| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Cardiothoracic Surgery, Mount Sinai School of Medicine, New York, New York
b Department of Neurology, Mount Sinai School of Medicine, New York, New York
Accepted for publication October 15, 2007.
* Address correspondence to Dr Filsoufi, Department of Cardiothoracic Surgery, Mount Sinai School of Medicine, 1190 Fifth Ave, Box 1028, New York, NY 10029 (Email: farzan.filsoufi{at}mountsinai.org).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentReferences |
|---|
Methods: We retrospectively analyzed prospectively collected data from 2985 patients (2064 men [67%]), with a mean age of 66 ± 11 years, who underwent CABG between January 1998 and December 2006. Stroke was defined as any new permanent focal neurologic deficit (early stroke, 
24 hours; delayed, > 24 hours postoperatively).
Results: The incidence of stroke was 1.6% (n = 48) and similar between conventional CABG (1.6%) and off-pump CABG (1.4%). Early stroke occurred in 25 patients (52%). Brain imaging was obtained in 44 patients (92%): 44 had computed tomography, 3 had magnetic resonance imaging. Results were positive in 33 of 44 patients (75%), showing large embolic stroke in 25 (76%), watershed in 5 (15%), and mixed pattern in 3 (9%). Chronic ischemic changes were found in 17 patients. Multivariate analysis revealed extensive aortic calcification (odds ratio [OR], 4.2), previous stroke (OR, 2.2), female sex (OR, 1.9), and congestive heart failure (OR, 2.6) as predictors of stroke. The hospital mortality rate after stroke was 16.7% (n = 8) compared with 1.5% (n = 44) in those without (p < 0.001). The mortality rate was higher in early stroke at 24% (6 of 25) compared with 9% (2 of 23) in late stroke. Survival of stroke patients was 87% at 1 year and 62% at 5 years and was significantly reduced compared with 96% and 85%, respectively, in patients without stroke (p<0.001).
Conclusions: Most strokes after CABG occurred early after surgery. This complication is associated with an increased hospital mortality and morbidity and reduced long-term survival. The infarction type had no impact on early and late outcome. Preoperative computed tomography scan seems warranted in patients at risk and without any previous history of stroke.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
24 hours) and delayed postoperative stroke. It has been suggested that this distinction is important, because the pathology, cause, and risk factors may vary, depending on the timing of this event [2–5]. This study was designed to determine the incidence of early and late stroke, the topography of cerebral lesions, independent predictors, and late outcome after the occurrence of this complication in patients undergoing CABG.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
|
The main outcome variable of this study was the occurrence of postoperative stroke. This complication was defined in accordance with the NYSDH data registry as any new permanent focal neurologic deficit. Early stroke was defined as a permanent neurologic deficit occurring either intraoperatively or within 24 hours postoperatively. Delayed stroke was defined as a neurologic event occurring more than 24 hours postoperatively throughout hospitalization in the cardiac surgery intensive care unit, step-down unit, regular floor, and Mount Sinai rehabilitation unit. No data are offered for patients who sustained a stroke after discharge from hospital.
Diagnosis of stroke was made by a neurologist and confirmed in most patients by postoperative computed tomography (CT) or magnetic resonance imaging (MRI). The medical records, operative notes, and imaging studies of all patients with stroke were thoroughly reviewed. The description of cerebral lesions was reported according to the Johns Hopkins Cardiac Surgery Stroke Database [8]. Embolic infarction was defined as infarct of a well-defined vascular territory, a watershed infarction was defined as an infarct between vascular territories, and a mixed pattern infarction consisted of signs of both. Patients without stroke served as control group.
Further outcome measures for this study were hospital mortality, major postoperative complications, including respiratory failure, renal failure, deep sternal wound infection, bleeding requiring reoperation, gastrointestinal complications, and unplanned reoperation; length of hospital stay, and late survival. The definition of these variables is reported in Table 1.
Follow-up survival information was obtained by cross-matching patient Social Security numbers with the Web-based Social Security Death Index (www.ssa.gov).
Intraoperative and Postoperative Management
All patients underwent a median sternotomy. Since 2003, we have routinely performed epiaortic scanning of the ascending aorta for detection of atherosclerotic plaques and calcification, particularly at the potential site of aortic cannulation and cross clamp. In the presence of localized atherosclerotic disease, two approaches were selected by the surgeon’s preference: (1) off-pump CABG (OPCABG); (2) identification of a disease-free segment of the aorta for aortic cannulation and cross clamp, followed by conventional CABG. In the presence of extensive aortic calcifications, off-pump CABG was the procedure of choice. Cardiopulmonary bypass (CPB) was used in 2569 procedures (86%), and the remaining 416 (14%) were OPCABG procedures.
