Recurrent Seizures Following Cardiac Surgery: Risk Factors and Outcomes in a Historical Cohort Study Rizwan A. Manji, MD, PhD, MBA, FRCSC,*†‡ Hilary P. Grocott, MD, FRCPC,*†‡ Jacqueline S. Manji, PhD,* Alan H. Menkis, DDS, MD, FRCSC,*‡ and Eric Jacobsohn, MBChB, MHPE, FRCPC*† Objectives: To determine the risk factors for and outcomes after recurrent seizures (RS) in patients following cardiac surgery. Design: A historical cohort study. Setting: A single-center university teaching hospital. Participants: Cardiac surgery patients from April 2003 to September 2010 experiencing postoperative seizures. Interventions: None. Measurements and Main Results: Patients were divided into an isolated seizure group and an RS group. Risk factors for RS were determined using logistic regression. Intermediate-term follow-up was conducted by phone. Of 7,280 consecutive patients undergoing cardiac surgery, 61 (0.8%) experienced postoperative seizure and 36 (59%) of those experienced at least 1 recurrence. Of these, 32 (89%) experienced RS within 24 hours of the first seizure, and 29 (81%) had grand mal seizures. Preoperative creatinine Z120 μmol/L (p ¼ 0.02), time until first seizure occurred (r4
hours; p ¼ 0.01), and procedures involving the thoracic aorta were associated with RS (R2 ¼ 0.53, p o 0.05). Patients with RS had longer intensive care unit stays (5.3 v 2.9 days, p ¼ 0.03) and longer mechanical ventilation duration (53.3 v 15.0 hours, p ¼ 0.01). At a median followup of 21 months for the RS group and 16 months for the isolated seizure group, restrictions, anticonvulsant use, morbidity, and mortality were similar between patients with isolated versus recurrent seizures. Conclusions: Higher preoperative creatinine, thoracic aortic surgery, and early seizure onset were associated with RS after cardiac surgery. When compared to isolated seizures, recurrence per se was not associated with significantly increased long-term morbidity or mortality. & 2015 Elsevier Inc. All rights reserved.
P
impact a patient’s quality of life regarding return-to-normal activity and the possibility of long-term anticonvulsant therapy. The purpose of this historical cohort study was to create an exploratory predictive model for seizure recurrence after cardiac surgery as well as to compare the outcomes for these patients compared to those with isolated seizure.
OSTOPERATIVE SEIZURES, although relatively uncommon, are a well-documented complication of cardiac surgery. The incidence of seizures after cardiac surgery varies between 0.5 and 7.6%.1 When seizures occur, recurrence rates of 40% to 66% have been reported.2–5 Seizures may be caused by thromboembolic ischemic stroke, cerebral air embolism, medication toxicity related to antibiotics, or other perioperative drugs such as tranexamic acid (TXA).6–8 The authors previously reported that administration of TXA, preoperative cardiac arrest, Acute Physiology and Chronic Health Evaluation (APACHE) II scores 420, previous cardiac surgery, openchamber procedure, cardiopulmonary bypass (CPB) time 4150 minutes, and preoperative neurologic disease are all independent risk factors for seizures after cardiac surgery.4 Importantly, approximately 60% of patients in that study had more than 1 seizure during their postcardiac surgery hospital stay, thereby defining recurrent seizure (RS). In the general non-surgical population, RS often is related to factors such as a history of epilepsy, subtherapeutic anticonvulsant levels, alcohol withdrawal, and structural brain abnormalities.9,10 However, associated risk factors and outcomes for RS after cardiac surgery are not known. Knowing these risk factors could influence in-hospital clinical decision making and
From the *Cardiac Sciences Program, Winnipeg Regional Health Authority and St. Boniface Hospital, †Department of Anesthesia and Perioperative Medicine; and ‡Department of Surgery, University of Manitoba, Winnipeg, Manitoba, Canada. Address reprint requests to Rizwan A. Manji, University of Manitoba, St. Boniface Hospital, CR3014 - 369 Tache Avenue, Winnipeg, Manitoba, Canada R2H 2A6. E-mail: [email protected] © 2015 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2015.03.020 1206
KEY WORDS: recurrent seizures, cardiac surgery, follow-up, anticonvulsants, tranexamic acid
METHODS
The local Research Ethics Board approved this study (January 2010) and waived the requirement for patient consent for the historical cohort component of the study. For follow-up, informed consent was obtained from patients prior to the telephone interview. This report represents a subset of patients from a previously published single-center historical cohort study performed on all patients who underwent cardiac surgery at the authors’ institution between April 1, 2003 and September 30, 2010 who had postoperative seizures.4 In the original study, the impact of TXA and other risk factors were compared between patients who did and did not experience postoperative seizure. For this present study, patients from the previous cohort who experienced post–cardiac surgery seizures were divided into 2 groups: those who had one postoperative isolated seizure and those who experienced RS (ie, more than one postoperative seizure while in the hospital). The reason for this division was to specifically explore the significance of recurrent seizures as compared to isolated seizures. Data collected from the authors’ cardiac surgery, perfusion service, and intensive care unit (ICU) databases included information on all consecutive cardiac surgery cases performed at their institution. The perfusion service database collects intraoperative variables, and the cardiac surgery and ICU databases collect demographic and comorbidity variables (including postoperative complications), ventilation time, and lengths of ICU and hospital stays. A concurrent pharmacy database maintains information related to the inventory of
Journal of Cardiothoracic and Vascular Anesthesia, Vol 29, No 5 (October), 2015: pp 1206–1211
1207
RECURRENT SEIZURES
