Nonischemic Postoperative Seizure Does Not Increase Mortality After Cardiac Surgery

Nonischemic Postoperative Seizure Does Not Increase Mortality After Cardiac Surgery

Natalia S. Ivascu, MD, Mario Gaudino, MD, Christopher Lau, MD, Alan Z. Segal, MD, William J. Debois, CCP, MBA, Monica Munjal, MS, and Leonard N. Girar...

107KB Sizes 0 Downloads 113 Views

Natalia S. Ivascu, MD, Mario Gaudino, MD, Christopher Lau, MD, Alan Z. Segal, MD, William J. Debois, CCP, MBA, Monica Munjal, MS, and Leonard N. Girardi, MD Departments of Anesthesiology, Cardiothoracic Surgery, and Neurology, New York Presbyterian Hospital, Weill Cornell Medical College, New York, New York

Background. Postoperative seizure (PS) is an infrequent, yet distressing, complication after cardiac surgery. We wished to determine the prognostic significance of these complicated neurologic events. Methods. The Weill Cornell Medical College Department of Cardiothoracic Surgery database and the New York State Department of Health Database were reviewed to identify all patients having PS after cardiac surgery between January 1, 2008, and December 31, 2011. Results. During the study period 3,518 patients had cardiac surgery at the index hospital; 45 of them had PS (1.27%). Overall, patients having PS had a significant increase in 30-day mortality when compared with those not having PS (6.7% versus 1.5%; p < 0.006). The incidence of major postoperative complications in those having PS was also significantly higher (53.3% versus 10.5%; p < 0.001). However, logistic regression failed to demonstrate PS as an independent predictor of perioperative

mortality. When the PS group was further stratified by the presence or absence of cerebrovascular accident, those having both PS and cerebrovascular accident had substantially increased morbidity and mortality (mortality, 0 of 33 versus 3 of 12; major morbidity, 12 of 12 versus 12 of 33; p < 0.01 for both), whereas PS patients without cerebrovascular accident did not have greater risk for either major adverse events or mortality. Conclusions. When PS is associated with acute cerebrovascular accident, a significant increase in postoperative morbidity and mortality can be expected. However, in those with isolated PS (without evidence of new neurologic injury), perioperative mortality and morbidity are comparable to those without any neurologic complications.

N

Murkin and associates [5] reported a series of 24 patients experiencing PS, of which 21 had no evidence of ischemic injury. All of the reported patients had full neurologic recovery and had otherwise uneventful postoperative courses [5]. Conversely, Goldstone and colleagues [6] reported that patients with PS had worse outcomes, including a fivefold increase in hospital mortality. Contradictions in the literature stimulated us to examine our own experience with acute seizure after cardiac surgery in the hopes of elucidating the prognostic implications of PS.

eurologic complications after cardiac surgery include a variety of different clinical presentations. Stroke, short- and long-term memory dysfunction, delirium, cognitive decline, and transient neurologic dysfunction all represent varying degrees of neurologic injury that can impact perioperative outcomes and long-term survival and quality of life [1–4]. To date a considerable amount of evidence exists regarding stroke and postoperative neurologic dysfunction. Although different risk factors have been linked to postoperative seizure (PS), including aortic atherosclerosis, cardiopulmonary bypass time, use of deep hypothermic circulatory arrest, and, most notably, high-dose tranexamic acid, these associations have not been consistent among published studies [5–9]. Far less is known about the incidence, risk factors, and prognostic role of PS. Recently two studies focused on PS after cardiac surgery. Although the incidence of PS was similar in both groups (1.2% to 1.3%), these studies drew conflicting conclusions on the importance of PS with respect to postoperative morbidity and mortality [5, 6].

Accepted for publication Feb 26, 2015. Address correspondence to Dr Ivascu, Department of Clinical Anesthesiology, 525 E 68th St, G 4313d, New York, NY 10065; e-mail: [email protected] med.cornell.edu.

Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier

(Ann Thorac Surg 2015;100:101–6) Ó 2015 by The Society of Thoracic Surgeons

Material and Methods The primary end point of this study was evaluation of the effect of PS on postoperative mortality. Analysis of the effect of PS on postoperative morbidity and definition of the determinants of PS were secondary end points. The Weill Cornell Medical College Department of Cardiothoracic Surgery’s contribution to the New York State Department of Health Database was reviewed to identify all patients having PS after cardiac surgery from January 2008 to December 2011. Our database is contemporaneously updated and maintained by clinical information analysts. Data collection is validated by means of external and internal controls. Our data are then 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.077

ADULT CARDIAC

Nonischemic Postoperative Seizure Does Not Increase Mortality After Cardiac Surgery

ADULT CARDIAC

102

IVASCU ET AL POSTOPERATIVE SEIZURE AND MORTALITY

further validated by the New York State Department of Health before it is included in the state-mandated Cardiac Surgery Database. Preoperative and postoperative variables are entered prospectively during and immediately after a patient’s hospital stay. The electronic inpatient hospital medical record is also queried to gather additional variables on patients identified as having PS. The Weill Cornell Medical College Institutional Review Board approved this study and waived the need for individual patient consent. Patients were included in the PS group if they had cardiac surgery requiring cardiopulmonary bypass and had either convulsive seizure activity or evidence of nonconvulsive seizure activity documented by electroencephalogram. Patients with a preexisting seizure disorder or those on preoperative anticonvulsant medication were excluded from this analysis. A consulting neurologist confirmed the diagnosis of PS in all cases. All patients experiencing PS were treated immediately with antiepileptic medications. Hemodynamically stable PS patients had a noncontrast head computed tomography once their seizure event resolved. A computed tomography scan was obtained in a delayed fashion in those too hemodynamically unstable for transport at the conclusion of the seizure. Antiepileptic medications were continued throughout the patient’s hospital stay and were discontinued by the neurologist at outpatient follow-up.

Anesthetic and Surgical Technique All patients underwent induction of anesthesia with midazolam, fentanyl, and either propofol or thiopental. Type of neuromuscular blockade (rocuronium, cisatracurium, or vecuronium) varied depending on the choice of the anesthesiologist. All patients were given 10-g loading dose of aminocaproic acid followed by an infusion of 1 g/h for the duration of the case. Patients undergoing deep hypothermic circulatory arrest (DHCA) or reoperative sternotomy were given an infusion of 2 g/h. Anesthesia was maintained throughout surgery with fentanyl and volatile anesthetic, according to standard protocol. Patients had a complete transesophageal echocardiographic assessment examination of their heart and aorta; aortic atherosclerosis was graded according to previously reported guidelines [10, 11]. Mean arterial pressure during cardiopulmonary bypass was maintained between 70 and 90 mm Hg, and the hematocrit was maintained at greater than 20%. Patients having cardiac surgery without the need for DHCA were cooled to a minimum temperature between 25 and 32 C during the period of cardiac arrest. Those requiring DHCA were cooled for a minimum of 30 minutes until the bladder or tympanic temperature reached 18 C; in all DHCA cases retrograde cerebral perfusion through the superior vena cava was used and thiopental (1,000 mg) or methohexital (500 mg) were given before commencing circulatory arrest [12]. Patients were warmed until a bladder temperature of 35 C was obtained. During the warming period a maximal temperature gradient of 10 C was maintained between the perfusate and the patient’s bladder temperature. An alpha-stat strategy was used for

Ann Thorac Surg 2015;100:101–6

acid-base management in all patients. Removal of air from the bypass circuit was performed according to standard techniques and was judged adequate on the basis of the absence of air within the cardiac chambers on intraoperative transesophageal echocardiography.

