302

Page 1

Cerebral Protection During Cardiac Surgery

Charles W. Hogue, M.D. Baltimore, Maryland

Despite many advances in perioperative care, cerebral injury from cardiac surgery remains an important source of
patient morbidity and mortality.1-3 This injury has a range of clinical manifestations, including stroke and
encephalopathy, which occur in 1–3% and 10–15% of patients, respectively.1,3,4 Postoperative cognitive dysfunction
(POCD) is more common, affecting 15–40% of patients 1 month after surgery.5,6 This form of cerebral injury
involves alterations in attention, concentration, and executive cognitive tasks that are detected only with detailed
psychometric testing. Debate is on-going regarding whether POCD increases the risk for long-term cognitive decline
and dementia, particularly in the elderly. Newman et al6 reported an association between POCD after CABG surgery
and cognitive impairment 5 years later. In a longitudinal study, Selnes et al7 compared cognitive function in patients
recovering from CABG surgery with that of a control group receiving medical management for coronary artery
disease. Surgery and control groups showed decrements in cognition over time, but no differences in testing results
were observed between the CABG surgery and the control groups 6 years after the initial evaluation. Van Dijk et al8
compared cognitive outcomes in patients 5 years after CABG surgery (performed with or without cardiopulmonary
bypass, CPB) with a group of subjects without cardiac disease. After adjusting for differences in age, education, and
co-morbidities, there was no difference in cognitive performance between CABG surgery patients and controls.
Evered et al9 performed neuropsychological evaluations in patients undergoing total hip replacement surgery
(THJR), patients undergoing CABG surgery, patients undergoing cardiac catheterization (CA), and non-procedural
controls. Although the incidence of POCD at day 7 was higher in CABG surgery patients than in other patients, at 3
months the incidence of POCD was similar between groups: 21% for CA under sedation, 16% for THJR surgery,
and 16% for CABG surgery (p=0.13). Accumulating data suggest that the impact of cardiac surgery on cognitive
function is short lived, with recovery by 3 to 6 months after surgery in most patients. Further cognitive decline more
likely results from the progression of coexisting cerebrovascular disease.

Etiology
Cerebral injury from cardiac surgery is believed to result primarily from cerebral embolism and cerebral
hypoperfusion.5 It is hypothesized that the clinical manifestations depend on the size and location of the injury (e.g.,
motor areas vs. areas involved with cognition). Embolism and hypoperfusion can contribute to injury in the same or
different areas in a given patient.10 Although, primary cerebral hemorrhage is believed to be a rare cause of cerebral
injury, sensitive magnetic resonance imaging (MRI) methods suggest that punctuate hemorrhages can occur after
surgery in some patients, such as those with endocarditis.11 Inflammatory processes resulting from CPB and organ
ischemia/reperfusion and genetic susceptibility may further contribute to injury.3,12-14

Insight into the mechanism(s) of perioperative cerebral injury have resulted from studies involving brain MRI or CT
imaging. These studies suggest that 30–50% of strokes after cardiac surgery result from cerebral macroembolism,
mostly arising from atherosclerosis of the ascending aorta.3 Cerebral microemboli (<200 µm) have multiple sources
and can be gaseous or composed of particulate material. Atherosclerotic burden of the aorta has been found to be
associated with the number of transcranial Doppler (TCD)-detected cerebral microembolic signals during CABG
surgery.5,15 Lipogenous emboli lodged in small arterioles of the brain are found at autopsy in humans after cardiac
surgery.16 Canine experiments suggest that fat arising from the cardiotomy suction aspirate is the major source of
these lipid emboli.17 Current CPB filters are inefficient at removing fatty material. Some small studies have
implicated cerebral microemboli as a cause of POCD by correlating the number of TCD-detected cerebral embolic
signals with cognitive decline after CABG surgery.18,19 These results were not confirmed in subsequent studies of
patients undergoing open-chamber valvular surgery, where more emboli are present.20,21 It is likely that the
composition and not the number of microemboli is more important for the manifestations of cerebral injury.

Clinical evidence suggests that cerebral injury from hypoperfusion may be more common in contemporary practice
due a to the growing number of patients with preexisting, and typically clinically asymptomatic, cerebrovascular

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

 302
Page 2

disease.5 This concern is reinforced by data showing that as many as 27–43% of patients experience regional or
global cerebral O2 desaturations during CPB, indicating cerebral O2 demand/delivery mismatch.22 The importance of
cerebral hypoperfusion as a cause of perioperative stroke is underscored by a study by Gottesman et al,23 who found
that 68% of 98 strokes after cardiac surgery were hypoperfusion-type watershed strokes based on diffusion-weighted
MRI. Cerebral hypoperfusion might be of particular concern during “off pump” surgery when displacement of the
heart results in systemic hypotension and cerebral venous hypertension. One study showed that 15% of patients
undergoing off-pump CABG surgery exhibited EEG evidence of cerebral hypoxemia and near infrared spectroscopy
(NIRS) desaturations.24

