999;67:72– 8. 3. Kouchoukos NT, Masetti P, Rokkas CK, Murphy SF. Hypothermic cardiopulmonary bypass and circulatory arrest for operations on the descending thoracic and thoracoabdominal aorta. Ann Thorac Surg 2002;74:S1885– 8. 4. Kawaharada N, Morishita K, Fukada J, et al. Stroke in surgery of the arteriosclerotic descending thoracic aortic aneurysms: influence of cross-clamping technique of the aorta. Eur J Cardiothorac Surg 2005;27:622–5.
MARISCALCO ET AL CVAS AFTER TEVAR
1881
5. Sullivan TM, Sundt TM, 3rd. Complications of thoracic aortic endografts: spinal cord ischemia and stroke. J Vasc Surg 2006;43(suppl A):85A– 8A. 6. Makaroun MS, Dillavou ED, Wheatley GH, Cambria RP. Gore TAG Investigators. Five-year results of endovascular treatment with the Gore TAG device compared with open repair of thoracic aortic aneurysms. J Vasc Surg 2008;47:912– 8. 7. Buth J, Harris PL, Hobo R, et al. Neurologic complications associated with endovascular repair of thoracic aortic pathology: Incidence and risk factors. a study from the European Collaborators on Stent/Graft Techniques for Aortic Aneurysm Repair (EUROSTAR) registry. J Vasc Surg 2007;46:1103–11. 8. Girardi LN, Krieger KH, Mack CA, Lee LY, Tortolani AJ, Isom OW. Does cross-clamping the arch increase the risk of descending thoracic and thoracoabdominal aneurysm repair? Ann Thorac Surg 2005;79:133– 8. 9. Gutsche JT, Cheung AT, McGarvey ML, et al. Risk factors for perioperative stroke after thoracic endovascular aortic repair. Ann Thorac Surg 2007;84:1195–1200. 10. Khoynezhad A, Donayre CE, Bui H, Kopchok GE, Walot I, White RA. Risk factors of neurologic deficit after thoracic aortic endografting. Ann Thorac Surg 2007;83:S882–92. 11. Fattori R, Nienaber CA, Rousseau H, et al. Results of endovascular repair of the thoracic aorta with the Talent Thoracic stent graft: the Talent Thoracic Retrospective Registry. J Thorac Cardiovasc Surg 2006;132:332–9. 12. Mitchell RS, Ishimaru S, Ehrlich MP, et al. First International Summit on Thoracic Aortic Endografting: roundtable on thoracic aortic dissection as an indication for endografting. J Endovasc Ther 2002;9(suppl II):II98 –105. 13. Katz ES, Tunick PA, Rusinek H, Ribakove G, Spencer FC, Kronzon I. Protruding aortic atheromas predict stroke in elderly patients undergoing cardiopulmonary bypass: experience with intraoperative transesophageal echocardiography. J Am Coll Cardiol 1992;20:70 –7. 14. Estrera AL, Rubenstein FS, Miller CC 3rd, Huynh TT, Letsou GV, Safi HJ. Descending thoracic aortic aneurysm: surgical approach and treatment using the adjuncts cerebrospinal fluid drainage and distal aortic perfusion. Ann Thorac Surg 2001;72:481– 6. 15. Cambria RP, Clouse WD, Davison JK, Dunn PF, Corey M, Dorer D. Thoracoabdominal aneurysm repair: results with 337 operations performed over a 15-year interval. Ann Surg 2002;236:471–9. 16. Scharrer-Pamler R, Kotsis T, Kapfer X, Gorich J, Orend KH, Sunder-Plassmann L. Complications after endovascular treatment of thoracic aortic aneurysms. J Endovasc Ther 2003;10:711– 8. 17. Demers P, Miller DC, Mitchell RS, et al. Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts. J Thorac Cardiovasc Surg 2004;127:664 –73. 18. Djaiani GN. Aortic arch atheroma: stroke reduction in cardiac surgical patients. Semin Cardiothorac Vasc Anesth 2006;10:143–57. 19. Bladin CF, Bingham L, Grigg L, Yapanis AG, Gerraty R, Davis SM. Transcranial Doppler detection of microemboli during percutaneous transluminal coronary angioplasty. Stroke 1998;29:2367–70. 20. Braekken SK, Endresen K, Russell D, Brucher R, Kjekshus J. Influence of guidewire and catheter type on the frequency of cerebral microembolic signals during left heart catheterization. Am J Cardiol 1998;82:632–7. 21. Urbanski PP, Lenos A, Blume JC, et al. Does anatomical completeness of the circle of Willis correlate with sufficient cross-perfusion during unilateral cerebral perfusion? Eur J Cardiothorac Surg 2008;33:402– 8. 22. Feezor RJ, Martin TD, Hess PJ, et al. Risk factors for perioperative stroke during thoracic endovascular aortic repairs (TEVAR). J Endovasc Ther 2007;14:568 –73. 23. Melissano G, Bertoglio L, Civilini E, et al. Results of thoracic endovascular grafting in different aortic segments. J Endovasc Ther 2007;14:150 –7.
