Postcardiac Surgical Cognitive Impairment in the Aged Using Diffusion-Weighted Magnetic Resonance Imaging

Transcription

1 Postcardiac Surgical Cognitive Impairment in the Aged Using Diffusion-Weighted Magnetic Resonance Imaging David J. Cook, MD, John Huston III, MD, Max R. Trenerry, PhD, Robert D. Brown, Jr, MD, Kenton J. Zehr, MD, and Thoralf M. Sundt III, MD Departments of Anesthesiology, Radiology, Psychology, Neurology, and Surgery, Mayo Clinic College of Medicine, Rochester, Minnesota; Division of Cardiac Surgery, University of Pittsburgh Presbyterian Medical Center Pittsburgh, Pennsylvania Background. Cardiac surgery is associated with cerebral dysfunction. While 1% to 2% of patients experience stroke, cognitive deficits are seen in more than half of patients. Given the high incidence of cognitive decline, it has become the endpoint of many cardiac surgery investigations. Because the elderly are at highest risk, this investigation sought to determine if there is a relationship between new ischemic changes demonstrated by diffusion-weighted magnetic resonance imaging (DW- MRI) and postoperative cognitive deficit in older patients. Methods. Fifty cardiac surgical patients (>65 years of age) underwent preoperative and postoperative neurocognitive examinations, including four to six week, postdischarge, follow-up. This evaluation assessed higher cortical function, memory, attention, concentration, and psychomotor performance. Objective evidence of acute cerebral ischemic events was identified using DW-MRI. Scans were analyzed by a neuroradiologist blinded to clinical status and cognitive outcomes. Results. Among patients with a mean age of 73 years, 88% demonstrated cognitive decline in the postoperative testing period while 32% showed evidence of acute perioperative cerebral ischemia by DW-MRI. At postdischarge follow-up, 30% of patients showed cognitive impairment. However, cognitive decline assessed postoperatively, or at a four to six week follow-up, was unrelated to the presence or absence of DW-MRI detected cerebral ischemia. Conclusions. Postoperative neurocognitive impairment, assessed by standard means, is unrelated to acute cerebral ischemia detected by DW-MRI. This strongly suggests that cognitive decline after cardiac surgery is a function of underlying patient factors rather than perioperative ischemic events. This observation has broad implications for future investigation of strategies to prevent cardiac surgery-related neurologic injury. (Ann Thorac Surg 2007;83: ) 2007 by The Society of Thoracic Surgeons Accepted for publication Nov 28, Address correspondence to Dr Cook, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, Minnesota 55905; cook.david@ mayo.edu. Improvements in cardiac surgical care have allowed for a broader focus in assessing outcomes. In addition to severe acute events such as stroke, it is increasingly common to include more functional endpoints such as cognitive decline. This is probably also attributable to the aging of the cardiac surgical population, a shift in the type of morbidity seen, popular awareness of cognitive decline after cardiac surgery, and its relative ease as a study endpoint given its high frequency [1 5]. The incidence of cognitive impairment is dependent on the age of the patient and the timing and intensity of the assessment, but commonly more than half of patients will show postcardiac surgical cognitive decline [3 5]. Accordingly, efforts to reduce neurologic injury such as performance of coronary artery bypass surgery without the use of cardiopulmonary bypass, modifications of standard techniques such as elimination of partial occlusion clamps, and the introduction of advanced technology such as arterial line filters or intraaortic baffles and screens have increasingly used sophisticated neurocognitive testing as an endpoint for assessment of neurologic injury [6 8]. For cardiac surgical neurologic morbidity, it is important to appreciate that cardiac surgery concentrates the risk factors for stroke and cognitive impairment in the general population. While the etiology of brain injury is multifactorial, most evidence suggests that outcome is largely dependent on intraoperative embolization and possibly the functional reserve of the patient s cerebral circulation [9 11]. Because stroke is infrequent, it is a difficult endpoint to study. In contrast, cognitive impairment is common and therefore most of the research on clinical neurologic outcomes in cardiac surgery over the last several years has relied heavily on cognitive decline. This focus has gained momentum because of public concern and the attention of the media to the problem [12, 13]. This investigation applies two very different tools to understand postcardiac surgical cognitive injury. First, older ( 65 years) cardiac surgical patients underwent a 2007 by The Society of Thoracic Surgeons /07/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 1390 COOK ET AL Ann Thorac Surg COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE 2007;83: Table 1. Demographics-Magnetic Resonance Imaging Condition Infarction (n 16) No Infarction (n 34) p Value Age (mean) HTN (%) Hyperlipidemia (%) CAD (%) IDDM (%) COPD (%) Hx of smoking (%) Previous MI (%) Hx of stroke (%) Hx of arrhythmia (%) CAD coronary artery disease; COPD chronic obstructive pulmonary disease; HTN hypertension; Hx history of; IDDM insulin dependent diabetes mellitus; MI myocardial infarction. 