Approaches to Cerebral Protection

Cerebral ischemic injury can be global or focal, as well as temporary (e.g., period of circulatory arrest) or permanent
(e.g., distal to embolus). Brain energy stores are rapidly depleted during ischemia, leading to depolarization and
release of excitatory neurotransmitters. Cerebral ischemia triggers activation of multiple pathways (“ischemic
cascade”) over a period of hours to days that determine the ultimate extent of injury and functional outcome.25 Other
injury pathways activated lead to altered calcium homeostasis, free radical production, activation of proteases, and
initiation of apoptosis. Vulnerability to injury varies between cell subtypes, with particular susceptibility occurring
in neurons located in the hippocampus, cortex, cerebellum, corpus striatum, and thalamus.26 The central area of
ischemic injury is surrounded by viable tissue that is vulnerable to infarction (“the ischemic penumbra”) from the
secondary events that follow the initial injury. Cerebral protection during cardiac surgery, thus, can be broadly
categorized into several basic strategies: 1) prevent injury from cerebral emboli; 2) ensure cerebral O 2
delivery/demand balance; and 3) prevent secondary brain injury to the ischemic penumbra.

Reducing Cerebral Emboli

Reducing cerebral embolism is the basis of various strategies for cerebral protection during cardiac surgery. Due to
CPB’s purported role in increasing cerebral embolism (and increasing potentially injurious inflammatory processes),
avoiding CPB by performing CABG surgery “off-pump” has been advocated as a means for reducing cerebral injury
from surgery.27 Most data supporting this approach have been anecdotal and retrospective. Several prospectively
randomized trials of off- versus on-pump CABG surgery have failed to demonstrate that avoiding CPB provides a
clear reduction in stroke rate.28 In a prospectively randomized study of low-risk patients undergoing CABG surgery,
there were no differences in the frequency of cognitive dysfunction 3 months or 5 years after surgery for patients
randomized to on- versus off-pump surgery.29 In the VA Randomized On/Off Bypass Study, there was no difference
between surgery groups in the rate of the composite outcome of death or complications (reoperation, new
mechanical support, cardiac arrest, coma, stroke, or renal failure) within 30 days of surgery.30 The rate of graft
patency at 1 year based on angiography was lower in the off-pump group (82.6% vs. 87.8%, p<0.01). In a review of
6665 patients followed for an average of 4.5 years, off-pump CABG was associated with an increased risk of repeat
revascularization and major vascular events (cardiac death, repeat revascularization, MI, or stroke) compared with
on-pump surgery.31 It has been argued that high-risk patients are a group that may benefit the most from CPB
avoidance. More recently, however, in a randomized study of 341 high-risk CABG surgery patients (Euroscore ≥ 5),
avoiding CPB did not affect the frequency of stroke (4.0% vs. 3.7%; RR, 1.08; 95% CI, 0.37-3.14; p=1.0).28 The
decision to carry out CABG surgery off-pump must be individualized, but the current data do not indicate that this
surgical approach substantially improves neurological outcomes.

Atherosclerosis of the ascending aorta is a known source of cerebral emboli that may cause stroke and possibly
POCD.5 In patients presenting with stroke, aortic atherosclerosis is associated with blood hypercoagulability, which
increases the risk for recurrent stroke and death.32 Two fundamental clinical issues related to atherosclerosis of the
aorta are its detection and its surgical management. Epiaortic ultrasound is more sensitive than direct palpation or
TEE for detecting atherosclerosis of the ascending aorta.33,34 This method allows the surgeon to identify and avoid
atheroma during aortic cannulation and cross-clamping. Approaches to surgical management of patients found to
have an atherosclerotic aorta include: a) converting to off-pump surgery; b) using alternative sites for CPB
cannulation; c) using the single cross-clamp technique to avoid partial aortic occlusion clamping for proximal

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

 302
Page 3

bypass graft anastamosis; d) using fibrillatory arrest rather than cardioplegic arrest to avoid cross-clamping; e)
avoiding proximal anastomosis by using arterial grafts; and f) replacing the ascending aorta under circulatory arrest.5
Improved neurological outcomes have been reported with epiaortic ultrasound-guided surgery and with practices
that minimize the risk for embolization.35-37 A meta-analysis of 7 studies found that the rate of stroke was 0.31% for
patients undergoing off-pump CABG with minimal aortic manipulations versus 1.35% for patients undergoing off-
pump CABG with a side-clamp or proximal graft anastomosis device (p=0.003).38

