Author + information
- Received June 28, 1999
- Revision received November 9, 1999
- Accepted December 29, 1999
- Published online April 1, 2000.
- ↵*Reprint requests and correspondence: Dr. Ryuzo Sakata, Kumamoto Central Hospital, 96 Tainoshima, Tamukae-machi, Kumamoto City, 862-0965, Japan
This study was performed to evaluate the frequency and risk factors associated with new aortal lesions induced by surgical manipulation and their correlation with postoperative stroke.
Little is known about the causative mechanism of intraoperative atheroembolism after cardiac surgery.
Epiaortic echocardiography was performed before cannulation and after decannulation in 472 patients undergoing cardiac surgery with extracorporeal circulation.
A new lesion in the ascending aortal intima was identified in 16 patients (3.4%) after decannulation. 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. Univariate analysis identified only the maximal thickness of the atheroma near the aorta manipulation site as a predictor of new lesions. The incidence of new lesions was 11.8% if the atheroma was ∼3 to 4 mm thick and as high as 33.3% if the atheroma was >4 mm, but only 0.8% when it was <3 mm. Total 10 patients (2.1%) sustained neurological complications. Arteriosclerosis obliterans, atherosclerosis of the aorta and new mobile lesions were identified as predictors of strokes.
This study demonstrated an association between new lesions created by surgical maneuvers and postoperative stroke. Embolic strokes were more likely to occur if new lesions were complicated with intimal disruption, especially of the mobile type. Modifications in surgical procedures will be needed if thick plaque (especially >4 mm) is noted near the manipulation site.
The dislodgement of material after vascular injuries caused by surgical manipulation has been recognized as a major cause of stroke after open heart surgery (1–5). Katz et al. (5) observed a mobile atheroma disappear after aortic cannulation during transesophageal echocardiographic (TEE) monitoring. The patient had a stroke postoperatively, adding to evidence that patients with mobile atheromatous disease are at higher risk of embolic stroke during cardiopulmonary bypass (5). Our previous studies have revealed that clamp- or cannula-induced new lesions, especially of the mobile type, are often complicated by postoperative stroke (6). However, there are few reports about the mechanism of intraoperative atheroembolism after cardiac surgery. We performed this study because postdecannulation imaging was hypothesized to be helpful in elucidating the mechanism of intraoperative atheroembolism and its associated risk factors.
Epiaortic two-dimensional echocardiography was performed before cannulation and after decannulation in patients undergoing open heart surgery with extracorporeal circulation (ECC). Of 496 patients undergoing open heart surgery at Kumamoto Central Hospital between January 1996 and December 1997, epiaortic two-dimensional echocardiography was not performed in 24 patients because of a lack of equipment availability and the emergency nature of the operation. A total of 472 patients (299 men and 173 women, mean age 65.6 ± 9.8 years [range 29 to 86]) were studied as described earlier. Three hundred and seventeen patients underwent isolated coronary artery bypass graft surgery (CABG), 32 underwent CABG in conjunction with valve surgery, 100 underwent valve surgery, 16 had an operation for congenital heart disease and 7 had miscellaneous cardiac surgery (isolated ventricular aneurysmectomy in two patients, VSP in 1, myxoma in three and pacemaker lead infection in 1). This study was approved by the institutional Ethics Committee and the Internal Review Board. All patients gave written, informed consent, and all procedures were in accordance with established institutional guidelines.
To identify the changes induced by the manipulation, the same sonographic probe was used for two-dimensional epiaortic ultrasonography of the ascending aorta before cannulation (pre-ECC) and after decannulation (post-ECC), as described in a previous study (6). For analytical purposes, the ascending aorta and proximal arch were divided into two segments, designated as the proximal and distal segments. The proximal segment included the aorta from the aortic annulus to 1 cm proximal to the innominate artery, and the rest of the ascending aorta and proximal arch were included in the distal segment. The proximal arch was meticulously dissected to obtain good visualization. This categorization was carried out to relate the proximal segment to the aortic clamp and the distal segment to the aortic cannulation.
In pre-ECC evaluation, we graded the severity of atherosclerosis, taking into account the maximal thickness of the lesion in the proximal and distal segments of the ascending aorta. The severity of the aortic atherosclerosis was assessed as grade 1 if the intima had no thickening or was minimal (<3 mm); grade 2 if the intimal thickening was between 3 and 4 mm; and grade 3 if the thickening was protruding >4 mm, often with an irregular surface or mobile components. Epiaortic ultrasonographic imaging was repeated after decannulation for comparison with data obtained before ECC, using the cannulation site and pulmonary artery as reference points.
