Author + information
- Received October 11, 2000
- Revision received March 15, 2001
- Accepted March 29, 2001
- Published online July 1, 2001.
- ↵*Reprint requests and correspondence:
Dr. Jan van der Linden, Department of Cardiothoracic Surgery and Anesthesiology, Karolinska Institutet, Huddinge University Hospital, SE-14186 Stockholm, Sweden
The aim of the study was to evaluate the risk from calcified atheromas in the ascending aorta, and the extent and topography of the disease in the development of stroke after cardiac surgery.
Postoperative stroke constitutes a serious problem in cardiac surgery, and atherosclerosis of the ascending aorta is an important risk factor.
Before surgical manipulation epiaortic echocardiographic ultrasound was performed to evaluate the ascending aorta in 921 consecutive patients undergoing cardiac surgery. The presence of calcification, location of atheroma, extent of the disease and clinical variables including postoperative stroke were recorded prospectively.
A total of 26.2% of the patients had atherosclerosis of the ascending aorta, and in 44.4% of them more than one of 12 possible segments was involved. Logistic regression showed that atherosclerotic disease in the ascending aorta was the most important predictive factor for postoperative stroke. The incidence of stroke was 1.8% in patients without atherosclerotic disease of the ascending aorta, and 8.7% in patients with the disease (p < 0.0001). Diabetes mellitus was also a predictive factor (p = 0.04). A new and unique finding of this study was that the middle-lateral segment is an independent predictive factor for postoperative stroke, with a relative risk of 26% (p = 0.04).
Patients with atheromatosis in the ascending aorta had an 8.7% incidence of postoperative stroke, in spite of minor surgical modifications. The risk depended on the presence, location and extent of the disease. Randomized trials evaluating alternative surgical strategies in coronary surgery are urgently needed in high risk patients.
Postoperative stroke remains one of the most serious problems in cardiac surgery, with atherosclerosis of the ascending aorta as an important risk factor (1–6). To date epiaortic ultrasound is the most accurate method for identification of atherosclerosis in the ascending aorta (7–10). The presence of atheromas may dictate modifications of the surgical technique, such as changes in the position of aortic cannula, cross-clamp, and proximal anastomoses, off-pump coronary artery bypass grafting (CABG) and replacement of the ascending aorta. However, not all surgeons are comfortable with off-pump CABG, and replacement of the ascending aorta may carry significant surgical risks even in experienced hands. The increased use of epiaortic ultrasound has caused uncertainty among surgeons regarding the implications of the findings and the necessity of surgical modifications. The aim of the present study was to evaluate the risk from calcified atheromas in the ascending aorta and the extent and topography of the disease in the development of postoperative stroke.
Patients and methods
The Hospital Ethical Committee approved the study, and informed written consent was obtained from all patients. We prospectively studied 921 consecutive patients undergoing elective cardiac surgery. After induction of anesthesia all patients were evaluated by an experienced anesthetist with transesophageal echo (TEE) with a multiplane probe (Hewlett Packard Sono 1500, Andover, Massachusetts) according to a standard protocol. Apart from the study of the heart, special attention was paid to the ascending aorta, aortic arch and descending aorta with regard to atherosclerotic changes. After a median sternotomy, the surgeon carefully examined the ascending aorta by digital palpation and by epiaortic ultrasound. Epiaortic ultrasound was performed with an ultrasound scanner (Site-Rite II, 9.0 MHz, Dymax Corporation, Pittsburgh, Pennsylvania). The probe’s advantage is the approximately 6-cm-long built-in hard-shell standoff, which results in a good resolution even of the surface in contact with the probe. A film of sterile gel was applied on the tip of the probe and a sterile plastic bag was wrapped around it. Two assessors, the surgeon and the anesthetist, interpreted the epiaortic ultrasound. In case of disagreement, a third assessor, without information about the conclusions of the first two, was asked for his opinion. Various minor procedural changes were adopted on the basis of the findings of the epiaortic ultrasound. These included changes in the position of the aortic cross-clamp, the placing of the aortic and cardioplegia cannulas and, in coronary cases, the site of the proximal anastomoses. Cases where deep hypothermic circulatory arrest was employed were excluded from the study. Clinical demographics were prospectively recorded (Table 1). Carotid artery disease was defined as either asymptomatic carotid stenosis or carotid bruit or carotid stenosis associated with a transient ischemic attack and/or a history of stroke. The mapping of the atherosclerotic disease in the ascending aorta was performed as follows: the ascending aorta was divided into three equal transverse (proximal, middle and distal third) and four equal longitudinal segments (anterior, posterior, lateral and medial, Fig. 1). The presence of ascending aortic atheroma (intimal thickening ≥0.5 mm) and calcification were recorded for every segment. Extent of disease was defined as the number of segments concomitantly involved. The preoperative (one to three days before the operation) and postoperative daily neurologic examination in the ward were performed by a surgical consultant. The first postoperative neurologic examination was performed in the intensive care unit the morning after the operation by a senior surgeon in cooperation with a senior anesthetist-intensivist. At the neurologic examination the following were checked and compared with the preoperative status: patients’ pupillary size and function, speech, orientation to time and place, memory function and power and symmetry of movements of feet, legs, hands, arms and facial muscles. The major outcome variable was postoperative stroke, defined as a clinically evident temporary or permanent new neurologic focal deficit during the same hospitalization. The neurologic deficit was confirmed by a computed tomography scan of the brain and clinically by a senior neurologist unaware of the study status.