10 test neurocognitive battery preoperatively, postoperatively before hospital discharge, and four to six weeks after surgery. In addition to cognitive testing, patients underwent brain diffusion weighted-magnetic resonance imaging (DW-MRI) after surgery. In contrast to standard neuroimaging modalities, which are typically negative in acute postoperative ischemia, impairment of water diffusion occurs minutes after onset of ischemia, resulting in a high signal intensity with DW-MRI [14]. The sensitivity and specificity of DWI for acute subcortical infarction has been reported at 95% and 94%, respectively, with positive and negative predictive values of 97% and 89% [15]. The resolution for ischemia is in the 2 to 3 millimeter range [16, 17]. The DW-MRI predicts acute and chronic neurologic outcomes in stroke and is able to detect subclinical ischemia [15]. Accordingly, this is the most sophisticated technology for objective assessment of acute postoperative structural neurologic injury. This technology is particularly useful in a cardiac surgical trial because a single postoperative scan can detect and differentiate acute and chronic cerebral ischemic events [15, 18]. Indirectly, a single scan can also reduce the likelihood of patient dropout in contrast to studies where multiple scans are required. Neurologic injury in cardiac surgery is assumed to be a continuum from frank cortical stroke, to encephalopathy, to cognitive deficit, with embolization thought to be the primary etiology. This investigation tests the hypothesis that postcardiac surgery cognitive decline is a function of subclinical acute ischemic injury, whether focal-embolic in origin or from hypoperfusion. Patients and Methods After Institutional Review Board approval and informed consent, 50 patients, age 65 years or greater undergoing elective cardiac surgery with cardiopulmonary bypass, were studied. Surgical procedures included coronary artery bypass grafting (CABG), valve procedures, and combined operations. Patients were excluded if they had history of head trauma; seizures, a preoperative National Institutes of Health stroke scale (NIHSS) score 1 or greater; stroke within one month of admission, if they were unable to participate in neurocognitive assessments or MRI, or if a preoperative Mini-Mental State Exam showed clinically evident cognitive impairment. A standardized 10-test neurocognitive battery was conducted preoperatively, prior to hospital discharge, and four to six weeks postoperatively. The NIHSS was administered on the same timetable; total score of 2 points or greater was considered clinically meaningful. Cognitive impairment was defined as a 20% or greater decline from baseline on two or more tests [19]. The battery was designed in collaboration with a clinical psychologist with expertise in neurocognitive assessment and included tests to interrogate domains of memory (Rey Auditory Verbal Learning Test*, Non-verbal Memory Test); attention, concentration and psychomotor performance (Symbol-Digit Modalities Test, Letter Cancellation Task, Trail Making Forms A and B*); and manual dexterity and fine motor/motor fatigue (Grooved Pegboard Test*, dominant and nondominant hands and Finger Tapping Test, dominant and nondominant hands). The tests chosen included all of those (indicated by *) recommend by the consensus conference on cognitive testing in cardiac surgery [19]. For robustness, each cognitive domain was tested with more than one test. The tests chosen reflected consideration of the cognitive domain, the sensitivity and reliability of the test, the time required, and the availability of standardized forms and duplicate versions. All testing and grading was conducted by trained psychometrists with results reviewed by a cognitive psychologist. DWI Imaging and Analysis Cerebral infarction was detected using DW-MRI prior to hospital discharge as early postoperatively as clinical status allowed. A high field strength (1.5T) MRI unit, equipped with high speed gradients capable of echo planar imaging, was used. Four sets of images, including T1 weighted sagittal, T2 weighted axial, trace weighted DWI axial, and apparent diffusion coefficient (ADC) Table 2. Surgical Data Number of Patients Cross-Clamp Time (Minutes) Bypass Time (Minutes) Ventilation Time (Hours) ICU Time (Hours) Duration of Hospitalization (Days) Total (n 50) Acute infarction (n 16) No infarction (n 34) ICU intensive care unit.