All operations were performed under standard cardiopulmonary bypass and moderate hypothermia (28 to 32°C). Arterial cannulations were performed in the ascending aorta (THI angled type aortic 21F or 24F perfusion cannula, Argyle Co., Missouri) or femoral or subclavian artery (William Harvey arterial perfusion cannulae, type 1858, Bardic Co, California), depending on the severity of atherosclerosis. Except for one patient, all anastomoses, including the proximal ones, were performed during a single period of aortic cross-clamping (Fogarty soft-jaw clamp, 86 mm, Baxter Co., California). Because palpation of the ascending aorta underestimates the severity of atherosclerosis, the decision to implement some modifications of the operative technique was made primarily on the basis of the qualitative sonographic information and secondarily on the surgeon’s feeling.
Postoperative neurologic events
Patients were examined preoperatively by the physician in charge. If a neurologic deficit was suspected, consultation with the neurologist was sought, but no neuropsychological studies were made preoperatively or postoperatively. Two independent neurologists assessed the development of perioperative stroke. Only permanent or reversible focal complications such as reversible ischemic neurologic deficit or transient ischemic attacks were considered neurologic events. Confusion, agitation, dementia, disorientation or psychosis were considered neurologic events only if new focal neurologic signs were also present. When clinical conditions permitted, computed tomographic (CT) scans of brain were obtained in patients with postoperative neurologic events, and in some patients, plain or angiographic magnetic resonance imaging, carotid duplex scanning and/or scintigraphy of the brain were performed in an effort to seek the cause (embolic or hypoperfusion) of the stroke.
To assess the association between the new lesions and the preoperative variables (gender, age, hypertension, diabetes, hyperlipidemia, chronic renal failure, arteriosclerosis obliterans [ASO], abdominal aortal aneurysm [AAA], history of smoking, ischemic heart disease) and the intraoperative variables (duration of aortic cross-clamping, severity of atherosclerosis of the ascending aorta, maximal thickness of the atheroma near the aorta manipulation site, calcification of plaque), univariate analysis was performed.
Univariate analysis was done to assess the association between stroke and history-related variables (gender, age, hypertension, diabetes, hyperlipidemia, cerebral vascular events, chronic renal failure, ASO, AAA, duration of ECC, duration of aortic cross-clamping, severity of atherosclerosis of the ascending aorta, carotid artery disease or presence of new mobile lesions in the aorta).
Univariate testing of variables was performed with the Fisher exact test for discrete variable comparisons. The Mann-Whitney U test was used for continuous variable comparisons.
For multiple comparisons, a significance level of 0.003 was set to ensure the whole type I error probability was <0.05, according to the Bonferroni method. All analyses were performed using the SAS Institute version 6.12 software (Cary, North Carolina).
Grade 1 severity of the atherosclerosis was present in 371 patients; grade 2 in 73; and grade 3 in 28. The results of the pre-ECC evaluation indicated that modifications to operative procedures would be required in 63 patients. Cross-clamping of the ascending aorta was precluded in 23 patients, and in these patients, CABG under hypothermic fibrillation in circulatory arrest or on the beating heart was performed. The arterial cannulation site was changed in 52 patients. Operative modifications were required more often in patients with a severely atherosclerotic aorta (p < 0.001).
A new lesion in the ascending aortal intima was identified in 16 (3.4%) of 472 patients after decannulation (Table 1). The echogenicity of new lesions was usually isoechoic to that of the intima nearby, suggesting that the mobile lesions were disrupted plaque or intima. However, owing to a lack of histologic confirmation, there is still a chance of new lesions with thrombus formation on top of the disrupted plaque, so the term “new lesion” was used instead of “new intimal lesion.” All lesions except one were distributed in the posterior portion of the aorta. New lesions were of the minor form with some irregularity of the intima in six patients, and of the severe form with mobility or disruption of the intima in 10 patients (Figs. 1 and 2). ⇓⇓ Six of the severe lesions were attributable to aortic clamping (five to cross-clamping and one to tangential clamping) and the other four to aortic cannulation. Of four new lesions that were thought to be related to aortic cannulation in three patients, the new lesions were located in the direction of the cannula jet and opposite the cannulation side, suggesting that the probable cause was the aortic cannula jet. Three patients in this group had postoperative stroke. No strokes occurred in patients with only new intimal irregularity or intimal tearing with no mobile component.