drugs dispensed to various clinical services, including anesthesia for cardiac surgery.4 A seizure was defined as a physician-documented event such as rhythmic tonic-clonic motion of localized body parts consistent with a focal seizure, or of all four limbs with a decreased level of consciousness consistent with a grand mal seizure that required prescribed intervention. Nonconvulsive seizures were diagnosed in patients with a persistent decrease in the level of consciousness and confirmed seizure activity on serial electroencephalography (EEG). A telephone interview was conducted more than a year after discharge to ascertain if there had been any hospital readmission, new neurologic diagnoses (eg, RS and stroke), the need for long-term anticonvulsant therapy, and/or any activity restrictions (such as driving) instituted that were related to seizures. Statistical Analysis All statistical analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, IL). Univariate analysis was performed with a Student’s 2-sample t-test (for normally distributed data), the Mann–Whitney U test (for non-normally distributed data) and Pearson’s chi-square test (for categoric data) in which potential recurrent seizure risk factors and outcomes were compared between groups. Due to a relatively small sample size, the authors opted to create an exploratory (ie, hypothesis-generating) model for risk factors for RS that could serve as a basis for future studies. As such, multivariable logistic regression was performed using clinically relevant variables (that had a univariate p value o0.1) to create a
model with the highest R2 value and largest area under the receiver operating curve (ROC). A 2-tailed p value r 0.05 was considered statistically significant. RESULTS
Sixty-one (0.8%) of 7,280 consecutive patients experienced a postoperative seizure (Fig 1). Of these 61 patients, 36 (59%) had at least 1 RS, 17 patients (28%) experienced more than 1 recurrence, and 3 (8%) had non-convulsive status epilepticus. Of the 36 patients with RS, 32 (89%) experienced RS within 24 hours of the first seizure, and 4 patients (11%) experienced RS 424 hours after the first seizure. Grand mal seizures accounted for the majority of patients with RS (29/36; 81%), and each seizure (except for the 3 patients with nonconvulsive status epilepticus) lasted less than 5 minutes. A comparative univariate analysis between patients with RS and patients with isolated seizure is outlined in Table 1. Importantly, patients with RS had higher preoperative creatinine levels, higher APACHE II scores, more thoracic aortic and deep hypothermic circulatory arrest (DHCA), more postoperative cerebrovascular accidents (CVA), and a shorter period of time from the end of surgery until the first seizure occurrence (ie, early onset of the first seizure). Of note, patients with isolated seizures and patients with RS had similar doses of TXA given, although the authors did not know the serum (or cerebrospinal fluid levels11) of TXA. Variables associated with RS that were considered biologically plausible and that had a p value o 0.1 were further examined in the authors’ multivariate model. These variables included preoperative creatinine, APACHE II score, open-
Fig 1. Flow diagram of the historical cohort observed in this study comparing patients with isolated seizures to patients with recurrent seizures (RS). The median follow-up period was 21 (6-30) months for the recurrent seizure group and 16 (11-28) months for the isolated seizure group.
1208
MANJI ET AL
Table 1. Univariate Comparison Between Patients With Isolated or Recurrent Seizure After Cardiac Surgery Variable
Number of patients per group Age (years) Female Preoperative seizure disorder Preoperative alcohol abuse Preoperative cerebrovascular disease Ejection fraction (%) Preoperative hemoglobin (g/L) Preoperative creatinine (μmol/L) APACHE II score Preoperative cardiac arrest Preanesthetic use of benzodiazepine Intraoperative use of benzodiazepine Intraoperative use of propofol TXA dose (mg/kg) CABG only (including redo CABG) Open-chamber procedure (valve procedure, aortic procedure, intracardiac procedure) Procedure involving aortic valve Procedure involving thoracic aorta Deep hypothermic circulatory arrest during surgery Cross-clamp time (min) Cardiopulmonary bypass time (min) Lowest mean arterial pressure (mmHg) during OR Duration of lowest mean arterial pressure (min) Postoperative cerebrovascular accident Postoperative cardiac arrest New postoperative renal dysfunction Time from end of surgery until first seizure (hours)
Isolated Seizure
Recurrent Seizure
p Values
25 68.7 (13.0) 8/25 (32%) 0/25 (0%) 2/25 (8%) 5/25 (20%) 55.0 (55.0-60.0) 125 (17) 92.0 (73.0-103.0) 17.7 (6.6) 0/25 (0%) 2/20 (10%) 1/20 (5%) 17/20 (85%) 68.4 (36.1) 9/25 (36%) 15/25 (60%) 6/25 (24%) 1/25 (4%) 0/25 (0%) 75.0 (51.5-134.0) 124.0 (85.0-200) 51.6 (6.9) 10.0 (10.0-20.0) 1/25 (4%) 2/25 (8%) 6/25 (24%) 10.5 (4.8-17.5)
36 70.6 (10.6) 10/36 (28%) 2/36 (6%) 1/36 (3%) 8/36 (22%) 55.0 (45.0-60.0) 121 (22) 103.0 (93.5-136.5) 22.3 (5.8) 3/36 (8%) 6/36 (17%) 2/36 (6%) 27/36 (75%) 74.9 (34.3) 7/36 (19%) 29/36 (81%) 16/36 (44%) 11/36 (31%) 8/36 (22%) 77.0 (50.5-130.0) 165.0 (92.3-218.3) 49.9 (10.9) 10.0 (10.0-15.0) 8/36 (22%) 0/36 (0%) 5/36 (14%) 4.1 (1.8-10.6)
0.53 0.78 0.51 0.56 1.00 0.64 0.54 0.01 0.01 0.07 0.70 1.00 0.51 0.51 0.24 0.08 0.10 0.02 0.02 0.96 0.23 0.53 0.95 0.05 0.16 0.33 0.03
NOTE. Continuous variables reported as mean (SD) for normally distributed data or median (IQR) for non-normally distributed data. Categoric data are reported as n/range. Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CABG, coronary artery bypass graft; OR, operating room; TXA, tranexamic acid.
chamber procedure, procedure involving the thoracic aorta, and time duration from the end of surgery until the first seizure occurred. Preoperative cardiac arrest was not included in this analysis as the event rate was too low; only 3 patients with RS experienced this variable. “Deep hypothermic circulatory arrest during surgery” was the same as “procedures involving the thoracic aorta.” Thus, only the thoracic aortic procedure variable (with the larger event rate) was examined. Postoperative CVA also was the same as the thoracic aortic procedure variable and, thus, was not examined separately. The multivariate model (Table 2) demonstrated that elevated preoperative creatinine (Z120 μmol/L), procedure involving the thoracic aorta and early seizure onset (r4 h from the end of surgery until the first seizure) were associated with RS [R2 ¼ 0.53, p o 0.05, area under the ROC curve 0.859 (95% confidence intervals, 0.756-0.963)].