3 Ann Thorac Surg COOK ET AL 2007;83: COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE 1391 maps were reviewed by the radiologist blinded to the clinical and cognitive status of the patients. If acute ischemic changes were present, the number, the volumes, and the anatomic locations were noted. Preoperative scanning was not indicated as the focus of the study was the onset of new neurocognitive deficits and their correlation with new anatomic injuries as identified by DW-MRI, and because the combination of diffusion and standard MRI in a single session allows acute and chronic cerebral ischemic events to be differentiated. Data Analysis The percentage of patients with new cerebral infarctions was estimated using a point-estimate and exact 95% confidence interval. A power analysis demonstrated that a study size of 50 patients would have 96% of the power to test our hypothesis at a 0.05 level. For patients identified with acute infarction, the number, size, and location were summarized. In all cases the cognitive function measures were treated as continuous variables. Separate analyses were performed for each cognitive function test. In all cases, distributional assumptions required for model fitting were assessed and transformations (eg, log) used as appropriate. Where appropriate, the Fisher exact test was used to test for a difference in the proportion of abnormal results based on a positive or negative diffusion. Results The study continued until 50 patients completed three episodes of neurocognitive testing and a postoperative DW-MRI. To reach fifty patients, 54 were enrolled. Patient demographics are provided in Table 1. Based on results of the postoperative DW-MRI, patients were divided into two groups, identified as positive or negative for acute cerebral infarction. The mean age of the study population was 73 5 years. The patients with and without acute infarction did not differ with regard to age, or any comorbid or preexisting health conditions. Surgical details are provided in Table 2. Three surgeons participated and conduct of cardiopulmonary bypass, physiologic management, and administration of Table 4. Anatomic Locations and Number of Lesions of Acute DW-MRI Changes in the Sixteen Patients Having Positive Scans Anatomic Location Right Left Frontal 17 6 Parietal 7 7 Temporal 5 0 Occipital 5 6 Cerebellum 4 4 Thalamic 1 0 Basal ganglia 0 1 DWI diffusion-weighted. Contingency analysis and McNemar test were unable to demonstrate any difference in sidedness overall or for any particular region. cardioplegia, etc, were consistent among surgeons. Proximal anastomoses during CABG were completed with a partial occlusion clamp except when in conjunction with aortic valve replacement. Transesophageal echocardiography was utilized in all patients undergoing valve surgery. Epiaortic ultrasonography was not utilized. There were no apparent differences in cross-clamp time, bypass time, or ventilation time between the ischemia and no cerebral ischemia groups. A breakdown of the study group by procedure is provided in Table 3 with notation regarding the incidence of neurocognitive deficit and acute ischemia by DW-MRI. Approximately one-half of the study group underwent CABG. The mean age of this group was similar to that of patients undergoing valve repair or Table 3. Procedure-Magnetic Resonance Imaging (MRI) CABG AVR or MVR CABG & AVR or MVR TVR Number of patients Age (mean) Postoperative NC deficit (%) 4 6 weeks NC Deficit (%) Acute infarction MRI (%) AVR aortic valve replacement; CABG coronary artery bypass grafting; MVR mitral valve repair or replacement; NC neurocognitive; TVR tricuspid valve repair. Fig 1. Diffusion-weighted magnetic resonance imaging slice showing typical small, focal regions expected in cerebral embolization. Two right parietal defects are shown (arrow).