Univariate analysis was performed to identify the risk factors associated with the new lesions induced by aortic manipulation and with the preoperative and intraoperative variables. Because new lesions caused by aortic cannulation appeared to be multifactorial, these data were not included in the analysis. Thus, only 12 new lesions caused by aortic cross-clamping were analyzed.
In 449 patients who underwent aortic cross-clamping, univariate analysis identified only the maximal thickness of the atheroma near the aorta manipulation site to be a predictor of new lesions (Table 2).
A total of 10 patients (2.1%) sustained neurologic complications. The clinical profiles of these patients with postoperative stroke are summarized in Table 3. Predictors of stroke were determined by univariate analysis (Table 4). Univariate analysis identified ASO, atherosclerosis grade of the ascending aorta and new mobile lesions as predictors of strokes. The preoperative presence of mobile plaque did not influence the occurrence of strokes (Table 4).
Hospital death occurred in six patients (1.3%), three of whom had a stroke and died from infections. The causes of death in the other patients included LOS in one, celiac artery embolization in one and gastrointestinal bleeding in one.
Vascular injuries after cardiac surgery have been reported in previous studies (7,8). These include disruption of the intima, laceration, perforation, dissecting aneurysm and false aneurysm. Black et al. (8) reported 50 instances of vascular injuries after cardiac prosthetic valve replacement. Of these, nine patients had injuries caused by surgical clamps. These investigators noted that preexisting abnormalities in the vessel wall indicated a predisposition to certain types of injury. Coelho et al. (9) reported the results of arteriographic and ultrasonic evaluation of vascular clamp injuries, using an in vitro human experimental model. They noted that intimal tears and flaps were observed in 14% and 26% of all atherosclerotic vessels, respectively, whereas clamp trauma is minimal in normal arteries. Ultrasound imaging is superior to arteriography in detecting intimal flaps and can show the majority of intimal flaps caused by vascular clamps.
The dislodgement of material after vascular injury caused by surgical manipulation has been recognized as a major cause of stroke after open heart surgery. Katz et al. (5) observed a mobile atheroma disappear after aortic cannulation during TEE monitoring. The patient had a stroke postoperatively, adding to the evidence that patients with mobile atheromatous disease are at higher risk of embolic strokes during cardiopulmonary bypass. However, there are few reports on the mechanisms that cause intraoperative atheroembolism after cardiac surgery. We performed this study because imaging after decannulation was hypothesized to be helpful in elucidating the mechanism of intraoperative atheroembolism and its associated risk factors. Of 449 patients undergoing open heart surgery with some kind of manipulation of the aorta, a new lesion in the ascending aortal intima was identified in 16 patients (3.6%) after decannulation. Probable causes of new lesions included aortic cannulation, the aortic cannula jet, aortic cross-clamping and tangential clamping.
In a study to determine whether atherosclerotic plaques in the aortic arch are a risk factor in recurrent brain infarction and vascular events, after adjustment for the other risk factors, aortic plaques ≥4 mm thick were found to be independent predictors of recurrent brain infarction and all vascular events (10). Using an in vitro human experimental model, Coelho et al. (9) concluded that the occurrence of significant trauma caused by vascular clamps in atherosclerotic arteries is high and occurs whether or not discrete plaques are clamped. In our study, the maximal thickness of the atheroma near the aorta manipulation site was identified as a predictor of new lesions. The incidence of new lesions was 11.8% if the atheroma was ∼3 to 4 mm and as high as 33.3% if the atheroma was >4 mm, suggesting that diseased plaque is dangerously fragile. Although no definite conclusion could be drawn owing to the small number of events, these results may help in formulating a strategy for modification of surgical procedures when dealing with a diseased aorta.
Aranki et al. (11) reported, in a multivariate logistic regression analysis of adverse outcomes in a retrospective study of 310 patients, that use of a partial occluding clamp was a significant predictor of adverse outcome. They noted that application of a partial occluding clamp is probably the most traumatic form of manipulation of the ascending aorta, with the total force of the clamp being concentrated onto a small area, possibly increasing the likelihood of an intimal tear and the dislodgement of plaque or atheromatous material that could embolize on removal of the clamp. At our institution, all anastomoses, including proximal ones, were performed during a single period of aortic cross-clamping in CABG, except in one patient. The only patient in whom a side-biting clamp was used to perform the proximal anastomosis developed new mobile type lesions and had a postoperative stroke. Despite the attempt to clamp a relatively less atherosclerotic anterior portion of the aorta, new mobile type lesions developed in the posterior portion of the same segment. It is highly likely that the application of a side-biting clamp changed the three-dimensional geometry of the aorta, stressing and disrupting the fragile plaque in the posterior wall. Careful application of a side-biting clamp is needed, even if the atherosclerosis involves the posterior part of the aorta. In a diseased aorta, the clamping site of a side-biting clamp, as well as a total cross clamp, should be carefully evaluated.