ADULT CARDIAC
Ann Thorac Surg 2009;88:1877– 81
Background. Cerebrovascular accidents are devastating and worrisome complications after thoracic endovascular aortic repair. The aim of this study was to determine cerebrovascular accident predictors after thoracic endovascular aortic repair. Methods. Between January 2001 and June 2008, 76 patients treated with thoracic endovascular aortic repair were prospectively enrolled. The study cohort included 61 men; mean age was 65.4 ⴞ 16.8 years. All patients underwent a specific neurologic assessment on an hourly basis postoperatively to detect neurologic deficits. Cerebrovascular accidents were diagnosed on the basis of physical examination, tomography scan or magnetic resonance imaging, or autopsy. Results. Cerebrovascular accidents occurred in 8 (10.5%) patients, including 4 transient ischemic attack and 4 major strokes. Four cases were observed within the first 24-hours. Multivariable analysis revealed that anatomic
incompleteness of the Willis circle (odds ratio [OR] 17.19, 95% confidence interval [CI] 2.10 to 140.66), as well as the presence of coronary artery disease (OR 6.86, 95 CI% 1.18 to 40.05), were independently associated with postoperative cerebrovascular accident development. Overall hospital mortality was 9.2%, with no significant difference for patients hit by cerebrovascular accidents (25.0% vs 7.3%, p ⴝ 0.102). Conclusions. Preexisting coronary artery disease, reflecting a severe diseased aorta and anomalies of Willis circle are independent cerebrovascular accident predictors after thoracic endovascular aortic repair procedures. A careful evaluation of the arch vessels and cerebral vascularization should be mandatory for patients suitable for thoracic endovascular aortic repair.
C
matous degree as well as the length of the procedure, and coverage of the left subclavian artery [5, 7, 9 –11]. The identification of CVAs predictors may contribute to explain the etiology, to optimize the risk stratification for this devastating complication, and finally improve procedural techniques in order to prevent and decrease the CVA incidence after TEVAR [9]. The aim of this study was to identify the CVAs determinants after TEVAR.
erebrovascular accidents (CVAs), and in particular stroke, have been reported to occur in 2% to 11% of the patients undergoing ascending-arch aortic repair, and in over 5% after descending or thoracoabdominal aortic operations [1– 4]. With the advent of endovascular repair, several studies have documented improved survival rates. However, despite endograft advances and technical improvements, a disturbing number of patients still have prolonged, complicated courses; in particular, a disappointing stroke rate of 3% to 9% has also been reported after thoracic endovascular aortic repair (TEVAR) [5–7]. Atherosclerotic disease of the thoracic aorta was renowned as one of the most important CVA risk factors after aortic open operations [8]. Underlying mechanisms of CVAs after TEVAR remain unclear though previous studies have identified a variety of factors associated with strokes, such as the extent of aortic disease, the atheroAccepted for publication Aug 11, 2009. Address correspondence to Dr Piffaretti, Department of Surgical SciencesVascular Surgery, Varese University Hospital, University of Insubria, Viale Guicciardini 9, Varese, 21100, Italy; e-mail: gabriele.piffaretti@ uninsubria.it.