Statistical Analysis Data were analyzed using IBM SPSS Statistics version 22 (IBM, Corp, Armonk, NY). Comparison of various subgroups was made using c2 test and Fisher’s exact test for categorical variables as appropriate. The Student’s t test was used for the comparison of the continuous variables. To analyze the subgroups and determine the factors contributing to mortality and PS, multivariable logistic regression was used. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. All tests required an alpha less than 0.05 for significance. Preoperative and intraoperative factors considered in the logistic regression were age, sex, surgical priority, Canadian Cardiovascular Society angina class, preoperative ejection fraction, preoperative serum creatinine, coronary artery disease, valvular disease, previous cardiac surgery, previous myocardial infarction, cerebrovascular disease, peripheral vascular disease, hemodynamic status at time of surgery, past or present congestive heart failure, chronic lung disease, aortic atherosclerosis, diabetes, renal and hepatic insufficiency, active endocarditis, and use of DHCA as defined in the New York State Department of Health Database. Postoperative factors considered were stroke (cerebrovascular accident [CVA]), Q wave myocardial infarction, deep sternal wound infection, reexploration for bleeding, sepsis, gastrointestinal events, dialysis-dependent renal failure, tracheal intubation longer than 72 hours, and unplanned cardiac reoperation or interventional procedure. All deaths occurring before hospital discharge or within 30 days of surgery were included as mortality.

Results A total of 3,518 patients had cardiac surgery during the study period. Forty-five of them had PS (1.27%). Preoperative characteristics are presented in Table 1. The PS group had a significantly greater percentage that were male, a greater percentage with advanced Canadian Cardiovascular Society class, worse renal function, a higher percentage of valvular heart disease, more previous open heart procedures, and higher incidences of endocarditis and current and previous congestive heart failure. Postoperative morbidity and mortality is demonstrated in Table 2. Patients with PS had a substantially higher incidence of early (10 of 45, 22.2% versus 10 of 3473, 0.3%; p < 0.001) and late CVA (2 of 45, 4.4% versus 25 of 3,473, 0.7%; p < 0.001) after surgery. They had a greater need for hemodialysis (4 of 45, 8.9% versus 61 of 3,473, 1.8%; p < 0.001) and prolonged respiratory support (20 of 45, 20% versus 206 of 3,473, 5.9%; p < 0.001) after surgery. Thirty-day and inhospital unadjusted mortality was significantly greater

IVASCU ET AL POSTOPERATIVE SEIZURE AND MORTALITY

103 ADULT CARDIAC

Ann Thorac Surg 2015;100:101–6

Table 1. Characteristics of the Overall Population and According to the Occurrence of Seizuresa Subgroups Variable

Entire Cohort (n ¼ 3,518)

Seizures (n ¼ 45)

No Seizures (n ¼ 3,473)

66.15  14.38 2,274 (64.6)

70.98  13.8 22 (48.9)

66.09  14.38 2,252 (64.8)

0.575 0.026

2,036 (57.9) 1,482 (42.1)

23 (51.1) 22 (48.8)

2,013 (57.9) 1,460 (42)

0.610

714 (20.3) 0.49  0.12 1.23  1.77 1,577 (44.8) 2,253 (64) 497 (14.1) 16 (0.5) 594 (16.9) 357 (10.1) 900 (25.6) 60 (1.7) 25 (0.7) 1,107 (31.5) 585 (16.6) 281 (8) 183 (5.2) 694 (19.7) 7 (0.2) 96 (2.7) 65 (1.8)

17 (37.8) 0.48  0.14 2.75  6.34 22 (48.9) 38 (84.4) 18 (40) 0 (0) 8 (17.8) 8 (17.8) 12 (26.7) 1 (2.2) 0 (0) 21 (46.7) 17 (37.8) 4 (8.9) 5 (11.1) 8 (17.8) 0 (0) 3 (6.7) 3 (6.7)

697 (20.1) 0.49  0.12 1.21  1.62 1,555 (44.8) 2,215 (63.8) 479 (13.8) 16 (0.05) 586 (16.9) 349 (10) 888 (25.6) 59 (1.7) 25 (0.7) 1,086 (31.3) 568 (16.4) 277 (8) 178 (5.1) 686 (19.8) 7 (0.2) 93 (2.7) 62 (1.8)

0.011 0.084 <0.001 0.581 0.004 <0.001 0.648 0.872 0.088 0.867 0.788 0.568 0.027 <0.001 0.822 0.072 0.741 0.763 0.103 0.016

Age, y (mean  SD) Males Surgical priority Elective Urgent/emergency CCS class Class 3/Class 4 Mean ejection fraction (mean  SD) Creatinine, mg/dL (mean  SD) Coronary artery disease Valvular disease Previous OHS Previous MI (<24 h) Previous MI (>24 h) Cerebrovascular disease Peripheral vascular disease Hemodynamically unstable Shock Congestive heart failure Previous congestive heart failure Chronic lung disease Extensive aortic atherosclerosis Diabetes Hepatic failure Renal failure Active endocarditis a

p Value

Data presented as n (%), unless otherwise noted.