The association between protruding aortic atheromas and spontaneously occurring embolic disease has been reported (12). Tunick et al. (12) reported that protruding atheromas seen by TEE predict future vascular events. Embolic events were more likely to occur when the debris was pedunculated and highly mobile than when it was sessile and immobile. Controversy exists as to whether the mobile component seen by ultrasonography comprises disrupted plaque or thrombus imposed on plaque. Some investigators have reported successful treatment of mobile thrombus by the use of anticoagulant agents or surgical removal (13,14). Histologic examination revealed thrombosis protruding into the aortic lumen, attached to an atherosclerotic lesion (13). However, in our hospital, histologic examination of surgically removed mobile lesions revealed atheromatous plaque with cholesterol crystal and macrophage invasion (not yet reported). In our study, the echogenicity of new lesions was usually isoechoic to that of the intima, suggesting that mobile lesions comprised disrupted plaque or intima. Taking these facts into account, and considering the relatively short time lag between aortic manipulation and reevaluation by echocardiography (10 to 15 min after decannulation), it is likely that, in our cases, the mobile component seen on the post-ECC images was not thrombus but was disrupted plaque.
In the present study, new mobile lesions created by surgical manipulation have been identified as predictors of strokes, in addition to the more familiar predictors, ASO and atherosclerosis of the ascending aorta, described in previous studies (1,6). However, no direct cause–effect relation between stroke and the new lesions was proved. It is not known whether mobile lesions indeed embolized. However, by univariate analysis, new mobile lesions were identified as risk factors of strokes. This suggests that new lesions could play a role in triggering strokes with or without preexisting atherosclerotic disease. This may result from the atheroembolism of mobile plaque or thromboembolism derived from newly disrupted lesions in the early or late postoperative stage. Although not all new lesions result in strokes, efforts not to create new lesions, including suitable strategies based on pre-ECC imaging, careful maneuvers and meticulous feedback by means of post-ECC imaging, may reduce the number of strokes.
This nonexperimental study was undertaken to clarify the causative mechanism of embolic stroke during open heart surgery. For obvious ethical reasons, data could not be obtained on the impact of surgical manipulation on the severely diseased aorta in which the risk of embolization was high. In addition, because efforts were made to avoid creating atherosclerotic lesions during manipulation, the frequency of new lesions was decreased, so the number of new lesions was too small to draw any definite conclusions.
Previous studies have demonstrated that the presence of triglyceride-rich lipoproteins may predict a plaque type particularly vulnerable to rupture, because the echolucency of carotid atherosclerotic plaques on computerized ultrasound B-mode images has been associated with a high incidence of brain infarcts as identified by CT scans (15). However, because our study was retrospective, plaque echogenicity was not included in this analysis. Measurement of echogenicity be means of computerized ultrasound B-mode images may enable more accurate identification of vulnerable plaque to be made in the future, before surgical manipulation.
Because intraoperative TEE was not used routinely in this series, TEE data were not included in this analysis. However, because we meticulously dissected the proximal arch before pre-ECC ultrasonography, we believe that inclusion of this data may reflect, to some extent, the degree of atherosclerosis of the arch. Continuous monitoring by TEE, combined with pre-ECC and post-ECC imaging by means of epiaortic ultrasonography, may help in clarifying the mechanism by which strokes are caused.
This study demonstrated the association between new lesions created by surgical maneuvers and postoperative strokes. Embolic strokes were more likely to occur if new lesions were complicated by intimal disruptions, especially of the mobile type. If thick plaque (especially >4 mm) is noted near the manipulation site, modifications in surgical procedures will be necessary to avoid creating new lesions.
- abdominal aortal aneurysm
- arteriosclerosis obliterans
- coronary artery bypass graft surgery
- computed tomographic
- extracorporeal circulation
- transesophageal echocardiography
- Received June 28, 1999.
- Revision received November 9, 1999.
- Accepted December 29, 1999.
- American College of Cardiology
- Katz E.S.,
- Tunick P.A.,
- Rusinek H.,
- et al.
- Tunick P.A.,
- Rosenzweig B.P.,
- Katz E.S.,
- et al.
- Laperche T.,
- Laurian C.,
- Roudaut R.,
- Steg P.G.
- Gronholdt M.L.,
- Nordestgaard B.G.,
- Wiebe B.M.,
- et al.