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;88:1877– 81) © 2009 by The Society of Thoracic Surgeons
Material and Methods Study Population Between January 2001 and June 2008 all consecutive patients undergoing TEVAR of the descending thoracic or thoracoabdominal aorta were enrolled into the study. All clinical and procedural data were prospectively collected and recorded onto a computerized database registry that remained consistent over the study period. Data entry was managed by physicians and anesthesiologists involved in patient care. In addition, demographic characteristics, clinical details, and perioperative variables were supplemented by review charts, operative notes, and diagnostic reports. The study group included 61 males and 15 females; mean age was 65.4 ⫾ 16.8 years (range, 17 to 87 years). 0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.08.020
ADULT CARDIAC
Predictive Factors for Cerebrovascular Accidents After Thoracic Endovascular Aortic Repair
1878
MARISCALCO ET AL CVAS AFTER TEVAR
ADULT CARDIAC
The underlying aortic pathologies included atherosclerotic descending thoracic aortic aneurysms (n ⫽ 35, ruptured n ⫽ 3), chronic type B dissecting aneurysm or complicated acute type B dissection (n ⫽ 16), penetrating ulcers (n ⫽ 14, ruptured n ⫽ 2), and traumatic pseudoaneurysms (n ⫽ 11, chronic n ⫽ 3). The study protocol was in compliance with the local Institutional Review Board and received full approval. All patients gave their consent to participate. The authors had full access to, and take full responsibility for, the integrity of the data. All authors have read and agree to the manuscript as written.
Patient Management All elective patients underwent preoperative evaluation with transthoracic echocardiography, whereas computedtomography angiography of the brain was performed in every patient in order to assess the integrity of the Willis circle and the dominance of the vertebral arteries. Briefly, intervention was performed under general anesthesia with orotracheal intubation; short-term antibiotic prophylaxis with vancomicyn (2 gr twice a day) and endovenous heparinization (70 units/kg) were administered on a routine basis. Cerebrospinal fluid drainage was used selectively (extensive aortic coverage, previous or synchronous abdominal aortic repair). Transcranial-Doppler was performed in every case of arch vessels debranching. In addition, intraoperative transesophageal echocardiography was also used for all dissection cases, but not for the aneurysmatic cases. Postoperative neurologic assessments were performed on an hourly basis by the surgeons to detect neurologic deficits. For every minimal suspect of neurologic deficit, a full neurologic examination was promptly performed by a neurologist. However, CVA was diagnosed on the basis of physical examination, tomographic scan, magnetic resonance imaging, or autopsy. Three different devices were used in our experience: Talent/Valiant (Medtronic Vascular-Santa Rosa, Santa Rosa, CA), Excluder/TAG (W. L. Gore and Associates, Flagstaff, AZ), and TX-1/TX-2 (Cook, Bloomington, IN). Generally, endograft with bare stents were not used for traumatic aneurysms or dissections. Controlled hypotension was used during the deployment of the endograft; the proximal attachment site of the endograft was routinely ballooned except for selective cases.
Ann Thorac Surg 2009;88:1877– 81
using a modified scale (I to IV levels) based on a previously described grade for atheromatous disease that has been shown to correlate with stroke risk [13]. The extent of the disease was classified into three categories as described by Estrera and colleagues [14]: extent 1 included the proximal third of the descending thoracic aorta, extent 2 targeted the inferior third, extent 3 was considered the entire descending thoracic aorta. Primary technical success rate was defined as successful deployment of the endograft without any type of endoleak at the final angiogram. The CVAs included both stroke or transient ischemic attack (TIA): stroke was defined as any new onset of a focal neurologic impairment of sudden onset, and lasting more than 24 hours (or leading to death), whereas TIA was considered a new focal neurologic symptom but lasting less than 24 hours. All focal neurologic deficit was confirmed by computed tomography, together with a full neurologic examination by a neurologist.
Data Analysis Clinical data were prospectively recorded and tabulated with Microsoft Excel (Microsoft Corp, Redmond, WA). Continuous variables were tested for normal distribution by the Kolmogorov-Smirnov test and compared between groups with the unpaired Student t test for normally distributed values; otherwise, the Mann-Whitney U test was employed. In case of dichotomous variables, group differences were examined by 2 or Fisher exact tests as appropriate. A stepwise logistic regression model was developed to identify patients and procedural variables associated with postoperative cerebrovascular accidents. The model was built using variables that demonstrated a p value less than 0.20 in univariable analysis. The strength of the association of variables with CVA was estimated by calculating the odds ratio (OR) and 95% confidence intervals (CI). The calibration of the model was obtained by calculating the Hosmer-Lemeshow goodness-of-fit test. Results are expressed as mean ⫾ SD for continuous variables and frequencies for the categoric ones. A 2-sided p value less than 0.05 was considered statistically significant. Statistical analysis was computed with SPSS, release 16.0 for Windows (SPSS Inc, Chicago, IL).