The median (IQR) duration of anticonvulsant therapy was significantly longer in the RS group (4.0 [3.0-5.8]) days) versus the isolated seizure group (0.3 [0.1-1.0] days; p o 0.001). However, at the discretion of the attending physicians, 29 (81%) of the 36 patients in the RS group and 19 (76%) of the 25 in the isolated seizure group had their anticonvulsant therapy discontinued during their hospitalization (p ¼ 0.69). Patients underwent extended periods of in-hospital observation (see below). No seizure occurred in-hospital after discontinuation of anticonvulsant therapy in either group. Patients in the RS group had significantly longer median (IQR) mechanical ventilation times than in the isolated seizure group (53.3 h [18.5 to 154.0] v 15.0 h [2.2 to 48.0] group; p ¼ 0.01), as well as significantly longer median (IQR) ICU stay (5.3 days [3.0 to 10.3] v 2.9 days [1.9 to 6.0]; p ¼ 0.03). However, the median (IQR) length of hospital stay was similar
Table 2. Logistic Regression of Potential Predictors of Recurrent Seizures After Cardiac Surgery Variable
Preoperative creatinine (Z120 μmol/L) Procedure involving thoracic aorta Time from end of surgery until first seizure r4 h
Unadjusted Odds Ratio
95% CI
Adjusted Odds Ratio
95% CI
2.6 10.6 4.6
0.7-9.6 1.3-88.2 1.1-19.2
11.3 Undefined† 11.0
1.6-81.8 Undefined† 1.9-65.5
NOTE. R2 ¼ 53%; Area under the ROC Curve ¼ 0.859 (0.756 – 0.963). †Undefined since odds ratio approaches infinity as almost all events occurred within the recurrent seizure group.
1209
RECURRENT SEIZURES
Table 3. In-Hospital Mortality and Live Discharge for All Patients Experiencing Seizure Post–Cardiac Surgery (Includes Patients With Isolated and Recurrent Seizures) Variable
Total no. of patients Age (years) Female COPD Ejection fraction (%) APACHE II score Preoperative creatinine (μmol/L) Major stroke (watershed or large vessel) on CT head New stroke on CT head New postoperative renal dysfunction Focal seizure Grand mal seizure Nonconvulsive seizure Recurrent seizure
In-hospital Mortality
Live Discharge
p Value
13 72.3 (8.2) 2/12 (17%) 4/12 (33%) 42.5 (32.5-55.0) 26.7 (7.0) 122.0 (100.8-161.5) 4/8 (50%) 4/8 (50%) 7/13 (54%) 2/13 (15%) 6/13 (46%) 1/13 (8%) 6/13 (46%)
48 69.2 (12.3) 16/48 (33%) 4/48 (8%) 55.0 (53.860) 18.9 (5.4) 96.5 (80.3-109.9) 5/40 (13%) 5/40 (13%) 5/48 (10%) 10/48 (21%) 30/48 (63%) 2/48 (4%) 30/48 (63%)
0.41 0.48 0.04 0.02 o0.001 0.02 0.03 0.03 0.01 1.00 1.00 0.57 0.53
NOTE. Continuous variables are expressed as mean (standard deviation) for normally distributed variables and median (IQR) for non-normally distributed variables. Categoric variables are reported as n/range. Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; CT, computed tomography; ICU, intensive care unit; OR, operating room.
between groups (17.0 days [12.0 to 34.5] in the RS group v 12.5 days [6.3 to 28.3] in the isolated seizure group; p ¼ 0.09). To investigate what factors, including RS, were associated with survival, the authors performed a comparative analysis of patients who died while in-hospital or were discharged alive after surgery (Table 3). All patients who experienced one or more seizures after cardiac surgery were included in this analysis (n ¼ 61). Patients who died were more likely to have chronic obstructive pulmonary disease, lower ejection fraction, higher APACHE II scores, elevated preoperative creatinine, postoperative CVA, and have postoperative renal dysfunction. Interestingly, RS was not associated with in-hospital mortality (p ¼ 0.53). As seen in Figure 1, 83% of patients in the RS group and 76% of patients in the isolated seizure group were discharged from the hospital. At a median (IQR) follow-up period of 21 (6-30) months for the RS group and 16 (11-28) months for the isolated seizure group (p ¼ 0.79), a similar percentage of patients died during the follow-up period, and a similar percentage were restricted from driving at follow-up. One person suffered a stroke in the RS group during follow-up whereas one person experienced a seizure in the isolated seizure group. Twenty-three percent of patients in the RS group were readmitted to the hospital during the follow-up period as compared to 16% in the isolated seizure group, and 17% of patients in the RS group were still on anticonvulsants at the time of follow-up compared to 11% in the isolated seizure group. DISCUSSION
The risk factors and outcomes for patients who experience RS after cardiac surgery have not yet been defined. With this study, the authors examined the differences between patients who experienced isolated seizure or RS after cardiac surgery. They observed an RS rate of 59% in patients who had a primary seizure (0.6% of all cardiac surgery patients in their cohort). Patients with higher preoperative creatinine, patients
who undergo a procedure involving the thoracic aorta, and patients who experience an initial seizure onset fewer than 4 hours postoperatively were more likely to have RS. In addition, most patients with RS had grand mal seizures (81%), and the majority occurred within 24 hours of surgery. Mechanisms underlying postoperative seizures are not well understood. One possible mechanistic explanation that links RS to the risk factors identified in this study and previously published data4 is higher levels of proconvulsant drugs in the cerebrospinal fluid (CSF) in combination with pre-existing brain abnormalities. For instance, TXA is a renally cleared, proconvulsant antifibrinolytic drug that is an independent risk factor for post–cardiac surgery seizures.1,2,4,5,12–14 In addition, TXA-associated seizures occur in a dose-dependent manner and have a higher incidence in patients with a previous history of renal dysfunction.3,4 It is possible (although not proven in this study) that patients with pre-existing brain abnormalities and a disrupted blood-brain barrier (secondary to stroke or other cerebral embolic phenomena from thoracic aortic surgery) with high CSF levels of TXA and renal dysfunction are more susceptible to seizure, particularly as sedation is decreased postoperatively. In addition, when these patients have a seizure and are resedated, it may be that circulating TXA levels are still high when attempts to wean sedation are undertaken a second time, thus provoking another seizure. Eventually, patients eliminate TXA from their CSF, and the blood-brain barrier is restored; this may explain why there is no further seizure inhospital or even postdischarge in most patients. It is noted that the hospital length of stay for the RS and isolated seizure groups was similar despite ICU and mechanical ventilation time (likely related to sedation to prevent recurrent seizures) being different. Although TXA levels were not statistically significantly different between the isolated seizure and RS groups in the authors’ study, they do not know what the CSF TXA concentrations were. As dosing of TXA is an easily modifiable variable, clinicians should consider using the lowest dose of TXA that is optimal for the patient, especially if the
1210
MANJI ET AL