Other interventions intended to reduce cerebral emboli during CPB include the use of in-line leukocyte-depleting
filters and modified aortic CPB cannulae, but data to support the benefit of either for reducing cerebral injury are
lacking.5 Mediastinal insufflation of CO2 (to increase the rate of absorption of intravascular emboli by substituting
the more soluble CO2 for air in the surgical wound) may reduce the number of arterial emboli.5 In a study of 80
patients undergoing open chamber cardiac surgery, flooding of the surgical field with CO2 resulted in preserved
P300 auditory evoked potentials (believed to be a sensitive marker for subtle cognitive changes) compared with
patients not receiving this intervention.39 However, neurocognitive outcomes did not differ between the CO2
insufflation and control groups. Cardiotomy suction aspirate is high in lipid content and is a source of cerebral lipid
microemboli in experimental canine CPB.5,17 Based on animal data and inferential data from humans, many centers
adopted a practice of either discarding or processing cardiotomy suction aspirate before returning the blood to the
CPB circuit. As platelets and coagulation factors are lost with either approach, this practice might increase the
frequency of blood transfusions when the suction volume is high. This possibility is an important consideration, as
transfusion of allogeneic packed red blood cells and platelets is associated with increased risk for stroke after cardiac
surgery.40 Rubens et al41 reported no difference in the frequency of POCD 1 week or 3 months after CABG and/or
aortic valve surgery between patients whose cardiotomy blood was processed with a cell saver compared with those
who had direct return to the CPB circuit. Djaiani et al,42 however, found that processing cardiotomy blood with a
continuous-flow cell saver led to a significantly lower rate of POCD 6 weeks after CABG surgery. These benefits
were not sustained after 1 year of follow-up.43 Neither study showed a difference in the number of TCD embolic
signals between cell-saver and control groups. In both studies, patients in the cell-saver group had higher rates of
transfusion than controls. Of note, the median volume of aspirated cardiotomy blood in the Djaiani et al43 study (800
ml; range, 175 to 3840 ml) was higher than that in the study by Rubens et al41 (636 ml; 95% CI, 566 to 706 ml).
Thus, the data are conflicting on whether the routine processing of cardiotomy suction aspirate with a cell saver
improves neurological outcomes. Any benefits might depend on the volume of aspirated blood and/or the type of
device used. Therefore, when large blood loss is not expected, avoiding the return of unprocessed cardiotomy
suction aspirate to the CPB circuit is reasonable.

Improving Cerebral Oxygen Balance

Blood Pressure Management: Mean arterial pressure (MAP) is often kept at low levels (50 to 60 mmHg) during
CPB based on data showing that cerebral blood flow (CBF) autoregulation is intact when α-stat pH management is
used.44 Whether this practice is appropriate for the rising proportion of patients with cerebrovascular disease is not
clear. Gottesman et al23, for example, found a relationship between a decrease in MAP by ≥10 mmHg from baseline
during CPB and a risk for watershed stroke detected by MRI. More recent data suggest that maintaining MAP
between 80 and 90 mmHg during CPB is associated with less delirium and less early cognitive dysfunction.45 Our
group has reported in animal and clinical studies that real-time continuous monitoring of CBF autoregulation using
transcranial Doppler or NIRS may provide a novel approach for individualizing MAP targets during CPB.46,47 That
is, this method allows for MAP to be optimized within an individual’s autoregulatory range, thus reducing the
potential of cerebral hypoperfusion. In a cohort of 232 patients, we found that the average MAP at the lower limit of
CBF autoregulation was 66 mmHg (95% CI, 65-58 mmHg). The range of pressures, though, at the autoregulation
threshold was 40 to 90 mmHg. Women tended to have a lower autoregulatory threshold than men, but predicting the
appropriate MAP for CPB was not accurate based on demographic data, including preoperative MAP. These data
suggest that monitoring CBF autoregulation in real-time might be necessary to identify the optimal MAP for CPB.