Results Clinical Data
Definitions Emergent intervention was performed within the first 24 hours from admission. The elective procedures performed on the same admission day were not considered emergent. Coronary artery disease was defined as documented coronary lesions treated with prior coronary revascularization (percutaneous or surgical intervention) or medical therapy on the basis of a previous coronary angiography. Aortic arch map was defined accordingly to the classification proposed by Mitchell and colleagues [12]. The severity of the atheromatous disease was assessed by preoperative computed tomography scans of the aortic arch, and scored
Mena follow-up was 23.9 months (range, 3 to 84). Among the 76 TEVAR cases (elective, n ⫽ 46), CVAs occurred in 8 (10.5%) patients, including 4 transient ischemic attack and 4 major strokes. Clinical onset was detected within the first 24 hours in 4 subjects (stroke/TIA, n ⫽ 2/2), 3 within the first 48 hours (stroke/TIA, n ⫽ 1/2), and the remaining one stroke on the 19th postoperative day. Five CVA cases occurred in the left hemisphere (stroke/TIA, n ⫽ 2/3; 1 TIA was bilateral), while 4 patients had a bilateral event (stroke/TIA, n ⫽ 3/1). The posterior circulation was involved in 5 cases (stroke/TIA, n ⫽ 2/3), of which 4 subjects had the coverage of the left subclavian
Ann Thorac Surg 2009;88:1877– 81
MARISCALCO ET AL CVAS AFTER TEVAR
Table 1. Demographic Data and Comorbidities
Age Mean Range Male sex (n, %) Comorbidities: Hypertension (n, %) Diabetes (n, %) Dyslipidemia (n, %) COPD (n, %) Chronic renal failure (n, %) Preoperative CVA (n, %) Cardiac history: Chronic atrial fibrillation (n, %) Coronary artery disease (n, %) Prior cardiac surgery (n, %) EuroSCORE logistic (n) Emergent cases (n, %) a
Table 3. Aneurysm and Procedure Details
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
67.8 ⫾ 8.3 55–79 7 (87.5)
65.1 ⫾ 17.6 17–87 54 (79.4)
0.999
8 (100.0) 0 (0.0) 1 (12.5) 4 (50.0) 3 (37.5) 2 (25.0)
55 (80.9) 5 (7.4) 17 (25.0) 30 (44.1) 12 (17.6) 9 (13.2)
0.337 0.999 0.672 0.999 0.188 0.326
1 (12.5)
11 (16.2)
0.630
5 (62.5)
16 (23.5)
0.033
10 (14.7) 20.9 ⫾ 19.6 8 (15.1%)
0.587 0.397 0.097
p Value 0.773
0 (0.0) 16.3 ⫾ 15.0 0 (0%)
For continuous variables: mean and SD; for categoric variables: n (%).
COPD ⫽ chronic obstructive pulmonary disease; CVA ⫽ cerebrovascular accident; EuroSCORE ⫽ European system for cardiac operative risk evaluation.
artery. None of the strokes were hemorrhagic. Seven CVAs were classified as embolic and the remaining one as low-flow ischemic. In detail, this latter patient developed a basilar artery thrombosis after the unintentional transient coverage of the left common carotid artery during TEVAR of an extent 3 aneurysm. This operator error was suddenly rescued by distally steering the endograft using the occluding balloon. The autopsy and the second look of the preoperative computed tomographic Table 2. Operative and Postoperative Data Variablea
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
Operative data: Operation time 165.6 ⫾ 124.2 135.9 ⫾ 89.9 (minutes) Bleeding (mL) 370.0 ⫾ 299.67 328.6 ⫾ 407.2 Contrast medium 151.3 ⫾ 57.4 114.3 ⫾ 48.3 (mL) Postoperative data: Transfusion (n, %) 3 (37.5) 21 (30.9) RBC unit (n) 1.1 ⫾ 2.0 0.8 ⫾ 1.0 ICU stay (n, %) 2 (25.0) 23 (33.8) Hospitalization (days) 28.0 ⫾ 34.3 9.8 ⫾ 10.2 Hospital mortality (n, %) 4 (50.0) 8 (11.8) a
p 0.454 0.523 0.060
0.702 0.951 0.999 0.083 0.019
For continuous variables: mean and SD; for categoric variables: n (%).