patient has renal dysfunction and is going to have thoracic aortic surgery. Obviously, TXA dosage must be balanced with the risks of bleeding and need for transfusions. As mentioned earlier, the authors’ study did not “prove” that there are indeed high TXA levels in CSF,11 but their study puts forward a possible hypothesis for future studies. Other factors of importance also include cerebral emboli from thoracic aortic surgery as well as inadequate cerebral protection during DHCA. Hence, care must be taken to minimize chances of cerebral emboli from thoracic aortic surgery and to ensure adequate hypothermia and consider selective cerebral perfusion during DHCA.15 Risk of RS after cardiac surgery and the need to continue antiseizure drugs (with side effects and drug interaction risks) are a complication that may restrict resumption of normal activity and may affect a patient’s quality of life. Intuitively, clinicians might equate an RS with a poor outcome and risk of further RS, thus compelling them to keep a patient on anticonvulsants for a prolonged period. In the authors’ study, however, most patients in the RS group were able to discontinue anticonvulsant therapy within days of their first seizure with no recurrence while in the hospital or at follow-up. This lends support to the hypothesis that RS are due to transient perioperative factors such as higher TXA levels in abnormal brains, and, thus, long-term treatment with anticonvulsants may not be warranted. Long-term follow-up showed no major significant differences in outcomes between patients with recurrent seizure and those with isolated seizure, suggesting that having an RS does not necessarily predict a poor prognosis. Lastly, upon comparative analysis of patients who died while in-hospital versus those who were discharged, the authors observed that patients having RS were not more likely to die. Factors important for death involved other comorbidities and complications such as poor ventricular function, severe respiratory disease, high APACHE II score, and stroke. It is important to note that in the postdischarge follow-up, patients with RS had mortality rates, readmission rates, requirements for anticonvulsant therapy, and driving restriction rates that generally were comparable to patients with isolated seizures, suggesting that RS do not increase morbidity or mortality. These data could influence in-hospital management of patients with RS.
As with any historical cohort study, there are limitations related to retrospectively analyzed data. However, the information in the authors’ databases is collected prospectively, and these databases undergo periodic audits to ensure validity. Additionally, the authors’ follow-up was by phone interview, and a small percentage of patients were lost to follow-up. Thus, it is possible that there were more patients who had RS postdischarge. Furthermore, the sample size was relatively small, and this may cause issues with model fit and instability. However, the purpose of the authors’ study was to develop an exploratory model and to suggest a possible mechanism that could identify important factors worth investigating in future studies. As such, the authors felt the best approach to the regression analysis was not using a stepwise approach but rather choosing variables that appeared clinically relevant and that had a p value o 0.1. Variables that were collinear (eg, DHCA with thoracic aortic surgery) were tested in the model and did not change the R2 value significantly; thus, the authors kept the variables with the larger event rates in the analysis. They obtained an R2 value of 0.53 and an area under the ROC curve of 0.859 suggesting that their model has some value in predicting seizure recurrence. Lastly, the treatment decisions for seizures were dependent on the individual physician caring for the patient at the time the seizure occurred and, as a result, were not standardized. Hence, it is possible that the RS rate may have been different if all patients with isolated seizure had been similarly treated. However, the authors’ data, analysis, and synthesis may provide clinicians with some additional information for clinical decision-making and long-term management as well as expanding on data currently available in the literature. In conclusion, the authors’ data suggest that several risk factors for recurrent postoperative seizure should be considered when caring for this patient subset. For patients who are administered anticonvulsant therapy, discontinuation of these medications might be considered for most patients while in the hospital followed by a period of observation. This may minimize accompanying side effects and resulting restrictions on activities. ACKNOWLEDGMENTS
The authors would like to acknowledge Brett M. Hiebert BSc (Statistics), MSc (Community Health Sciences) who provided assistance with statistical analyses.
REFERENCES 1. Sharma V, Katznelson R, Jerath A, et al: The association between tranexamic acid and convulsive seizures after cardiac surgery: A multivariate analysis in 11,529 patients. Anaesthesia 69:124-130, 2014 2. Bell D, Marasco S, Almeida A, et al: Tranexamic acid in cardiac surgery and postoperative seizures: A case report series. Heart Surg Forum 13:E257-E259, 2010 3. Kalavrouziotis D, Voisine P, Mohammadi S, et al: High-dose tranexamic acid is an independent predictor of early seizure after cardiopulmonary bypass. Ann Thorac Surg 93:148-154, 2012 4. Manji RA, Grocott HP, Leake J, et al: Seizures following cardiac surgery: The impact of tranexamic acid and other risk factors. Can J Anaesth 59:6-13, 2012 5. Martin K, Knorr J, Breuer T, et al: Seizures after open heart surgery: Comparison of epsilon-aminocaproic acid and tranexamic acid. J Cardiothorac Vasc Anesth 25:20-25, 2011
6. Hunter GR, Young GB: Seizures after cardiac surgery. J Cardiothorac Vasc Anesth 25:299-305, 2011 7. Lecker I, Orser BA, Mazer CD: “Seizing” the opportunity to understand antifibrinolytic drugs. Can J Anaesth 59:1-5, 2012 8. Newman MF, Mathew JP, Grocott HP, et al: Central nervous system injury associated with cardiac surgery. Lancet 368: 694-703, 2006 9. Samokhvalov AV, Irving H, Mohapatra S, et al: Alcohol consumption, unprovoked seizures, and epilepsy: A systematic review and meta-analysis. Epilepsia 51:1177-1184, 2010 10. Vigevano F, Arzimanoglou A, Plouin P, et al: Therapeutic approach to epileptic encephalopathies. Epilepsia 54(Suppl 8):45-50, 2013 11. Lecker I, Wang DS, Romaschin AD, et al: Tranexamic acid concentrations associated with human seizures inhibit glycine receptors. J Clin Invest 122:4654-4666, 2012
RECURRENT SEIZURES
12. Montes FR, Pardo DF, Carreño M, et al: Risk factors associated with postoperative seizures in patients undergoing cardiac surgery who received tranexamic acid: A case-control study. Ann Card Anaesth 15:6-12, 2012 13. Murkin JM, Falter F, Granton J, et al: High-dose tranexamic acid is associated with nonischemic clinical seizures in cardiac surgical patients. Anesth Analg 110:350-353, 2010
1211
14. Keyl C, Uhl R, Beyersdorf F, et al: High-dose tranexamic acid is related to increased risk of generalized seizures after aortic valve replacement. Eur J Cardiothorac Surg 39:e114-e121, 2011 15. Chan SK, Underwood MJ, Ho AM, et al: Cannula malposition during antegrade cerebral perfusion for aortic surgery: Role of cerebral oximetry. Can J Anaesth 61:736-740, 2014
hours; p ¼ 0.01), and procedures involving the thoracic aorta were associated with RS (R2 ¼ 0.53, p o 0.05). Patients with RS had longer intensive care unit stays (5.3 v 2.9 days, p ¼ 0.03) and longer mechanical ventilation duration (53.3 v 15.0 hours, p ¼ 0.01). At a median followup of 21 months for the RS group and 16 months for the isolated seizure group, restrictions, anticonvulsant use, morbidity, and mortality were similar between patients with isolated versus recurrent seizures. Conclusions: Higher preoperative creatinine, thoracic aortic surgery, and early seizure onset were associated with RS after cardiac surgery. When compared to isolated seizures, recurrence per se was not associated with significantly increased long-term morbidity or mortality. & 2015 Elsevier Inc. All rights reserved.