The Kolmogorov-Smirnov test was used to compare the distribution of each variable with the normal distribution. The variable “age” did not follow normal distribution and therefore underwent a logarithmic transformation. Chi-square analysis was used to compare categorical variables. Logistic multiple regression analysis was used to identify the predictive value of various factors in the development of postoperative stroke. It was also used to calculate the relative risk of stroke for each of the previously stated segments of the ascending aorta, taking into account other factors that may influence the development of stroke. Results were expressed as mean ± standard error or median and 25th/75th percentiles. Differences were considered significant at a probability level of p < 0.05. Data were analyzed with SPSS version 8.0 statistical program (SPSS Inc., Chicago, Illinois).
The preoperative patient characteristics, epiaortic ultrasound and TEE data are summarized in Table 1. Atherosclerotic disease in the ascending aorta was detected in 26.2% of patients. The details of the epiaortic ultrasound findings are summarized in Table 2. Most often involved was the distal-anterior segment (48.4%). Adaptations of the surgical technique because of atherosclerotic disease in the ascending aorta included changes in the choice of site of the aortic cannula (26.1%), aortic cross-clamp (23.2%), proximal aortic-vein anastomoses (14.0%), cannula for cardioplegia (5.4%) and elimination of the need for proximal anastomoses (3.0%). Adaptation of the surgical technique was undertaken in 101 of the 921 patients with significant new findings on the epiaortic ultrasound not detectable on palpation.
Patients with atherosclerotic disease in the ascending aorta identified by epiaortic ultrasound had a higher incidence of atherosclerotic disease in the aortic arch (70.1% vs. 17.4%, p < 0.0001) and in the descending aorta (73.9% vs. 26.6%, p < 0.0001). They also had a higher incidence of carotid artery disease (18.3% vs. 9.3%, p < 0.0001) and intermittent claudication (9.6% vs. 5.8%, p = 0.05).
The incidence of postoperative stroke was 3.5%. Logistic regression showed that atherosclerotic disease in the ascending aorta as identified by epiaortic ultrasound was the most important predictive factor for postoperative stroke (Table 3). The incidence of stroke was 1.8% in patients without and 8.7% in patients with atherosclerotic disease in the ascending aorta (p < 0.001). Diabetes mellitus was also an independent predictive risk factor for stroke (p = 0.04).
In patients with atherosclerotic disease in the ascending aorta the presence or absence of calcification was not a predictive risk factor for stroke. The incidence of stroke without calcification was 7.6% (9/118) and with calcification 9.6% (10/104) (p = 0.64). Patients with more than six segments involved (more than half of the ascending aorta) were at a significantly higher risk of postoperative stroke than patients who had fewer than four segments involved (Table 4).
The incidence of postoperative stroke for every diseased segment is presented in Table 5. However, univariate analysis could not reliably predict the origin of the embolic material that caused the stroke. Because other segments were often involved, the embolic material might not necessarily have originated from the diseased segment. Table 5also shows the calculated incidence of postoperative stroke for single-segment disease. This incidence was calculated through a logistic regression analysis that took into consideration the presence of diabetes mellitus and the extent of the disease. The middle-lateral segment was an independent predictive factor for postoperative stroke, with a relative risk of 26% (p = 0.04). The proximal-lateral segment also showed a high relative risk for stroke, although it failed to reach statistical significance (26%, p = 0.08).
Our study shows that the presence of calcification in the ascending aorta does not predispose to postoperative stroke. It is possible that postoperative stroke results mainly from cerebral embolization of atheromatous debris and not from calcified material.
As one may expect, the extent of the atherosclerotic disease in the ascending aorta correlated with postoperative stroke. Thus, patients with atherosclerotic disease occupying more than 50% of the ascending aorta (more than 6 out of 12 possible segments) had a 33.3% incidence of postoperative stroke.