CVA ⫽ cerebrovascular accident; red blood cells.
ICU ⫽ intensive care unit;
RBC ⫽
Variablea TAA characteristics: diameter (mm) length (mm) TAA etiology: Atherosclerotic (n, %) Chronic dissection (n, %) Traumatic (n, %) Penetrating ulcer (n, %) Atheroma degree: Normal (n, %) ⬍3 mm (n, %) 3–5 mm (n, %) ⬎5 mm (n, %) EG characteristics: Diameter (mm) Coverage length (mm) Number of EG Free-flow (n, %) EG landing zone: LCA (n, %) LSA (n, %) prox DTA (n, %) dist DTA (n, %) LSA coverage (n, %) Debranching (n, %) a
CVA (n ⫽ 8)
No CVA (n ⫽ 68)
6.1 ⫾ 1.5 17.4 ⫾ 9.3
5.5 ⫾ 1.9 13.4 ⫾ 10.6
5 (62.5) 1 (12.5) 0 (0.0) 2 (14.3)
30 (44.1) 15 (22.1) 11 (16.2) 12 (17.6)
0 (0.0) 4 (50.0) 3 (37.5) 1 (12.5)
40 (58.8) 21 (30.9) 7 (10.3) 0 (0.0)
37.0 ⫾ 3.0 24.1 ⫾ 10.6 1.6 ⫾ 0.5 3 (37.5)
35.6 ⫾ 4.3 18.3 ⫾ 17.9 1.3 ⫾ 0.6 15 (22.1)
1 (12.5) 17 (22.4) 4 (50.0) 0 (0.0) 4 (50.0) 1 (12.5)
9 (13.2) 14 (20.6) 35 (51.5) 10 (14.7) 26 (38.2) 11 (16.2)
p 0.183 0.161 0.505
⬍0.001
0.492 0.037 0.035 0.385 0.552
0.705 0.999
For continuous variables: mean and SD; for categoric variables: n (%).
CVA ⫽ cerebrovascular accident; DTA ⫽ descending thoracic aorta; EG ⫽ endograft; LCA ⫽ left carotid artery; LSA ⫽ left subclavian artery; TAA ⫽ thoracic aortic aneurysm.
imaging revealed that the thrombosis occurred on a preexisting, and preoperatively underestimated, stenosis of the same vessel. In the group of patients with Willis incompleteness (n ⫽ 6), no one had ispilateral CVA. Univariable analysis concerning CVAs is represented in Tables 1 and 2. Briefly, no differences were observed in age distribution (67.8 ⫾ 8.3 vs 65.1 ⫾ 17.6 years, p ⫽ 0.773) and gender predominance (male, 87.5 % vs 79.4%, p ⫽ 0.999) between patients affected by CVAs and patients without them. In addition, no differences were noted in terms of comorbidities as well as in the logistic European system for cardiac operative risk evaluation (16.3 ⫾ 15.0 vs 20.9 ⫾ 19.6, p ⫽ 0.397). However, patients with CVAs had a threefold incidence of coronary artery disease compared with patients without them (62.5% vs 23.5%, p ⫽ 0.033). Interestingly, patients with CVAs revealed a higher level anatomic incompleteness of the Willis circle (37.5% vs 4.4%, p ⫽ 0.014). The overall in-hospital mortality was 9.2%, with no significant difference for patients hit by CVAs (25.0% vs 7.3%, p ⫽ 0.102). These patients had also a longer length of hospital stay, although it was not statistically significant (p ⫽ 0.083). Requirement of transfusion and intensive care unit stay were similar in the two groups (p ⫽ 0.702 and p ⫽ 0.999, respectively).