Conventional CABG
After systemic heparin was administered to achieve an activated clotting time (ACT) of 480 seconds or more, CPB was established between the ascending aorta and the right atrium using a two-stage cannula. During CPB, a minimum flow of 2.2 L/min/m2 and a perfusion pressure exceeding 60 mm Hg were maintained in all patients. After aortic cross-clamp and cardioplegic arrest, distal anastomoses were performed first, followed by proximal anastomoses using the single cross-clamp technique [9]. The aortic cross-clamp was released thereafter, and patients were weaned from CPB after a short reperfusion. After the completion of CPB, protamine was given based on the heparin level. The decision to administer 
-aminocaproic acid or aprotinin for antifibrinolytic therapy was determined by the preference of the cardiac surgeon and anesthesiologist.
Off-Pump CABG
Off-pump CABG was performed according to the surgeon’s preference. This technique, however, was favored in elderly patients, particularly those with significant comorbidities such as renal dysfunction. In patients undergoing OPABG, heparin was administered to achieve an ACT of 300 seconds. Coronary stabilizer and cardiac positioning devices were used to improve the exposure of the coronary arteries and to facilitate the distal anastomoses under beating heart conditions. Intracoronary shunts were inserted to avoid myocardial ischemia during the construction of the distal anastomoses. Proximal anastomoses were created using a side-bite clamp. Protamine was administered according to the heparin level after the completion of the coronary anastomoses.
Statistical Analyses
Normally distributed continuous variables are presented as mean ± standard deviation (SD) and otherwise as median ± interquartile range (IQR). Categoric variables are shown as the percentage of the sample. A value of p < 0.05 was considered statistically significant for all used tests.
To identify preoperative and perioperative factors (Table 1) associated with the occurrence of postoperative stroke, data were explored by contingency table analyses to look for evidence that some values should be grouped and for evidence of linear trend in continuous variables. The 
2 test, Fisher exact test, and Cochran-Armitage test for trend were used to identify factors that significantly influenced the risk of stroke when considered one at a time. These factors were then entered in a stepwise logistic regression analysis to identify a set of independent variables associated with stroke. The cutoff for inclusion and exclusion into the stepwise logistic regression analysis was p = 0.10. The Hosmer-Lemeshow goodness-of-fit test was used to assess the performance of the analysis [10]. The odds ratio (OR), corresponding 95% confidence interval, and the p value are reported for each independent factor. Long-term survival was analyzed using Kaplan-Meier survival curves. Differences in patient characteristics were controlled by Cox proportional hazard analysis. The statistical analysis was performed using SPSS 15.0 software (SPSS Inc, Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
|
Imaging Study Findings
Of the 48 patients with postoperative stroke, 44 (92 %) had brain imaging studies (Table 3), of which 44 patients had CT scanning and 3 had MRI. The first imaging study for all patients was performed on a mean of 5.7 ± 4.0 days postoperatively. The results of radiologic studies were negative for 11 patients (25%), of which 5 studies were within the first 48 hours after surgery. The remaining 33 patients (75%) had evidence of a new cerebral infarction. Ischemic stroke was noted in all 33 patients (100%) who had positive imaging findings, with an additional hemorrhagic transformation [petechial hemorrhage around the periphery of embolic strokes in 7 (21%)]. We did not observe any primary hemorrhagic strokes in the study.
|
Finally, chronic ischemic changes were found in 17 of 44 patients (38%) with brain imaging studies. These changes were classified as lacunar infarctions in 6 patients (35%), large vessel infarctions in 3 (18%), and white matter ischemic changes in 8 (47 %). Of interest was that only 5 of these patients reported a history of stroke before the surgical procedure.
Predictors of Stroke
Differences in patient’s characteristics in univariate analysis are shown in Table 2. Patients with stroke were more likely to be women and presented with risk factors such as congestive heart failure, previous cerebrovascular accident, peripheral vascular disease, hypertension, and ascending aortic calcification. The mean EuroSCORE for stroke patients was 12 ± 11 compared with 8 ± 7 for the control group (p = 0.007). Other operative characteristics between the two groups were similar (Table 4).
|
|
|
|
Late Survival
Long-term survival of discharged patients with postoperative stroke was significantly decreased compared with those without this complication. Actuarial survival rates at 1, 3, and 5 years were 87.1% ± 5.4%, 76.5% ± 7.4%, and 62.2% ± 9.6% for patients with stroke compared with 95.8% ± 0.4%, 91.6% ± 0.6%, and 85.0% ± 0.8%, respectively, for the control group (p < 0.001; Fig 1).