4 1392 COOK ET AL Ann Thorac Surg COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE 2007;83: Fig 2. Incidence of cognitive decline in patients with and without acute cerebral infarction. ( e ischemia; no ischemia.) replacement. The incidence of neurocognitive deficit, either immediately postoperatively or at four to six weeks, did not differ among procedure groups. Detectable acute ischemia was somewhat more frequent among open chamber procedures, although this did not reach statistical significance. For early neurocognitive outcomes, the Fisher exact p value testing for differing outcomes between surgical groups was 1.000; at four to six week follow-up, the p value was Surgical groups also had similar incidences of acute ischemic change (p for the Fisher exact test looking for any between group differences.). Table 4 summarizes DW-MRI results indicating the number and location of ischemic foci. Figure 1 shows a diffusion-weighted MRI slice typical of those showing acute ischemic change. The mean timing of DW-MRI was days postoperatively. Defects were small, focal, and multiple. There were 63 ischemic regions in 16 patients. The group mean was 4 5 regions per patient; three of 16 patients had greater than five areas of focal infarction. Of the 63 defects, only three were greater than 10 mm in diameter. The total ischemic volume was less than 1,000 mm 3 in 11 of 16 patients. Contingency analysis and the McNemar test were unable to demonstrate any difference in sidedness overall or for any particular region (Table 4). Of the 50 patients, four patients had an NIHSS score 2 or greater on the initial postoperative assessment. Three of the four had a negative DW-MRI. The fourth patient, with a NIHSS score of 15, showed three small ischemic foci: left centrum semi ovalis (75 mm 3 ), left cerebellar hemisphere (135 mm 3 ), and (128 mm 3 ), in the left occipital lobe. At the four to six week follow-up the NIHSS score for this patient was 0. Overall, 88% of patients demonstrated a decline in cognitive status in the postoperative testing period. At four to six week evaluation, 30% of patients showed cognitive decline relative to preoperative testing; however, only nine of 15 patients showed the same deficits as at their early postoperative testing. As shown in Figure 2, the incidence of neurocognitive dysfunction was similar at both time points regardless of the presence or absence of anatomic abnormalities detectable by DW-MRI. Sixteen of 50 patients showed new postoperative cerebral infarction. Of the 34 patients showing no evidence of acute perioperative infarction, 88% had cognitive deficits on in-hospital postoperative testing. Similarly, of the 16 patients showing acute ischemic changes, 88% also showed early new cognitive deficits (Fig 2). As such, the incidence of postoperative cognitive decline was the same whether or not acute cerebral ischemia occurred. Neurocognitive assessment at four to six weeks also failed to show any relationship between neurocognitive decline and acute ischemic change. Overall, 30% of patients met the criteria for cognitive decline at both predischarge testing and at four to six week follow-up; however, of the 15 patients demonstrating neurocognitive decline at four to six weeks only two had a positive early postoperative DW-MRI. Conversely, among the 16 patients with demonstrable ischemia by DW-MRI, only two had demonstrable cognitive decline at four to six weeks. Therefore, the incidence of neurocognitive decline at four to six week follow-up is the same whether or not the patient experienced perioperative cerebral ischemia. As such, there does not appear to be any relationship between early or persistent cognitive change and acute cerebral ischemic events after cardiac surgery. Table 5. Neurocognitive and DWI Results Predischarge and at Four to Six Weeks DWI T-A T-B LC NVM SD FTd FTnd GPd GPnd Rey Postoperative (n 34) p (n 16) week (n 34) p (n 16) DWI diffusion-weighted I; FTd/nd finger tapping dominant and nondominant; GPd/nd grooved peg board test, dominant and nondominant; LC letter cancelation test; NVM nonverbal memory test; Rey auditory verbal learning test; T-A trials A test; T-B trials B test; SD symbol digit test. Percentage of patients in early postoperative and four to six week follow-up with cognitive decline on 10 tests divided into groups with and without acute cerebral ischemic events.