ADULT CARDIAC
Variablea
1879
1880
MARISCALCO ET AL CVAS AFTER TEVAR
Morphologic and Procedural Data
ADULT CARDIAC
No differences between groups were detected in terms of thoracic aortic disease etiology (P ⫽ 0.505), aortic diseased diameter and extension (p ⫽ 0.183 and p ⫽ 0.161, respectively), as wells as the endograft landing zone (p ⫽ 0.552; Table 3). In contrast, the atheroma degree involving the aortic arch, the length of the covered aorta, and the number of endografts were all significantly different between the two groups (p ⬍ 0.001, p ⫽ 0.037, and p ⫽ 0.035, respectively). Of note, aortic arch involvement (p ⫽ 0.148), endograft with a proximal free-flow (p ⫽ 0.385), coverage of the left subclavian artery (p ⫽ 0.705), or debranching (p ⫽ .999), did not achieve statistical significance (Table 3). The operation time was also similar between the two groups (165.6 ⫾ 124.2 vs 135.9 ⫾ 89.9 minutes, p ⫽ 0.454). Finally, the interaction among coronary artery disease, anomalies in the Willis circle, and atheroma degree were tested and found significant at the univariable level. Patients with a documented coronary artery disease revealed a higher atheroma degree (41.7% vs 15.0%, p ⫽ 0.011) as well as subjects with a documented anatomic incompleteness of the Willis circle (50.0% vs 7.1%, p ⫽ 0.014).
Multivariable Analysis Multivariable analysis showed that anatomic incompleteness of the Willis circle (OR 17.19, 95% CI 2.10 to 140.66; p ⫽ 0.008), as well as the presence of coronary artery disease (OR 6.86, 95 CI% 1.18 to 40.05; p ⫽0.032), were independently associated with postoperative CVAs. The interaction coronary artery disease–atheroma degree was not significant at multivariable level (OR 1.67; 95% CI 0.36 to 7.69; p ⫽ 0.513). With all patients considered, the variables of the model correctly predicted 92% of the observed cases. The Hosmer-Lemeshow goodness-of-fit test was not significant for lack of fit (p ⫽ 0.674), indicating that there was not statistically significant departure from a perfect fit. In addition, CVA occurrence was also an independent predictor of hospital mortality (OR 14.71, 95% CI 2.05 to 105.83).
Comment Neurologic complications and stroke are still considered the most devastating and frequently fatal complications after open thoracic aortic surgery [1– 4]. More recently, data coming from multicenter trials or single-center experiences [5, 9, 15] reported that stroke was the most common neurologic complication also after TEVAR, somewhat higher than spinal cord ischemia, leading to a relevant mortality rate. The clinical analysis, the pattern of brain infarction, and the timing of CVAs, led most of the authors to conclude that cerebral embolization was the primary mechanism of strokes complicating TEVAR, in relation to the difficult advancement and manipulation of the devices in the aortic arch curvature [7, 10, 11, 16]. The stroke rate reported in the present study confirmed the results of the extensive experience published by Demers and colleagues [17]. These authors reported that their high 7% stroke rate was the major drawback over 103 TEVAR cases and suggested a possible causative mechanism to the primitive, first-generation de-
Ann Thorac Surg 2009;88:1877– 81
vices, which required manipulation of large and stiff introducers and delivery sheaths in diseased aortic arch curvatures. In their recent overview, Sullivan and Sundt III [5] reported a stroke incidence of 2.2%; in our experience stroke was the dominant neurologic complication with a similar incidence rate of 5.5%, but in contrast with other available studies the total CVAs accounted for an overall 10.5% rate [5, 6, 9, 10, 16]. This finding underlined that this type of complication still occurs with a disturbingly high frequency also with more recent devices, therefore reflecting the potential combination of different factors [9, 10, 18 –20]. The most important finding of the present study was the presence of coronary artery disease as an independent predictor of CVAs. Our data confirmed the existing experience coming from cardiac operations, and stressed that the combination of a severely diseased aorta with a prolonged arch endovascular handling should be considered a relevant risk condition and should make the operators aware for perioperative stroke. As a matter of fact, coronary artery disease reflected the atheroma degree as confirmed at the univariable level calculating the interaction pattern between these two variables. We suggest that coronary artery disease should be considered a more complex marker that incorporated the role of an underlying high-loaded atheromatous thoracic aorta, as was confirmed at the univariable level calculating the interaction pattern with the arch atheroma degree. This confirmed the interesting data from Gutsche and colleagues [9], who used preoperative computed tomography angiography and intraoperative transesophageal echocardiography to classify