CCS ¼ Canadian Cardiovascular Society;

MI ¼ myocardial infarction;

in patients having PS (3 of 45, 6.7% versus 53 of 3,473, 1.5%; p < 0.006). The median time from arrival to the cardiothoracic intensive care unit to first seizure was 13.5 hours. One patient had a PS before leaving the operating room, and 2 additional patients had delayed presentation of their PS on postoperative days 6 and 12, respectively. Excluding these 3 patients, the mean time to seizure presentation was 15.3 hours. A majority of PS patients had between one and five seizures (40 of 45; 89%); 17 patients (17 of 45; 38%) had only a single seizure. Five of 45 (11%) had six or more seizure episodes. Computed tomography scans were obtained once the patient was deemed stable, and ultimately 41 of 45 PS patients (91.1%) underwent imaging. Multivariate logistic regression analysis was performed to investigate the significance of perioperative risk factors in predicting PS. Preoperative renal insufficiency, valvular disease, and previous cardiac surgery were significant predictors of PS (Table 3). Postoperative CVA was also independently associated with PS. Logistic regression analysis was used to identify risk factors contributing to perioperative mortality (Table 4). Postoperative seizure was not an independent predictor of postoperative mortality.

OHS ¼ open heart surgery.

A separate analysis of PS patients according to the occurrence of perioperative (early plus late) CVA revealed that when PS was associated with a documented CVA, mortality was significantly higher (p < 0.003; Table 5). When comparing PS with non-PS cases and dividing the groups according to the occurrence of perioperative CVA, it becomes apparent that in the absence of CVA the occurrence of PS did not impact the unadjusted postoperative mortality and morbidity, and was associated only with an increased rate of prolonged intubation (Table 6). In the group with CVA, PS did not have a significant association with in-hospital outcomes (Table 6).

Comment Seizure after cardiac surgery remains an enigma. Despite extensive investigation of seizure disorders and epilepsy in other fields [13], there is little consensus regarding the etiology, incidence, and long-term prognosis of this dramatic event after cardiovascular surgery. In the seminal paper by Roach and colleagues [7], PS was grouped with intellectual deterioration and memory

ADULT CARDIAC

104

IVASCU ET AL POSTOPERATIVE SEIZURE AND MORTALITY

Ann Thorac Surg 2015;100:101–6

Table 2. Major Postoperative Events in the Overall Population and According to the Occurrence of Seizuresa Subgroups Entire Cohort (n ¼ 3,518)

Variable Postoperative complications None Stroke <24 h Stroke >24 h Deep sternal wound infection Reexploration for bleeding Gastrointestinal complications Dialysis-dependent renal failure Intubation >72 h 30-day status Alive Dead a

3,130 20 27 7 98 18 65 226

(89) (0.6) (0.1) (0.2) (2.8) (0.5) (1.8) (6.4)

3,462 (98.4) 56 (1.6)

Seizures (n ¼ 45)

No Seizures (n ¼ 3,473)

p Value

21 10 2 0 3 1 4 20

3,109 10 25 7 95 17 61 206

(89.5) (0.3) (0.7) (0.2) (2.7) (0.5) (1.8) (5.9)

<0.001 <0.001 0.004 0.763 0.111 0.106 <0.001 <0.001

3,420 (98.5) 53 (1.5)

0.006

(46.7) (22.2) (4.4) (0) (6.7) (2.2) (8.9) (44.4)

42 (93.3) 3 (6.7)

Data presented as n (%).