5 Ann Thorac Surg COOK ET AL 2007;83: COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE 1393 Table 5 shows the neurocognitive test results postoperatively and at follow-up, dividing patients into two groups based on the presence or absence of cerebral infarction. There was no difference in the incidence of cognitive decline between groups in the early postoperative assessment on any of the ten tests. At follow up, between-group comparison shows differences in cognitive decline in five of 10 tests (Table 5). For trails A and B, the Symbol Digit test and Grooved Pegboard, both dominant and nondominant hand, patients who showed a positive postoperative DW-MRI had better cognitive outcomes than patients who did not (Table 5), further indicating an absence of a relationship between infarction and postoperative cognitive status. Comment The principal finding of this study is the absence of correlation between postoperative cognitive dysfunction and objective evidence of structural ischemia as detected by diffusion-weighted MRI. This finding has important implications both for the understanding of the mechanisms of cognitive dysfunction postoperatively and for the design of trials of techniques, technologies, and pharmacologic agents directed toward reducing cerebroembolic injury or improving cognitive outcomes. Concern regarding cognitive decline after cardiac surgery rose to the cardiac surgical consciousness after a 1987 report identifying a greater than 60% incidence [3]. Over the ensuing two decades attention to this complication increased as patients, families, nurses, and physicians have become aware of it, and particularly after a 2001 publication demonstrating that cognitive dysfunction early after cardiac surgery predicted decline over the next five years [5]. Immediately, the lay press raised the public and professional awareness [12] spurring clinical investigation and commercial activity directed toward improving cognitive outcomes. In addition to the high public profile of cognitive decline, postcardiac surgical cognitive dysfunction has had appeal as an investigational endpoint (in contrast to a low frequency event like stroke) because outcome trials, of adequate statistical power, can be conducted with much smaller numbers of patients. As such, cognitive endpoints have been used in studies of cardiac surgical techniques [6, 20], neuroprotective agents [21, 22], anesthesia and cardiopulmonary bypass (CPB) management [23, 24], as well as cardiac surgical devices [7]. Perhaps the most important trial using a cognitive endpoint tested the hypothesis that elimination of cardiopulmonary bypass would improve neurocognitive outcome [6]. However, that large randomized trial [6] could not show a meaningful effect of eliminating bypass on cognitive outcome. A second randomized study on the effect of off-pump surgery in CABG did conclude that elimination of CPB improved neurologic outcome [25]. In a study of approximately 160 patients (80 per group), the off-pump group had better scores in 3 of 13 cognitive assessment at six weeks, and had better scores in 2 of 13 tests at six months. While this study concluded that off-pump has better cognitive outcomes, this seems to be an overstatement. It might have been equally valid to conclude that conduct of CABG on- or off-pump has minimal effect on cognitive results, or better that cognitive outcome is primarily a function of factors unrelated to cardiopulmonary bypass. In fact, attempts to reduce adverse cognitive outcomes with multiple perioperative interventions (including mean arterial pressure [26], bypass temperature [24, 27, 28], CO 2 and glucose management [19, 23, 29], pulsatility [19], drug neuroprotectants [21, 22], surgical techniques [6, 20, 30], and surgical devices) have failed. To understand this paradox, we hypothesized that cognitive impairment after cardiac surgery was a function of subclinical cerebral infarction detectable by DW-MRI imaging. The findings of this study cannot support this hypothesis, calling into question the use of cognitive decline as an endpoint in the assessment of strategies to reduce cerebral embolic load or to minimize intraoperative cerebral ischemia. We found that cognitive decline after cardiac surgery is unrelated to acute ischemic events. With intensive testing, nearly 90% of patients showed cognitive deficits in the early postoperative period; however, only 32% of patients showed any evidence of cerebral ischemia on DW-MRI. The incidence of cognitive decline was the same in patients who did, and did not, show cerebral ischemia. The postdischarge follow-up data also showed no evidence of a relationship between perioperative ischemia and postoperative cognitive decline. The addition of a nonsurgical control group, such as interventional cardiology patients, might have added interest to this investigation; from a nonsurgical control group