the severity of atherosclerotic disease in the aortic arch and proximal descending aorta, and noted that a computed tomographic angiography atheroma grade of IV (⬎ 5 mm) was a significant predictor for perioperative stroke. Similarly, Buth and colleagues [7] suggested that prolonged instrumentation of the aortic arch in patients with severe atheromatous disease may disrupt vulnerable atheroma, causing embolic phenomena also in a later postoperative period. Another important result of our study was the incomplete integrity of the Willis circle as an independent predictor of postoperative CVAs. Few studies reported on the potential correlation between anatomic incompleteness of the Willis circle and postoperative CVAs. Urbanski and colleagues [21] suggested that an anatomic incompleteness of the Willis circle did not correlate with insufficient cerebral protection during unilateral cerebral perfusion for cardiac operations. Our findings deserve few comments: most of the patients with Willis incomplete integrity had also the highest grade of arch atheroma, one patient with Willis incomplete integrity had an unintentional coverage of the left common carotid artery, and a further one had a posterior stroke after a diseased left subclavian artery was debranched. Hence, we believe the anatomic incompleteness of the Willis should be questioned as a key factor for postoperative CVAs. The landing zone in TEAVR is another interesting issue because of its potential relationship with CVAs. The study of Feezor and colleagues [22] showed that proximal thoracic aortic repair involving landing zones “0 –2” was associated
with increased risk of stroke, particularly for posterior circulation-based strokes. In addition, the management of the left subclavian artery still remains a debated issue; Buth colleagues [7] reported a 26% rate of left subclavian artery overstenting without revascularization in the European collaborators on stent/graft techniques for aortic aneurysm repair series with no significant correlation with postoperative stroke. This finding was in contrast with the Talent registry; the left subclavian artery occlusion without previous revascularization (59.3%) was reported as the only significant CVA predictor [11]. In our analysis debranching of the aortic arch vessels was not associated to an increased risk of CVAs, and according to several authors, in our practice the intentional coverage of the left subclavian artery was anticipated by revascularization only in selected cases [23]. Although half of our strokes had the left subclavian artery covered, two of them had bilateral strokes, and one of the patients had preventive left subclavian artery revascularization. This finding was similar to the results of the Gore TAG trial, in which four of the five patients who developed CVAs had left carotid-left subclavian artery bypass, and out of 28 patients having carotid-left subclavian artery intervention 4 had strokes (14%) compared with 1% of patients who did not require this adjunctive operation [5, 6]. Again, we believe that postoperative CVAs should be considered as multifactorial in origin; therefore, debranching and left subclavian artery revascularization should be selectively performed. This study is limited in its interpretation because data came from a single center only, and further it is limited in statistical terms from the different number of observations between groups. Although data are consonant with the ones reported by other authors [9, 22], a type II error must be considered in view of the limited number of patients in the CVA groups. A further limitation is the absence of a systematic postoperative imaging, in order to detect additional subclinical brain accidents. In conclusion, our study revealed that coronary artery disease, potentially reflecting a severely diseased aorta, and incompleteness of the Willis circle, are independent predictors of CVAs after TEVAR. A careful evaluation of the arch vessels and cerebral vascularization should be performed for patients who are scheduled to receive TEVAR, in order to optimize the risk of intervention.
References 1. Goldstein LJ, Davies RR, Rizzo JA, et al. Stroke in surgery of the thoracic aorta: incidence, impact, etiology, and prevention. J Thorac Cardiovasc Surg 2001;122:935– 45. 2. Okita Y, Ando M, Minatoya K, Kitamura S, Takamoto S, Nakajima N. Predictive factors for mortality and cerebral complications in arteriosclerotic aneurysm of the aortic arch. Ann Thorac Surg 1999;67:72– 8. 3. Kouchoukos NT, Masetti P, Rokkas CK, Murphy SF. Hypothermic cardiopulmonary bypass and circulatory arrest for operations on the descending thoracic and thoracoabdominal aorta. Ann Thorac Surg 2002;74:S1885– 8. 4. Kawaharada N, Morishita K, Fukada J, et al. Stroke in surgery of the arteriosclerotic descending thoracic aortic aneurysms: influence of cross-clamping technique of the aorta. Eur J Cardiothorac Surg 2005;27:622–5.