deficit and defined as type II (minor) neurologic events, with an incidence of 0.3%. However, because of the design of the statistical analysis grouping the variables together, no inference on the specific effect of PS on outcome could be made in this study. In the few subsequent series published on the topic, the incidence of PS varied from 0.9% to 3.6% [5, 6, 14, 15]. This large heterogeneity is likely the result of differences in patient population, type of surgery, intraoperative medications, and more importantly, the definition of seizure and the type of monitoring. Although the pathogenesis of PS is probably multifactorial, it is generally accepted that solid or gaseous intraoperative microembolization, as well as cerebral inflammation and edema, plays an important role. However, there is no uniformity with regard to the risk factors for PS. It has been suggested that aortic atherosclerosis, open-chamber procedures, use of DHCA, cardiopulmonary bypass and cross-clamp time, use of tranexamic acid, and preoperative critical status predispose patients to PS, but this association was not uniformly reported [5–9]. Considerable controversy also exists regarding the prognostic significance of PS in cardiac surgery patients. Murkin and associates [5] in a retrospective study of 669 patients denied any detrimental effect of PS on inhospital outcome. A few years later Manji and colleagues [8], in a study on 5,958 patients, found significantly

Table 3. Factors Associated With Postoperative Seizures Variable Renal insufficiency Valvular surgery Reoperative surgery Perioperative stroke CI ¼ confidence interval;

OR (95% CI) 1.0 4.4 2.28 4.01

(1.01–1.07) (2.16–8.81) (1.0–5.2) (1.9–8.5)

OR ¼ odds ratio.

p Value 0.003 0.019 <0.001 <0.001

increased mortality among PS cases. However, as the authors did not perform a logistic regression analysis to identify the predictors of mortality and in view of the significant differences in baseline profile between the PS and non-PS groups, the independent effect of PS on outcome in this study could not be ascertained. Goldstone and colleagues [6] recently described a fivefold increase in postoperative mortality for patients who had PS after cardiac surgery. Even in this study, no regression analysis for postoperative mortality was performed, so that the independent prognostic effect of PS could not be evaluated. Moreover, the high operative mortality for both PS and control patients (29% and 6%, respectively) does not appear representative of the current results of cardiac surgery [14]. Finally, in the only prospective investigation on PS in cardiac surgery patients, Gofton and coworkers [15] recently reported no mortality and an uneventful neurologic recovery for PS patients. The conclusions drawn by these authors, however, were from a very limited cohort of 3 PS patients of the 101 having continuous electroencephalographic monitoring. Of note, all of the current studies in the literature suffer from major methodological or sample size limitations.

Table 4. Predictors of 30-Day Mortality Variable

OR (95% CI)

Age Reoperative surgery Previous myocardial infarction Peripheral vascular disease Endocarditis Stroke Gastrointestinal complications Renal failure CI ¼ confidence interval;

1.0 4.4 2.28 4.01 4.7 6.8 13.4 7.97

(1.01–1.07) (2.16–8.81) (1.0–5.2) (1.9–8.5) (1.26–17.7) (1.37–34.09) (3.53–50.97) (2.64–24.08)

OR ¼ odds ratio.

p Value 0.010 <0.001 0.049 <0.001 0.021 0.019 <0.001 <0.001

IVASCU ET AL POSTOPERATIVE SEIZURE AND MORTALITY

105 ADULT CARDIAC

Ann Thorac Surg 2015;100:101–6

Table 5. Major Postoperative Events Among Seizures Patients According to Presence or Absence of Perioperative Strokea Subgroups Seizures Cohort (n ¼ 45)

Variable Postoperative complications None Reexploration for bleeding Gastrointestinal complications Dialysis-dependent renal failure Intubation >72 h 30-day status Alive Dead a

21 3 1 4 20

Stroke (n ¼ 12)

(46.7) (6.7) (2.2) (8.9) (44.4)

0 2 1 2 8

42 (93.3) 3 (6.7)

(0) (16.7) (8.3) (16.7) (66.7)

9 (75) 3 (25)

No Stroke (n ¼ 33)

p Value

(63.6) (3) (0) (6.1) (36.4)

<0.001 0.105 0.094 0.269 0.070

33 (100) 0 (0)

0.003

21 1 0 2 12

Data presented as n (%).