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
An interesting finding of our study was that the incidence of stroke decreased significantly during the study period, from 2% in 1998 to 2002 to 1% in 2003 to 2006, despite a worsening in the profile of patients referred for surgery as shown by a significant increase in mean predicted death by the EuroSCORE. Several studies have suggested the potential benefit of epiaortic scanning for the detection of atherosclerotic lesions of the ascending aorta in the operation room before manipulation and cannulation to decrease the effect of this risk factor on the occurrence of stroke [5, 11, 12]. Although our study was not designed to determine the benefit of epiaortic scanning, we believe that this reduction in the incidence of stroke is related to the implementation of this measurement in our daily practice. This assumption is further corroborated by the reduction in the incidence of stroke that was mainly observed in early stroke compared with delayed stroke.
In our study, all cerebrovascular accidents were ischemic, with 20% of cases in which we observed hemorrhagic conversion. More than half of the strokes were diagnosed within the first 24 hours after the surgery, indicating that intraoperative events are at the origin of most neurologic incidents. These findings are consistent with the Johns Hopkins experience, where 74% of neurologic complications were identified on the day of operation [8]. However, in another study reported by Hogue and colleagues [2], only 35% of strokes were diagnosed within 24 hours of operation whereas "the majority of events were diagnosed after an initial uneventful recovery from surgery."
There are two main explanations for these different findings: (1) delayed diagnosis of stroke in the Hogue experience and (2) differences in the study population and the nature of data collection. It has been suggested that prospectively collected data lead to a greater incidence of stroke, particularly early events, whereas retrospective studies are associated with a lower rate of strokes. Data collection in our study, similar to that of Johns Hopkins University, was performed prospectively and included purely CABG patients, whereas the study by Hogue and colleagues included patients having valvular procedures with or without CABG.
Most previous studies on stroke after cardiac operations do not provide information on imaging studies. In our series, these studies revealed several unique and important findings.
These latter findings contribute to better understanding of potential mechanisms of stroke after CABG, namely thromboembolic events and perioperative hypoperfusion. In most patients with large artery infarction, the origin of stroke is embolic due to detachment of atheromatous debris from the ascending aorta or the aortic arch or the head vessels [8, 14].
In contrast, watershed infarction may be related to different pathophysiologies. The "watershed" or "border zone" terminology (the territory between major cerebral arteries) refers to imaging findings and does not allow determining the exact mechanism of stroke [3]. Watershed stroke may occur after global cerebral hypoperfusion secondary to low perfusion pressure during cardiopulmonary bypass or to extensive distal cerebrovascular disease, or after multiple microemboli of atheroma or cholesterol [2, 8, 15]. Hypoperfusion may lead to decreased wash out of the microemboli of the border zone region [3, 16]. In our study using CT scan, we noted watershed stroke in 24% of patients.
It has been suggested that more specific and sensitive imaging modalities such as MRI, may identify a greater number of patients with this type of stroke. A recent study by Gottesman and colleagues [17] showed that in patients with postoperative stroke, only 22% presented watershed stroke in CT scan, whereas MRI found as many as 48% presented with this type of stroke.
Another interesting finding of our imaging studies is that 17 of 48 patients (38%) with a postoperative new neurologic event presented with chronic cerebral infarction on postoperative CT scanning. At the preoperative interview, however, only 5 (10%) of these patients presented with a history of stroke. Similarly, Salazar and colleagues [8] found that 13% of stroke patients reported a history of stroke, whereas 42% presented with chronic cerebral changes on imaging studies [8]. In a similar study of 421 patients undergoing CABG, preoperative MRI showed ischemic brain abnormalities in 50% of patients (small brain infarction in 30% and multiple infarctions in 20% of patients) [18]. The authors also reported a correlation between the severity of cerebrovascular lesions on preoperative MRI and postoperative stroke and neurocognitive dysfunction.
Previous studies have identified history of stroke as an important risk factor for the occurrence of new postoperative neurologic events after cardiac operations [1]. The increased risk may be as high as 40% [3]. Considering the devastating effect of stroke on surgical outcome and quality of life of our patients, these findings have a major clinical implication and suggest that in patients with multiple risk factors for stroke and without an apparent history of neurologic incidents, clinicians should have a low threshold to perform a preoperative head CT scan or MRI to detect any episode of old stroke. Even though the findings of such examination would not alter surgical indication in most patients, the positive finding of a chronic neurologic event could be used for a more accurate preoperative risk stratification, patient information, and informed consent as well as systemic application of all perioperative preventive measurements. Such patients may particularly benefit from maintaining high cerebral perfusion pressure during CPB.