MARISCALCO ET AL CVAS AFTER TEVAR
1881
5. Sullivan TM, Sundt TM, 3rd. Complications of thoracic aortic endografts: spinal cord ischemia and stroke. J Vasc Surg 2006;43(suppl A):85A– 8A. 6. Makaroun MS, Dillavou ED, Wheatley GH, Cambria RP. Gore TAG Investigators. Five-year results of endovascular treatment with the Gore TAG device compared with open repair of thoracic aortic aneurysms. J Vasc Surg 2008;47:912– 8. 7. Buth J, Harris PL, Hobo R, et al. Neurologic complications associated with endovascular repair of thoracic aortic pathology: Incidence and risk factors. a study from the European Collaborators on Stent/Graft Techniques for Aortic Aneurysm Repair (EUROSTAR) registry. J Vasc Surg 2007;46:1103–11. 8. Girardi LN, Krieger KH, Mack CA, Lee LY, Tortolani AJ, Isom OW. Does cross-clamping the arch increase the risk of descending thoracic and thoracoabdominal aneurysm repair? Ann Thorac Surg 2005;79:133– 8. 9. Gutsche JT, Cheung AT, McGarvey ML, et al. Risk factors for perioperative stroke after thoracic endovascular aortic repair. Ann Thorac Surg 2007;84:1195–1200. 10. Khoynezhad A, Donayre CE, Bui H, Kopchok GE, Walot I, White RA. Risk factors of neurologic deficit after thoracic aortic endografting. Ann Thorac Surg 2007;83:S882–92. 11. Fattori R, Nienaber CA, Rousseau H, et al. Results of endovascular repair of the thoracic aorta with the Talent Thoracic stent graft: the Talent Thoracic Retrospective Registry. J Thorac Cardiovasc Surg 2006;132:332–9. 12. Mitchell RS, Ishimaru S, Ehrlich MP, et al. First International Summit on Thoracic Aortic Endografting: roundtable on thoracic aortic dissection as an indication for endografting. J Endovasc Ther 2002;9(suppl II):II98 –105. 13. Katz ES, Tunick PA, Rusinek H, Ribakove G, Spencer FC, Kronzon I. Protruding aortic atheromas predict stroke in elderly patients undergoing cardiopulmonary bypass: experience with intraoperative transesophageal echocardiography. J Am Coll Cardiol 1992;20:70 –7. 14. Estrera AL, Rubenstein FS, Miller CC 3rd, Huynh TT, Letsou GV, Safi HJ. Descending thoracic aortic aneurysm: surgical approach and treatment using the adjuncts cerebrospinal fluid drainage and distal aortic perfusion. Ann Thorac Surg 2001;72:481– 6. 15. Cambria RP, Clouse WD, Davison JK, Dunn PF, Corey M, Dorer D. Thoracoabdominal aneurysm repair: results with 337 operations performed over a 15-year interval. Ann Surg 2002;236:471–9. 16. Scharrer-Pamler R, Kotsis T, Kapfer X, Gorich J, Orend KH, Sunder-Plassmann L. Complications after endovascular treatment of thoracic aortic aneurysms. J Endovasc Ther 2003;10:711– 8. 17. Demers P, Miller DC, Mitchell RS, et al. Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts. J Thorac Cardiovasc Surg 2004;127:664 –73. 18. Djaiani GN. Aortic arch atheroma: stroke reduction in cardiac surgical patients. Semin Cardiothorac Vasc Anesth 2006;10:143–57. 19. Bladin CF, Bingham L, Grigg L, Yapanis AG, Gerraty R, Davis SM. Transcranial Doppler detection of microemboli during percutaneous transluminal coronary angioplasty. Stroke 1998;29:2367–70. 20. Braekken SK, Endresen K, Russell D, Brucher R, Kjekshus J. Influence of guidewire and catheter type on the frequency of cerebral microembolic signals during left heart catheterization. Am J Cardiol 1998;82:632–7. 21. Urbanski PP, Lenos A, Blume JC, et al. Does anatomical completeness of the circle of Willis correlate with sufficient cross-perfusion during unilateral cerebral perfusion? Eur J Cardiothorac Surg 2008;33:402– 8. 22. Feezor RJ, Martin TD, Hess PJ, et al. Risk factors for perioperative stroke during thoracic endovascular aortic repairs (TEVAR). J Endovasc Ther 2007;14:568 –73. 23. Melissano G, Bertoglio L, Civilini E, et al. Results of thoracic endovascular grafting in different aortic segments. J Endovasc Ther 2007;14:150 –7.
ADULT CARDIAC
Ann Thorac Surg 2009;88:1877– 81