Moreover, major differences in terms of operative mortality, incidence, and definition of stroke and PS exits among the different investigations, making comparison of these results quite challenging. In our study, the incidence of postoperative seizure was 1.3%; multivariate analysis revealed that elevated preoperative creatinine, valvular surgery, and reoperative surgery were independent preoperative predictors of PS. Similar to previous reports [5, 8] univariate analysis of our data revealed a significantly higher incidence of postoperative complications and death in PS versus nonPS patients (p < 0.001; see Table 3). However, a more comprehensive logistic regression model excluded PS as an independent predictor of mortality. The more thorough statistical analysis highlights the confounding role of the differences in preoperative risk profile and the importance of the use of a logistic regression model in studies of this type.

The mortality of both the PS and non-PS groups in our series (6.7% and 1.5%, respectively) seems more representative of contemporary surgical series, compared with some of the previous papers [6]. It is apparent from our study that postoperative CVA is a major confounding factor when studying PS. When the PS group was separated by the presence or absence of perioperative CVA, we found that all the postoperative deaths and the great majority of the complications of the PS groups occurred among PS patients who also had CVA (Table 5). Patients who experienced seizure in the absence of CVA had outcomes similar to the nonseizure counterparts (Table 6). In fact, when prolonged tracheal intubation was excluded, there was no significant difference in the postoperative course between the nonischemic PS patients and the non-PS patients (p ¼ 0.62). It seems unlikely that PS had a real effect on pulmonary function, but rather the sedating effects of antiepileptic

Table 6. Major Postoperative Events in Patients With and Without Perioperative Stroke According to the Occurrence of Seizuresa

Variable

No Stroke Cohort (n ¼ 3,471)

Postoperative complications None 3,130 Reexploration 95 for bleeding Gastrointestinal 17 complications Dialysis-dependent 63 renal failure Intubation >72 h 209 30-day status Alive 3,421 Dead 50 a

Data presented as n (%).

NS ¼ not significant.

(90.2) (2.7)

Subgroups

Subgroups

Seizures (n ¼ 33)

No Seizure (n ¼ 3,438)

p Value

Stroke Cohort (n ¼ 47)

21 (63.6%) 1 (3)

3,109 (90.4) 94 (2.7)

<0.001 0.917

0 (0) 3 (6.4)

0 (0) 2 (16.7)

0 (0) 1 (2.9)

NS 0.091

Seizures (n ¼ 12)

No Seizures (n ¼ 35)

p Value

(0.5)

0 (0)

17 (0.5)

0.686

1 (2.1)

1 (8.3)

0 (0)

0.084

(1.8)

2 (6.1)

61 (1.8)

0.066

2 (4.3)

2 (16.7)

0 (0)

0.064

(6)

12 (36.4)

197 (5.7)

<0.001

17 (36.2)

8 (66.7)

9 (25.7)

0.011

(98.6) (1.4)

33 (100) 0 (0)

3,388 (98.5) 50 (1.5)

0.485

41 (87.2) 6 (12.8)

9 (75) 3 (25)

32 (91.4) 3 (8.6)

0.141

ADULT CARDIAC

106

IVASCU ET AL POSTOPERATIVE SEIZURE AND MORTALITY

drugs and the more conservative postoperative management of PS patients compared with the rest of the non-PS population were the causes of the increased duration of mechanical ventilation. Our observations support the suggestion that perioperative CVA, and not PS, is the important prognostic determinant. The clinical significance of a seizure after cardiac surgery is determined by the presence of a concomitant CVA. It appears that nonischemic seizures represent a distinct, benign form of PS. Our study may also offer intensivists and cardiac surgeons the ability to reassure patients and their families of the relatively benign nature of this event, especially when an acute CVA is not associated with PS. There are some inherent limitations with our study: because both seizure and stroke were rare complications in our overall population, we cannot rule out that, despite the large number of cases, our sample size was too small to detect associations between comorbidities, surgery type, or other preoperative characteristics. Moreover the initial diagnosis of seizure was mainly clinical, and continuous electroencephalogram monitoring was not available on all postoperative patients, so that we cannot exclude that some episodes of brief, self-limited, or nonconvulsive seizures were missed. Despite these weaknesses, our series is one of the largest studies specifically addressing the incidence, etiology, and prognostic significance of seizure after cardiac surgery and is the only one published to date that used logistic regression analysis to clarify the independent effect of PS on postoperative mortality. The heterogeneity of the literature on this topic clearly indicates the need for prospective, multiinstitutional, specifically designed studies on the topic.