Our study revealed four independent predictors of postoperative stroke after CABG, namely, extensive aortic calcification, congestive heart failure, previous cerebrovascular accident, and female sex. When independent predictors of stroke were identified depending on the timing of stroke, congestive heart failure was a predictor of early stroke, female sex was a predictor of delayed stroke, and extensive aortic calcification was a predictor of both early and delayed stroke. This finding was also supported by Hogue and colleague’s [2] study, in which female sex was associated with a 1.6-fold increased risk for stroke whereas it was increased by 1.9-fold in our series [2]. Other studies have also investigated the importance of aortic calcifications on the occurrence of stroke [12, 19]. A large study by van der Linden and colleagues [12] showed that the presence of calcification in the ascending aorta increased the risk of stroke significantly: 6.4% for patients with diseased ascending aorta compared with 1.5% for those without [12].
Other risk factors such as age or diabetes mellitus, which have been previously reported, were not found to be independent predictors of stroke in our analysis [2]. The lack of an independent association between age and the occurrence of stroke after CABG might be because age acts as a surrogate for other independent factors such as aortic calcification, previous cerebrovascular accident, or congestive heart failure.
Outcome
In our study, the mortality rate in patients with stroke was more than 10-fold increased compared with those without this complication. This finding is in accordance with a study reported by Almassi and colleagues [1] who observed a sixfold increase in death in patients who sustained a stroke after CABG [1]. In our observation, early stroke was associated with a higher mortality rate compared with delayed stroke (24% vs 9%); however, it is likely that this difference did not reach statistical significance because of the limited sample size. Hogue and colleagues [2] also observed a 41% mortality rate after early stroke compared with 13% in patients with delayed stroke, confirming our findings [2].
The exact reason for higher mortality rate associated with early stroke remains unclear. McKhann and colleagues [5] have suggested that watershed stroke is associated with a higher mortality rate than embolic stroke [5]. In our study, however, the increased mortality rate observed after early stroke was not related to the topography, extent, and pattern (large embolic vs watershed) of this complication. One potential explanation might be that patients with less severe early strokes and minor symptoms are diagnosed later during hospitalization, whereas those with a more extensive stroke are diagnosed earlier.
In addition, major complications, particularly respiratory failure and sepsis, occurred in 46% of patients with stroke, leading to a prolonged mean hospital stay of 32 days. Stroke also affected the quality of life of our patients, which is reflected by the fact that they were more likely discharged to nursing homes and rehabilitation facilities rather than home.
Stroke patients also had a significantly shorter life expectancy compared with those without this complication. In our experience, stroke patients had a 1-year and 5-year survival rate of 87.1% ± 5.4% and 62.2% ± 9.6% compared with 95.8% ± 0.4% and 85.0% ± 0.8% in nonstroke patients. These findings suggest that future research efforts should be directed to the implementation of preventive strategies to further decrease the occurrence of this complication and to the identification of patients at risk. Furthermore, after the occurrence of stroke, closer follow-up is necessary to decrease the higher mortality rate associated with this complication.
Limitations
This is a retrospective analysis of prospectively collected data and conclusions are necessarily limited in their application. Clinical outcomes after stroke focused on hospital morbidity and mortality, with no information on quality of life, late complications, cause of death, or cost-analysis. Databases may further underreport events and risk factors, and our analysis was only adjusted for data that have been collected. Our study did not examine some previously reported risk factors for the occurrence of stroke, such as atrial fibrillation, low cardiac output syndrome, or use of inotropic medication [1, 2].
Conclusion
Stroke remains a rare but devastating complication after CABG. Most strokes occur within 24 hours after CABG and are ischemic in nature. Imaging studies play a significant role in better understanding the etiology and topography of cerebral lesions. We were able to identify four independent predictors of this complication: female sex, previous stroke, congestive heart failure, and extensive aortic calcification. Finally, stroke is associated with an increased risk of hospital mortality and morbidity and decreased long-term survival.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Takami, K. Tajima, S. Terazawa, N. Okada, K. Fujii, and Y. Sakai Safer aortic crossclamping during short-term moderate hypothermic circulatory arrest for cardiac surgery in patients with a bad ascending aorta J. Thorac. Cardiovasc. Surg., April 1, 2009; 137(4): 875 - 880. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Gottesman and C. W. Hogue Invited Commentary Ann. Thorac. Surg., March 1, 2008; 85(3): 870 - 871. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