P
impact a patient’s quality of life regarding return-to-normal activity and the possibility of long-term anticonvulsant therapy. The purpose of this historical cohort study was to create an exploratory predictive model for seizure recurrence after cardiac surgery as well as to compare the outcomes for these patients compared to those with isolated seizure.
OSTOPERATIVE SEIZURES, although relatively uncommon, are a well-documented complication of cardiac surgery. The incidence of seizures after cardiac surgery varies between 0.5 and 7.6%.1 When seizures occur, recurrence rates of 40% to 66% have been reported.2–5 Seizures may be caused by thromboembolic ischemic stroke, cerebral air embolism, medication toxicity related to antibiotics, or other perioperative drugs such as tranexamic acid (TXA).6–8 The authors previously reported that administration of TXA, preoperative cardiac arrest, Acute Physiology and Chronic Health Evaluation (APACHE) II scores 420, previous cardiac surgery, openchamber procedure, cardiopulmonary bypass (CPB) time 4150 minutes, and preoperative neurologic disease are all independent risk factors for seizures after cardiac surgery.4 Importantly, approximately 60% of patients in that study had more than 1 seizure during their postcardiac surgery hospital stay, thereby defining recurrent seizure (RS). In the general non-surgical population, RS often is related to factors such as a history of epilepsy, subtherapeutic anticonvulsant levels, alcohol withdrawal, and structural brain abnormalities.9,10 However, associated risk factors and outcomes for RS after cardiac surgery are not known. Knowing these risk factors could influence in-hospital clinical decision making and
From the *Cardiac Sciences Program, Winnipeg Regional Health Authority and St. Boniface Hospital, †Department of Anesthesia and Perioperative Medicine; and ‡Department of Surgery, University of Manitoba, Winnipeg, Manitoba, Canada. Address reprint requests to Rizwan A. Manji, University of Manitoba, St. Boniface Hospital, CR3014 - 369 Tache Avenue, Winnipeg, Manitoba, Canada R2H 2A6. E-mail: [email protected] © 2015 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2015.03.020 1206
KEY WORDS: recurrent seizures, cardiac surgery, follow-up, anticonvulsants, tranexamic acid
METHODS
The local Research Ethics Board approved this study (January 2010) and waived the requirement for patient consent for the historical cohort component of the study. For follow-up, informed consent was obtained from patients prior to the telephone interview. This report represents a subset of patients from a previously published single-center historical cohort study performed on all patients who underwent cardiac surgery at the authors’ institution between April 1, 2003 and September 30, 2010 who had postoperative seizures.4 In the original study, the impact of TXA and other risk factors were compared between patients who did and did not experience postoperative seizure. For this present study, patients from the previous cohort who experienced post–cardiac surgery seizures were divided into 2 groups: those who had one postoperative isolated seizure and those who experienced RS (ie, more than one postoperative seizure while in the hospital). The reason for this division was to specifically explore the significance of recurrent seizures as compared to isolated seizures. Data collected from the authors’ cardiac surgery, perfusion service, and intensive care unit (ICU) databases included information on all consecutive cardiac surgery cases performed at their institution. The perfusion service database collects intraoperative variables, and the cardiac surgery and ICU databases collect demographic and comorbidity variables (including postoperative complications), ventilation time, and lengths of ICU and hospital stays. A concurrent pharmacy database maintains information related to the inventory of
Journal of Cardiothoracic and Vascular Anesthesia, Vol 29, No 5 (October), 2015: pp 1206–1211
1207
RECURRENT SEIZURES
drugs dispensed to various clinical services, including anesthesia for cardiac surgery.4 A seizure was defined as a physician-documented event such as rhythmic tonic-clonic motion of localized body parts consistent with a focal seizure, or of all four limbs with a decreased level of consciousness consistent with a grand mal seizure that required prescribed intervention. Nonconvulsive seizures were diagnosed in patients with a persistent decrease in the level of consciousness and confirmed seizure activity on serial electroencephalography (EEG). A telephone interview was conducted more than a year after discharge to ascertain if there had been any hospital readmission, new neurologic diagnoses (eg, RS and stroke), the need for long-term anticonvulsant therapy, and/or any activity restrictions (such as driving) instituted that were related to seizures. Statistical Analysis All statistical analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, IL). Univariate analysis was performed with a Student’s 2-sample t-test (for normally distributed data), the Mann–Whitney U test (for non-normally distributed data) and Pearson’s chi-square test (for categoric data) in which potential recurrent seizure risk factors and outcomes were compared between groups. Due to a relatively small sample size, the authors opted to create an exploratory (ie, hypothesis-generating) model for risk factors for RS that could serve as a basis for future studies. As such, multivariable logistic regression was performed using clinically relevant variables (that had a univariate p value o0.1) to create a
model with the highest R2 value and largest area under the receiver operating curve (ROC). A 2-tailed p value r 0.05 was considered statistically significant. RESULTS
Sixty-one (0.8%) of 7,280 consecutive patients experienced a postoperative seizure (Fig 1). Of these 61 patients, 36 (59%) had at least 1 RS, 17 patients (28%) experienced more than 1 recurrence, and 3 (8%) had non-convulsive status epilepticus. Of the 36 patients with RS, 32 (89%) experienced RS within 24 hours of the first seizure, and 4 patients (11%) experienced RS 424 hours after the first seizure. Grand mal seizures accounted for the majority of patients with RS (29/36; 81%), and each seizure (except for the 3 patients with nonconvulsive status epilepticus) lasted less than 5 minutes. A comparative univariate analysis between patients with RS and patients with isolated seizure is outlined in Table 1. Importantly, patients with RS had higher preoperative creatinine levels, higher APACHE II scores, more thoracic aortic and deep hypothermic circulatory arrest (DHCA), more postoperative cerebrovascular accidents (CVA), and a shorter period of time from the end of surgery until the first seizure occurrence (ie, early onset of the first seizure). Of note, patients with isolated seizures and patients with RS had similar doses of TXA given, although the authors did not know the serum (or cerebrospinal fluid levels11) of TXA. Variables associated with RS that were considered biologically plausible and that had a p value o 0.1 were further examined in the authors’ multivariate model. These variables included preoperative creatinine, APACHE II score, open-
Fig 1. Flow diagram of the historical cohort observed in this study comparing patients with isolated seizures to patients with recurrent seizures (RS). The median follow-up period was 21 (6-30) months for the recurrent seizure group and 16 (11-28) months for the isolated seizure group.