the effect of surgery might be partially isolated from patient factors. However, that design would not have added to the testing of the hypothesis that postoperative cognitive decline is a function of perioperative ischemic events detectable by diffusion-mri. Cognitive assessment, before and after surgery, correlated with new, perioperative, diffusion MRI findings was both the necessary and sufficient condition to test the hypothesis. One could criticize the DW-MRI technology as insufficiently sensitive to detect very small cerebral ischemic events. While this is theoretically possible, diffusion MRI is the best neuroimaging modality to detect acute cerebral ischemic events; it is highly sensitive and specific. Imaging was conducted at a time when acute ischemic events would be evident. Furthermore, the fact that patients demonstrated the same incidence of cognitive decline in the presence or absence of ischemia, both acutely and at follow-up, suggests much more strongly that cognitive decline is due to a process other than acute ischemic events, rather than that the DWI-MRI technology is not sufficiently sensitive to detect those events. The study might have been somewhat strengthened by a follow-up DW-MRI at six weeks. A second MRI would have revealed any new events occurring between discharge and four to six week follow-up. That said, a second MRI could not change the conclusion that postoperative cognitive decline is not a function of perioper-

6 1394 COOK ET AL Ann Thorac Surg COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE 2007;83: ative ischemia. Any perioperative ischemia that occurred was documented by the postoperative MRI. Since there was no relationship between cognitive decline and perioperative ischemic events, it would not be relevant if a later MRI showed resolution of ischemic change or new defects. It might also be argued that combining open and closed ventricle procedures makes it more difficult to interpret our results and that the study might have been better had a single type of surgery constituted our study population. We would argue that including patients having both open and closed ventricle procedures adds to the strength of this investigation. The hypothesis was to determine if perioperative cerebral ischemic events, particularly embolic, are related to neurocognitive outcome. To test that hypothesis it does not matter what type of cardiac surgery was done. In fact, that there were no differences in NC outcome between open and closed chamber procedures lends further, indirect, evidence to our speculation that preoperative cerebral status is more important than intraoperative events. Other than ruling out acute ischemia, this investigation was not designed to identify the cause of postoperative cognitive impairment. However, because of identification of significant cognitive decline after noncardiac major surgery [31, 32], similar longer term outcomes in interventional cardiology and CABG patients [33], and the absence of a meaningful effect of eliminating cardiopulmonary bypass [34], the logical working hypothesis should be that acute cognitive decline is a function of the patient s preexisting neurologic status interacting with the stresses associated with the perioperative period. Hospitalization stresses may unmask preexisting cognitive impairment [35] or eliminate compensatory mechanisms in older patients. Perioperative stresses can impair performance on cognitive testing either in themselves or because they might unmask underlying processes such as early dementia. This is consistent with patterns of early postoperative improvement [36], and improvement between three months and one year after surgery [4, 37 39]. While the Newman study [5] concluded that cardiac surgery and bypass may be responsible for long-term cognitive decline, an alternative interpretation might be that postoperative cognitive decline is a marker of senile cerebrovascular disease rather than the result of cardiac surgery per se. This certainly bears consideration if we appreciate that cardiology and cardiac surgery concentrates chronic vascular disease, hypertension, and diabetes into its practice and that cognitive deterioration in the elderly is often vascular in origin [34, 40]. It can be safely assumed that the cardiac surgical population carries a burden of occult chronic cerebral vascular disease equal to that documented in the Cardiovascular Heath Study [41] and the Rotterdam Scan Study [40]. Apart from shifting our thinking away from acute ischemic events as a cause of postoperative cognitive decline, a 32% incidence of acute cerebral infarction deserves comment. In all positive scans, the etiology of the events appeared embolic in origin. These findings are in congruence with other reports where diffusion MRI was done in much smaller numbers of patients [39, 42 44]. Taken together, a 30% to 50% incidence of