Earlier studies with epiaortic ultrasound have shown that the bulk of the disease is found in the upper distal part of the ascending aorta (1,4,11). As far as we know, the impact of the topography of the disease in the ascending aorta on the incidence of postoperative neurologic deficits has not been thoroughly studied. In the present study the distal anterior segment was the one most commonly affected by the disease. This is where surgeons usually place the aortic cannula. This finding definitely influenced the surgical technique, because the most frequent change consisted of the positioning of the arterial cannula away from that segment. The most frequently affected zones were the anterior and posterior aspects of the aorta, especially distally (Fig. 1). The reason for this remains unknown. The medial and lateral aspects, especially proximally, were involved in rather more extensive disease. Disease in the proximal third of the ascending aorta and in the mid-lateral segment was associated with a higher incidence of postoperative stroke. One may suspect that this was due to the association of this distribution with more extensive disease. However, a multivariate analysis taking into account the presence of diabetes and the extent of the disease, both of which were independent predictive factors for postoperative stroke, showed that disease in the mid-lateral segment was a significant independent risk factor for postoperative stroke. Blood flow pattern at the level of the aortic valve is asymmetric, showing highest flow velocities along the noncoronary leaflet with a counterclockwise rotation of 90° between commissures during systole (12). The resulting higher blood velocity in the lateral aspect of the ascending aorta during systole may be involved in the explanation of this finding. Furthermore, the blood jet through the aortic valve may strike the mid-lateral segment of the ascending aorta. This may dislodge loose atherosclerotic material, which may have been slightly mobilized by earlier surgical maneuvers (such as clamping, proximal anastomosis or cardioplegia needle insertion), and cause embolization into the brain vessels. This view is supported by the findings by Ura and coworkers (13)that clamping- or cannula-induced new lesions, especially of the mobile type, are more often complicated by postoperative stroke. Moreover, a recent study identified new lesions in the ascending aortal intima after decannulation in 16 out of 472 patients (3.4%) (14). New lesions were severe, with mobile lesions or disruption of the intima in 10 patients. Six of the severe lesions were related to aortic clamping and the other four to aortic cannulation. Three patients in this group had postoperative stroke.
Given the high incidence of the disease in patients undergoing cardiac surgery (26%), the implications of this study are intriguing. The 8.7% incidence of postoperative stroke in patients with atherosclerotic disease in the ascending aorta in spite of various modifications in surgical technique is especially worthy of note. The presence of disease in the ascending aorta may dictate more radical modifications in the surgical plan. An aggressive approach is the routine replacement of the ascending aorta under hypothermic circulatory arrest (15). However, the reported 8.6% 30-day mortality rate for this procedure may be held against this approach. The use of off-pump CABG without proximal anastomoses on the ascending aorta may be an alternative because clamping is avoided. However, the drawback is that this may result in incomplete revascularization.
The significance of minor modifications in surgical technique in view of the epiaortic ultrasound information is obscure. Admittedly, clinical studies have indicated less Doppler-detected cerebral embolization (16)and improved neuropsychologic outcome (17,18), but so far a significant effect on clinical stroke rate has not been demonstrated. Although epiaortic ultrasound probably prevented the cannulation of a diseased distal anterior segment, it has led to a major change of surgical strategy in only 3% of patients with diseased ascending aorta in our institution.
The study’s major limitation is that the outcome variable was a clinically evident neurologic deficit instead of deficits in a full neuropsychologic evaluation. The study may also be criticized for not evaluating the exact thickness of atheromas and the presence of ulceration, both reported as risk factors for stroke when evaluating the aortic arch (19,20). However, the primary aims included the evaluation of location and extent of significant (≥0.5 mm) disease in the ascending aorta. Moreover, the presence of severe atheroma is far easier and quicker to evaluate than the presence of ulceration.
In conclusion, significant atherosclerosis in the ascending aorta is a common finding in patients undergoing cardiac surgery. The value of minor modifications in surgical technique in view of the abnormal findings on epiaortic ultrasound is obscure. The incidence of postoperative stroke in spite of these modifications was 8.7%. The presence of calcium in the atherosclerotic plaques may not play a significant role in the development of stroke. The involvement by the disease of more than half of the ascending aorta increases the risk of stroke to 33%. Notably though, single-segment disease carries an 8% risk of stroke. This makes single-segment disease a risk factor that is not easy to ignore. Presence of disease in the middle-lateral aspect of the ascending aorta is an independent risk factor for stroke, with a calculated incidence of 26%. Diabetes mellitus is also a risk factor for postoperative stroke.
☆ This study was supported by grants from Karolinska Institutet.
- coronary artery bypass grafting
- cardiopulmonary bypass
- transesophageal echocardiography
- Received October 11, 2000.
- Revision received March 15, 2001.
- Accepted March 29, 2001.
- American College of Cardiology
- Davila-Roman V.G.,
- Barzilai B.,
- Wareing T.H.,
- et al.
- Pignoli P.,
- Tremoli E.,
- Poli A.,
- et al.
- Davila-Roman V.G.,
- Phillips K.J.,
- Daily B.B.,
- et al.
- Ura M.,
- Sakata R.,
- Nakayama Y.,
- Goto T.
- Katz E.S.,
- Tunick P.A.,
- Rusinek H.,
- et al.