1208
MANJI ET AL
Table 1. Univariate Comparison Between Patients With Isolated or Recurrent Seizure After Cardiac Surgery Variable
Number of patients per group Age (years) Female Preoperative seizure disorder Preoperative alcohol abuse Preoperative cerebrovascular disease Ejection fraction (%) Preoperative hemoglobin (g/L) Preoperative creatinine (μmol/L) APACHE II score Preoperative cardiac arrest Preanesthetic use of benzodiazepine Intraoperative use of benzodiazepine Intraoperative use of propofol TXA dose (mg/kg) CABG only (including redo CABG) Open-chamber procedure (valve procedure, aortic procedure, intracardiac procedure) Procedure involving aortic valve Procedure involving thoracic aorta Deep hypothermic circulatory arrest during surgery Cross-clamp time (min) Cardiopulmonary bypass time (min) Lowest mean arterial pressure (mmHg) during OR Duration of lowest mean arterial pressure (min) Postoperative cerebrovascular accident Postoperative cardiac arrest New postoperative renal dysfunction Time from end of surgery until first seizure (hours)
Isolated Seizure
Recurrent Seizure
p Values
25 68.7 (13.0) 8/25 (32%) 0/25 (0%) 2/25 (8%) 5/25 (20%) 55.0 (55.0-60.0) 125 (17) 92.0 (73.0-103.0) 17.7 (6.6) 0/25 (0%) 2/20 (10%) 1/20 (5%) 17/20 (85%) 68.4 (36.1) 9/25 (36%) 15/25 (60%) 6/25 (24%) 1/25 (4%) 0/25 (0%) 75.0 (51.5-134.0) 124.0 (85.0-200) 51.6 (6.9) 10.0 (10.0-20.0) 1/25 (4%) 2/25 (8%) 6/25 (24%) 10.5 (4.8-17.5)
36 70.6 (10.6) 10/36 (28%) 2/36 (6%) 1/36 (3%) 8/36 (22%) 55.0 (45.0-60.0) 121 (22) 103.0 (93.5-136.5) 22.3 (5.8) 3/36 (8%) 6/36 (17%) 2/36 (6%) 27/36 (75%) 74.9 (34.3) 7/36 (19%) 29/36 (81%) 16/36 (44%) 11/36 (31%) 8/36 (22%) 77.0 (50.5-130.0) 165.0 (92.3-218.3) 49.9 (10.9) 10.0 (10.0-15.0) 8/36 (22%) 0/36 (0%) 5/36 (14%) 4.1 (1.8-10.6)
0.53 0.78 0.51 0.56 1.00 0.64 0.54 0.01 0.01 0.07 0.70 1.00 0.51 0.51 0.24 0.08 0.10 0.02 0.02 0.96 0.23 0.53 0.95 0.05 0.16 0.33 0.03
NOTE. Continuous variables reported as mean (SD) for normally distributed data or median (IQR) for non-normally distributed data. Categoric data are reported as n/range. Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CABG, coronary artery bypass graft; OR, operating room; TXA, tranexamic acid.
chamber procedure, procedure involving the thoracic aorta, and time duration from the end of surgery until the first seizure occurred. Preoperative cardiac arrest was not included in this analysis as the event rate was too low; only 3 patients with RS experienced this variable. “Deep hypothermic circulatory arrest during surgery” was the same as “procedures involving the thoracic aorta.” Thus, only the thoracic aortic procedure variable (with the larger event rate) was examined. Postoperative CVA also was the same as the thoracic aortic procedure variable and, thus, was not examined separately. The multivariate model (Table 2) demonstrated that elevated preoperative creatinine (Z120 μmol/L), procedure involving the thoracic aorta and early seizure onset (r4 h from the end of surgery until the first seizure) were associated with RS [R2 ¼ 0.53, p o 0.05, area under the ROC curve 0.859 (95% confidence intervals, 0.756-0.963)].
The median (IQR) duration of anticonvulsant therapy was significantly longer in the RS group (4.0 [3.0-5.8]) days) versus the isolated seizure group (0.3 [0.1-1.0] days; p o 0.001). However, at the discretion of the attending physicians, 29 (81%) of the 36 patients in the RS group and 19 (76%) of the 25 in the isolated seizure group had their anticonvulsant therapy discontinued during their hospitalization (p ¼ 0.69). Patients underwent extended periods of in-hospital observation (see below). No seizure occurred in-hospital after discontinuation of anticonvulsant therapy in either group. Patients in the RS group had significantly longer median (IQR) mechanical ventilation times than in the isolated seizure group (53.3 h [18.5 to 154.0] v 15.0 h [2.2 to 48.0] group; p ¼ 0.01), as well as significantly longer median (IQR) ICU stay (5.3 days [3.0 to 10.3] v 2.9 days [1.9 to 6.0]; p ¼ 0.03). However, the median (IQR) length of hospital stay was similar
Table 2. Logistic Regression of Potential Predictors of Recurrent Seizures After Cardiac Surgery Variable
Preoperative creatinine (Z120 μmol/L) Procedure involving thoracic aorta Time from end of surgery until first seizure r4 h
Unadjusted Odds Ratio
95% CI
Adjusted Odds Ratio
95% CI
2.6 10.6 4.6
0.7-9.6 1.3-88.2 1.1-19.2
11.3 Undefined† 11.0
1.6-81.8 Undefined† 1.9-65.5
NOTE. R2 ¼ 53%; Area under the ROC Curve ¼ 0.859 (0.756 – 0.963). †Undefined since odds ratio approaches infinity as almost all events occurred within the recurrent seizure group.