acute focal ischemia indicates that further improvements in surgical practice, or perioperative care, are indicated. Our findings have obvious implications for the use of neurocognitive outcomes as an endpoint for investigations of cardiac surgical techniques or technologies to reduce brain injury. If the intent of those investigations is to reduce cerebral ischemic events, the analyses might be better conducted with DW-MRI because there does not appear to be a solid relationship between acute perioperative cerebral ischemia and cognitive outcomes. The frequency of events on DW-MRI is much greater than clinical stroke and it may be more appropriate for assessment of perioperative events than cognitive testing. It would be erroneous to conclude that cognitive outcomes are unimportant. Even after a successful surgery, these issues remain a concern of patients and families and are a burden to our health care system. As such, cognitive outcomes must be better understood and improved. Our findings only indicate that acute embolic cerebral infarction and its prevention is potentially a misdirection in solving postoperative cognitive decline. Improving cognitive outcomes is less likely to be addressed by technologies and techniques reducing cerebral infarction risks than by better risk stratification and a better understanding of the relationship between perioperative stressors, cognitive decline, and chronic cerebrovascular disease. References 1. Redmond JM, Greene PS, Goldsborough MA, et al. Neurologic injury in cardiac surgical patients with a history of stroke. Ann Thorac Surg 1996;61: Roach GW, Kanchuger M, Mangano CM, et al. Adverse cerebral outcomes after coronary bypass surgery. Multicenter Study of Perioperative Ischemia Research Group and the Ischemia Research and Education Foundation Investigators. N Engl J Med 1996;335: Shaw PJ, Bates D, Cartlidge NE, et al. Long-term intellectual dysfunction after coronary artery bypass graft surgery: a six month follow-up study. Q Rev Med 1987;62: McKhann GM, Goldsborough MA, Borowicz LM Jr, et al. Cognitive outcome after coronary artery bypass: a one-year prospective study. Ann Thorac Surg 1997;63: Newman MF, Kirchner JL, Phillips-Bute B, et al. Longitudinal assessment of neurocognitive function after coronaryartery bypass surgery. N Engl J Med 2001;344: Van Dijk D, Jansen EW, Hijman R, et al. Cognitive outcome after off-pump and on-pump coronary artery bypass graft surgery: a randomized trial. JAMA 2002;287: Schmitz C, Weinreich S, White J, et al. Can particulate extraction from the ascending aorta reduce neurologic injury in cardiac surgery? J Thorac Cardiovasc Surg 2003; 126: Cook DJ, Orszulak TA, Zehr KJ, et al. Effectiveness of the Cobra aortic catheter for dual-temperature management during adult cardiac surgery. J Thorac Cardiovasc Surg 2003;125: Borger MA, Ivanov J, Weisel RD, Rao V, Peniston CM. Stroke during coronary bypass surgery: principal role of cerebral macroemboli. Eur J Cardiothorac Surg 2001;19:

7 Ann Thorac Surg COOK ET AL 2007;83: COGNITIVE CHANGE, BRAIN ISCHEMIA, AND AGE Fearn SJ, Pole R, Wesnes K, Faragher EB, Hooper TL, McCollum CN. Cerebral injury during cardiopulmonary bypass: emboli impair memory. J Thorac Cardiovasc Surg 2001;121: Djaiani G, Fedorko L, Borger M, et al. Mild to moderate atheromatous disease of the thoracic aorta and new ischemic brain lesions after conventional coronary artery bypass graft surgery. Stroke 2004;35:e Gorman C. Hearts and minds; doctors knew about the mental fog that can set in after a bypass. But who knew it could come back? Time Magazine February 19, 2001:157, Grady D. Mental decline is linked to heart bypass surgery. New York Times February 8, van Everdingen KJ, van der Grond J, Kappelle LJ, Ramos LM, Mali WP. Diffusion-weighted magnetic resonance imaging in acute stroke. Stroke 1998;29: Singer MB, Chong J, Lu D, Schonewille WJ, Tuhrim S, Atlas SW. Diffusion-weighted MRI in acute subcortical infarction. Stroke 1998;29: Nakamura H, Yamada K, Kizu O, et al. Effect of thin-section diffusion-weighted MR imaging on stroke diagnosis. AJNR Am J Neuroradiol 2005;26: Mullins ME, Schaefer PW, Sorensen AG, et al. CT and conventional and diffusion-weighted MR imaging in acute stroke: study in 691 patients at presentation to the emergency department. Radiology 2002;224: Ozsunar Y, Sorensen AG. Diffusion- and perfusionweighted magnetic resonance imaging in human acute ischemic stroke: technical considerations. Top Magn Reson Imaging 2000;11: Murkin JM, Newman SP, Stump DA, Blumenthal JA. Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery. Ann Thorac Surg 1995;59: Svensson LG, Nadolny EM, Penney DL, et al. Prospective randomized neurocognitive and S-100 study of hypothermic circulatory arrest, retrograde brain perfusion, and antegrade brain perfusion for aortic arch operations. Ann Thorac Surg 2001;71: Kong RS, Butterworth J, Aveling W, et al. Clinical trial of the neuroprotectant clomethiazole in coronary artery bypass graft surgery: a randomized controlled trial. Anesthesiology 2002;97: Butterworth J, Legault C, Stump DA, et al. A randomized, blinded trial of the antioxidant pegorgotein: no reduction in neuropsychological deficits, inotropic drug support, or myocardial ischemia after coronary artery bypass surgery. J Cardiothorac Vasc Anesth 1999;13: Butterworth J, Wagenknecht LE, Legault C, et al. Attempted control of hyperglycemia during cardiopulmonary bypass fails to improve neurologic or neurobehavioral outcomes in patients without diabetes mellitus undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2005;130: Grigore AM, Mathew J, Grocott HP, et al. Prospective randomized trial of normothermic versus hypothermic cardiopulmonary bypass on cognitive function after coronary artery bypass graft surgery. Anesthesiology 2001;95: Al-Ruzzeh S, George S, Bustami M, et al. Effect of off-pump coronary artery bypass surgery on clinical, angiographic, neurocognitive, and quality of life outcomes: randomised controlled trial. BMJ 2006;332: Gold JP, Charlson ME, Williams-Russo P, et al. Improvement of outcomes after coronary artery bypass. A randomized trial comparing intraoperative high versus low mean arterial pressure. J Thorac Cardiovasc Surg 1995;110: Heyer EJ, Adams DC, Delphin E, et al. Cerebral dysfunction after coronary artery bypass grafting done with mild or moderate hypothermia. J Thorac Cardiovasc Surg 1997;114: Mora CT, Henson MB, Weintraub WS, et al. The effect of temperature management during cardiopulmonary bypass on neurologic and neuropsychologic outcomes in patients undergoing coronary revascularization. J Thorac Cardiovasc Surg 1996;112: Bashien G, Townes BD, Nessly ML, et al. A randomized study of carbon dioxide management during hypothermic cardiopulmonary bypass. Anesthesiology 1990;72: Lund C, Hol PK, Lundblad R, et al. Comparison of cerebral embolization during off-pump and on-pump coronary artery bypass surgery. Ann Thorac Surg 2003;76: Rasmussen LS, Johnson T, Kuipers HM, et al. Does anaesthesia cause postoperative cognitive dysfunction? A randomised study of regional versus general anaesthesia in 438 elderly patients. Acta Anaesthesiol Scand 2003;47: Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. (ISPOCD investigators. International Study of Post- Operative Cognitive Dysfunction). Lancet 1998;351: Wahrborg P, Booth JE, Clayton T, et al. Neuropsychological outcome after percutaneous coronary intervention or coronary artery bypass grafting: results from the Stent or Surgery (SoS) Trial. Circulation 2004;110: van Dijk EJ, Breteler MM, Schmidt R, et al. The association between blood pressure, hypertension, and cerebral white matter lesions: cardiovascular determinants of dementia study. Hypertension 2004;44: Rankin KP, Kochamba GS, Boone KB, Petitti DB, Buckwalter JG. Presurgical cognitive deficits in patients receiving coronary artery bypass graft surgery. J Int Neuropsychol Soc 2003;9: Mullges W, Berg D, Schmidtke A, Weinacker B, Toyka KV. Early natural course of transient encephalopathy after coronary artery bypass grafting. Crit Care Med 2000;28: Vingerhoets G, Van Nooten G, Vermassen F, De Soete G, Jannes C. Short-term and long-term neuropsychological consequences of cardiac surgery with extracorporeal circulation. Eur J Cardiothorac Surg 1997;11: Selnes OA, Grega MA, Borowicz LM Jr, Royall RM, McKhann GM, Baumgartner WA. Cognitive changes with coronary artery disease: a prospective study of coronary artery bypass graft patients and nonsurgical controls. Ann Thorac Surg 2003;75: Knipp SC, Matatko N, Wilhelm H, et al. Evaluation of brain injury after coronary artery bypass grafting. A prospective study using neuropsychological assessment and diffusionweighted magnetic resonance imaging. Eur J Cardiothorac Surg 2004;25: Vermeer SE, Prins ND, den Heijer T, Hofman A, Koudstaal PJ, Breteler MM. Silent brain infarcts and the risk of dementia and cognitive decline. N Engl J Med 2003;348: Price TR, Manolio TA, Kronmal RA, et al. Silent brain infarction on magnetic resonance imaging and neurological abnormalities in community-dwelling older adults. The Cardiovascular Health Study. CHS Collaborative Research Group. Stroke 1997;28: Restrepo L, Wityk RJ, Grega MA, et al. Diffusion- and perfusion-weighted magnetic resonance imaging of the brain before and after coronary artery bypass grafting surgery. Stroke 2002;33: Stolz E, Gerriets T, Kluge A, Klovekorn WP, Kaps M, Bachmann G. Diffusion-weighted magnetic resonance imaging and neurobiochemical markers after aortic valve replacement: implications for future neuroprotective trials? Stroke 2004;35: Wityk RJ, Goldsborough MA, Hillis A, et al. Diffusion- and perfusion-weighted brain magnetic resonance imaging in patients with neurologic complications after cardiac surgery. Arch Neurol 2001;58:571 6.