References 1. Selnes OA, Gottesman RF, Grega MA, Baumgartner WA, Zeger SL, McKhann GM. Cognitive and neurologic outcomes after coronary-artery bypass surgery. N Engl J Med 2012;366:250–7. 2. Sun X, Lindsay J, Monsein LH, Hill PC, Corso PJ. Silent brain injury after cardiac surgery: a review: cognitive dysfunction and magnetic resonance imaging diffusion-weighted imaging findings. J Am Coll Cardiol 2012;60:791–7.

Ann Thorac Surg 2015;100:101–6

3. Dacey LJ, Likosky DS, Leavitt BJ, et al, Northern New England Cardiovascular Disease Study Group. Perioperative stroke and long-term survival after coronary bypass graft surgery. Ann Thorac Surg 2005;79:532–6. 4. Almassi GH, Sommers T, Moritz TE, et al. Stroke in cardiac surgical patients: determinants and outcome. Ann Thorac Surg 1999;68:391–7. 5. Murkin JM, Falter F, Granton J, Young B, Burt C, Chu M. High-dose tranexamic acid is associated with nonischemic clinical seizures in cardiac surgical patients. Anesth Analg 2010;110:350–3. 6. Goldstone AB, Bronster DJ, Anyanwu AC, et al. Predictors and outcomes of seizures after cardiac surgery: a multivariable analysis of 2,578 patients. Ann Thorac Surg 2011;91: 514–8. 7. Roach GW, Kanchuger M, Mangano CM, et al. Multicenter Study of Perioperative Ischemia Research Group and the Ischemia Research and Education Foundation Investigators. Adverse cerebral outcomes after coronary bypass surgery. N Engl J Med 1996;335:1857–63. 8. Manji RA, Grocott HP, Leake J, et al. Seizures following cardiac surgery: the impact of tranexamic acid and other risk factors. Can J Anaesth 2012;59:6–13. 9. Martin K, Knorr J, Breuer T, et al. Seizures after open heart surgery: comparison of ε-aminocaproic acid and tranexamic acid. J Cardiothorac Vasc Anesth 2011;25:20–5. 10. Shanewise JS, Cheung AT, Aronson S, et al. ASE/SCA guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations of the American Society of Echocardiography Council for Intraoperative Echocardiography and the Society of Cardiovascular Anesthesiologists Task Force for Certification in Perioperative Transesophageal Echocardiography. Anesth Analg 1999;89:870–84. 11. Hartman GS, Peterson J, Konstadt SN, et al. High reproducibility in the interpretation of intraoperative transesophageal echocardiographic evaluation of aortic atheromatous disease. Anesth Analg 1996;82:539–43. 12. Girardi LN, Shavladze N, Sedrakyan A, Neragi-Miandoab S. Safety and efficacy of retrograde cerebral perfusion as an adjunct for cerebral protection during surgery on the aortic arch. J Thorac Cardiovasc Surg 2014;148:2927–33. 13. Nevalainen O, Ansakorpi H, Simola M, et al. Epilepsyrelated clinical characteristics and mortality: a systematic review and meta-analysis. Neurology 2014;83:1968–77. 14. LaPar DJ, Filardo G, Crosby IK, et al. The challenge of achieving 1% operative mortality for coronary artery bypass grafting: a multi-institution Society of Thoracic Surgeons Database analysis. J Thorac Cardiovasc Surg 2014;148: 2686–96. 15. Gofton TE, Chu MW, Norton L, et al. A prospective observational study of seizures after cardiac surgery using continuous EEG monitoring. Neurocrit Care 2014;21:220–7.