1209
RECURRENT SEIZURES
Table 3. In-Hospital Mortality and Live Discharge for All Patients Experiencing Seizure Post–Cardiac Surgery (Includes Patients With Isolated and Recurrent Seizures) Variable
Total no. of patients Age (years) Female COPD Ejection fraction (%) APACHE II score Preoperative creatinine (μmol/L) Major stroke (watershed or large vessel) on CT head New stroke on CT head New postoperative renal dysfunction Focal seizure Grand mal seizure Nonconvulsive seizure Recurrent seizure
In-hospital Mortality
Live Discharge
p Value
13 72.3 (8.2) 2/12 (17%) 4/12 (33%) 42.5 (32.5-55.0) 26.7 (7.0) 122.0 (100.8-161.5) 4/8 (50%) 4/8 (50%) 7/13 (54%) 2/13 (15%) 6/13 (46%) 1/13 (8%) 6/13 (46%)
48 69.2 (12.3) 16/48 (33%) 4/48 (8%) 55.0 (53.860) 18.9 (5.4) 96.5 (80.3-109.9) 5/40 (13%) 5/40 (13%) 5/48 (10%) 10/48 (21%) 30/48 (63%) 2/48 (4%) 30/48 (63%)
0.41 0.48 0.04 0.02 o0.001 0.02 0.03 0.03 0.01 1.00 1.00 0.57 0.53
NOTE. Continuous variables are expressed as mean (standard deviation) for normally distributed variables and median (IQR) for non-normally distributed variables. Categoric variables are reported as n/range. Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II score; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; CT, computed tomography; ICU, intensive care unit; OR, operating room.
between groups (17.0 days [12.0 to 34.5] in the RS group v 12.5 days [6.3 to 28.3] in the isolated seizure group; p ¼ 0.09). To investigate what factors, including RS, were associated with survival, the authors performed a comparative analysis of patients who died while in-hospital or were discharged alive after surgery (Table 3). All patients who experienced one or more seizures after cardiac surgery were included in this analysis (n ¼ 61). Patients who died were more likely to have chronic obstructive pulmonary disease, lower ejection fraction, higher APACHE II scores, elevated preoperative creatinine, postoperative CVA, and have postoperative renal dysfunction. Interestingly, RS was not associated with in-hospital mortality (p ¼ 0.53). As seen in Figure 1, 83% of patients in the RS group and 76% of patients in the isolated seizure group were discharged from the hospital. At a median (IQR) follow-up period of 21 (6-30) months for the RS group and 16 (11-28) months for the isolated seizure group (p ¼ 0.79), a similar percentage of patients died during the follow-up period, and a similar percentage were restricted from driving at follow-up. One person suffered a stroke in the RS group during follow-up whereas one person experienced a seizure in the isolated seizure group. Twenty-three percent of patients in the RS group were readmitted to the hospital during the follow-up period as compared to 16% in the isolated seizure group, and 17% of patients in the RS group were still on anticonvulsants at the time of follow-up compared to 11% in the isolated seizure group. DISCUSSION
The risk factors and outcomes for patients who experience RS after cardiac surgery have not yet been defined. With this study, the authors examined the differences between patients who experienced isolated seizure or RS after cardiac surgery. They observed an RS rate of 59% in patients who had a primary seizure (0.6% of all cardiac surgery patients in their cohort). Patients with higher preoperative creatinine, patients
who undergo a procedure involving the thoracic aorta, and patients who experience an initial seizure onset fewer than 4 hours postoperatively were more likely to have RS. In addition, most patients with RS had grand mal seizures (81%), and the majority occurred within 24 hours of surgery. Mechanisms underlying postoperative seizures are not well understood. One possible mechanistic explanation that links RS to the risk factors identified in this study and previously published data4 is higher levels of proconvulsant drugs in the cerebrospinal fluid (CSF) in combination with pre-existing brain abnormalities. For instance, TXA is a renally cleared, proconvulsant antifibrinolytic drug that is an independent risk factor for post–cardiac surgery seizures.1,2,4,5,12–14 In addition, TXA-associated seizures occur in a dose-dependent manner and have a higher incidence in patients with a previous history of renal dysfunction.3,4 It is possible (although not proven in this study) that patients with pre-existing brain abnormalities and a disrupted blood-brain barrier (secondary to stroke or other cerebral embolic phenomena from thoracic aortic surgery) with high CSF levels of TXA and renal dysfunction are more susceptible to seizure, particularly as sedation is decreased postoperatively. In addition, when these patients have a seizure and are resedated, it may be that circulating TXA levels are still high when attempts to wean sedation are undertaken a second time, thus provoking another seizure. Eventually, patients eliminate TXA from their CSF, and the blood-brain barrier is restored; this may explain why there is no further seizure inhospital or even postdischarge in most patients. It is noted that the hospital length of stay for the RS and isolated seizure groups was similar despite ICU and mechanical ventilation time (likely related to sedation to prevent recurrent seizures) being different. Although TXA levels were not statistically significantly different between the isolated seizure and RS groups in the authors’ study, they do not know what the CSF TXA concentrations were. As dosing of TXA is an easily modifiable variable, clinicians should consider using the lowest dose of TXA that is optimal for the patient, especially if the
1210
MANJI ET AL
patient has renal dysfunction and is going to have thoracic aortic surgery. Obviously, TXA dosage must be balanced with the risks of bleeding and need for transfusions. As mentioned earlier, the authors’ study did not “prove” that there are indeed high TXA levels in CSF,11 but their study puts forward a possible hypothesis for future studies. Other factors of importance also include cerebral emboli from thoracic aortic surgery as well as inadequate cerebral protection during DHCA. Hence, care must be taken to minimize chances of cerebral emboli from thoracic aortic surgery and to ensure adequate hypothermia and consider selective cerebral perfusion during DHCA.15 Risk of RS after cardiac surgery and the need to continue antiseizure drugs (with side effects and drug interaction risks) are a complication that may restrict resumption of normal activity and may affect a patient’s quality of life. Intuitively, clinicians might equate an RS with a poor outcome and risk of further RS, thus compelling them to keep a patient on anticonvulsants for a prolonged period. In the authors’ study, however, most patients in the RS group were able to discontinue anticonvulsant therapy within days of their first seizure with no recurrence while in the hospital or at follow-up. This lends support to the hypothesis that RS are due to transient perioperative factors such as higher TXA levels in abnormal brains, and, thus, long-term treatment with anticonvulsants may not be warranted. Long-term follow-up showed no major significant differences in outcomes between patients with recurrent seizure and those with isolated seizure, suggesting that having an RS does not necessarily predict a poor prognosis. Lastly, upon comparative analysis of patients who died while in-hospital versus those who were discharged, the authors observed that patients having RS were not more likely to die. Factors important for death involved other comorbidities and complications such as poor ventricular function, severe respiratory disease, high APACHE II score, and stroke. It is important to note that in the postdischarge follow-up, patients with RS had mortality rates, readmission rates, requirements for anticonvulsant therapy, and driving restriction rates that generally were comparable to patients with isolated seizures, suggesting that RS do not increase morbidity or mortality. These data could influence in-hospital management of patients with RS.
As with any historical cohort study, there are limitations related to retrospectively analyzed data. However, the information in the authors’ databases is collected prospectively, and these databases undergo periodic audits to ensure validity. Additionally, the authors’ follow-up was by phone interview, and a small percentage of patients were lost to follow-up. Thus, it is possible that there were more patients who had RS postdischarge. Furthermore, the sample size was relatively small, and this may cause issues with model fit and instability. However, the purpose of the authors’ study was to develop an exploratory model and to suggest a possible mechanism that could identify important factors worth investigating in future studies. As such, the authors felt the best approach to the regression analysis was not using a stepwise approach but rather choosing variables that appeared clinically relevant and that had a p value o 0.1. Variables that were collinear (eg, DHCA with thoracic aortic surgery) were tested in the model and did not change the R2 value significantly; thus, the authors kept the variables with the larger event rates in the analysis. They obtained an R2 value of 0.53 and an area under the ROC curve of 0.859 suggesting that their model has some value in predicting seizure recurrence. Lastly, the treatment decisions for seizures were dependent on the individual physician caring for the patient at the time the seizure occurred and, as a result, were not standardized. Hence, it is possible that the RS rate may have been different if all patients with isolated seizure had been similarly treated. However, the authors’ data, analysis, and synthesis may provide clinicians with some additional information for clinical decision-making and long-term management as well as expanding on data currently available in the literature. In conclusion, the authors’ data suggest that several risk factors for recurrent postoperative seizure should be considered when caring for this patient subset. For patients who are administered anticonvulsant therapy, discontinuation of these medications might be considered for most patients while in the hospital followed by a period of observation. This may minimize accompanying side effects and resulting restrictions on activities. ACKNOWLEDGMENTS
The authors would like to acknowledge Brett M. Hiebert BSc (Statistics), MSc (Community Health Sciences) who provided assistance with statistical analyses.
REFERENCES 1. Sharma V, Katznelson R, Jerath A, et al: The association between tranexamic acid and convulsive seizures after cardiac surgery: A multivariate analysis in 11,529 patients. Anaesthesia 69:124-130, 2014 2. Bell D, Marasco S, Almeida A, et al: Tranexamic acid in cardiac surgery and postoperative seizures: A case report series. Heart Surg Forum 13:E257-E259, 2010 3. Kalavrouziotis D, Voisine P, Mohammadi S, et al: High-dose tranexamic acid is an independent predictor of early seizure after cardiopulmonary bypass. Ann Thorac Surg 93:148-154, 2012 4. Manji RA, Grocott HP, Leake J, et al: Seizures following cardiac surgery: The impact of tranexamic acid and other risk factors. Can J Anaesth 59:6-13, 2012 5. Martin K, Knorr J, Breuer T, et al: Seizures after open heart surgery: Comparison of epsilon-aminocaproic acid and tranexamic acid. J Cardiothorac Vasc Anesth 25:20-25, 2011
6. Hunter GR, Young GB: Seizures after cardiac surgery. J Cardiothorac Vasc Anesth 25:299-305, 2011 7. Lecker I, Orser BA, Mazer CD: “Seizing” the opportunity to understand antifibrinolytic drugs. Can J Anaesth 59:1-5, 2012 8. Newman MF, Mathew JP, Grocott HP, et al: Central nervous system injury associated with cardiac surgery. Lancet 368: 694-703, 2006 9. Samokhvalov AV, Irving H, Mohapatra S, et al: Alcohol consumption, unprovoked seizures, and epilepsy: A systematic review and meta-analysis. Epilepsia 51:1177-1184, 2010 10. Vigevano F, Arzimanoglou A, Plouin P, et al: Therapeutic approach to epileptic encephalopathies. Epilepsia 54(Suppl 8):45-50, 2013 11. Lecker I, Wang DS, Romaschin AD, et al: Tranexamic acid concentrations associated with human seizures inhibit glycine receptors. J Clin Invest 122:4654-4666, 2012
RECURRENT SEIZURES
12. Montes FR, Pardo DF, Carreño M, et al: Risk factors associated with postoperative seizures in patients undergoing cardiac surgery who received tranexamic acid: A case-control study. Ann Card Anaesth 15:6-12, 2012 13. Murkin JM, Falter F, Granton J, et al: High-dose tranexamic acid is associated with nonischemic clinical seizures in cardiac surgical patients. Anesth Analg 110:350-353, 2010
1211
14. Keyl C, Uhl R, Beyersdorf F, et al: High-dose tranexamic acid is related to increased risk of generalized seizures after aortic valve replacement. Eur J Cardiothorac Surg 39:e114-e121, 2011 15. Chan SK, Underwood MJ, Ho AM, et al: Cannula malposition during antegrade cerebral perfusion for aortic surgery: Role of cerebral oximetry. Can J Anaesth 61:736-740, 2014