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

 302
Page 4

Red Cell Transfusion: Preoperative and intraoperative anemia has been linked in retrospective analysis with adverse
outcomes, including stroke, particularly when Hct is <21% during CPB.48,49 In a prospectively randomized study,
hemodilution to a Hct of 15–18% during CPB was found to be associated with increased neurocognitive impairment
in elderly patients compared with a Hct ≥27%.50 Whether treatment of anemia with transfusion of packed red blood
cells (PRBC) improves patient outcome was evaluated in a prospectively randomized clinical trial of 502 patients
undergoing cardiac surgery with CPB, the TRACS study.51 Patients were randomly assigned to a liberal or
restrictive transfusion group (Hct ≥30% or ≥ 24%, respectively). The hemoglobin level in the liberal transfusion
group (10.5 g/dL; 95% CI, 10.4-10.6) was higher than that in the restrictive group (9.1 g/dL; 95% CI, 9.0-9.2;
p<0.001). The frequency of the primary composite end-point of 30-day all-cause mortality or severe morbidity
(cardiogenic shock, ARDS, or acute renal injury requiring dialysis or hemofiltration) was similar between groups
[10% liberal vs. 11% restrictive; between-group difference, 1% (95% CI,−6% to 4%); p=0.85]. Non-leukocyte-
depleted blood was used in this study, but the PRBC units were stored for a median of 3 days before transfusion.
These variables have been associated with reduced risk from PRBC transfusion.52,53 Further, cell salvage was not
used in the study patients. Independent of transfusion strategy, the number of transfused PRBC units was an
independent risk factor for clinical complications or death at 30 days [hazard ratio for each additional unit
transfused, 1.2 (95% CI, 1.1-1.4); p=0.002]. Transfusion of ≥ 5 units of PRBC was associated with mortality. The
authors concluded that for patients undergoing cardiac surgery, outcomes do not differ with use of a liberal
transfusion strategy or a restrictive strategy. The Society of Thoracic Surgeons and Society of Cardiovascular
Anesthesiologists clinical practice guidelines opine that “For patients on CPB with risk for critical end-organ
ischemia/injury, it is not unreasonable to keep hemoglobin level at 7 g/dL or more.”54

Temperature Management: Hypothermia is the most widespread cerebral protection strategy during cardiac surgery,
a practice supported by a plethora of laboratory data.5 Clinically, hypothermia has been demonstrated to improve
neurologic recovery in comatose survivors of cardiac arrest.26 Despite its widespread use, there is little clinical
evidence that hypothermia is neuroprotective during cardiac surgery.55 The ineffectiveness of hypothermia might be
explained by the absence of hypothermia during all periods of cerebral risk or inadvertent hyperthermia with re-
warming.5 The latter might result from the proximity of the aortic cannula (returning warm blood) to cerebral
vessels and/or inaccurate patient temperature monitoring that underestimates brain temperature. Rapid re-warming
and postoperative hyperthermia are associated with risk for neurological complications.56,57 Re-warming to 34oC
rather than to 37oC resulted in improved neurocognitive outcome after CABG surgery.58 However, at 3 months after
CABG surgery, the rates of POCD were no different whether temperature was maintained at 37oC (with a
circulating warming blanket) or at 34oC during CPB without rewarming.59 Our group has demonstrated that CPB re-
warming is associated with impaired CBF autoregulation in a high proportion of patients.60 These results indicate
that CBF is pressure-passive in some patients during re-warming at a time when systemic vascular resistance is
typically low and cerebral metabolic rate increased. These conditions might lead to a CBF that is inadequate for
metabolic demand. In that study, there was an association between impaired CBF autoregulation and stroke.

Neuromonitoring: NIRS provides an estimate of the adequacy of regional cerebral O2 supply. Cerebral O2
desaturations detected with NIRS are associated with postoperative cognitive decline and stroke.61,62 In a
randomized study of 200 CABG surgical patients, NIRS-guided interventions were associated with less major organ
injury (death, MI, stroke) and shorter ICU length of stay versus standard care.63 Algorithms have been proposed for
treating cerebral O2 desaturation.64 Interventions include ensuring adequate CPB flow or cardiac output when not on
CPB, increasing MAP, avoiding hypocarbia (decreasing gas-inflow), deepening anesthesia, increasing the FiO2,
instituting pulsatile CPB flow, considering transfusion if indicated, administering anti-convulsant drugs when
indicated, and considering hypothermia.

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

 302
Page 5

Protecting the Ischemic Penumbra

The ischemic penumbra is vulnerable to infarction because multiple injury pathways that comprise the ischemic
cascade referenced earlier are activated. Microcirculatory changes resulting from brain injury, particularly
inflammatory and hemostatic activation, can contribute to cerebral hypoperfusion due to the “no reflow”
phenomenon. Thus, measures aimed at neuroprotective hemodynamic management may be critical to protection
from secondary brain injury. Other measures for protecting the ischemic penumbra are mostly pharmacologic,
including avoidance of hyperglycemia.

Blood Glucose Control: Hyperglycemia worsens experimental cerebral injury and is associated with poor
neurological outcome after stroke in humans.65 Hyperglycemia worsens ischemic damage by multiple mechanisms.
The relationship between serum glucose and stroke severity appears to be “U” shaped. Experimentally, the nadir
glucose level associated with best neurological outcome appears to be in the 120 mg/dL range.66 Initial studies
showing that intensive insulin treatment improves outcomes of critically ill patients were rapidly extrapolated to
patients undergoing cardiac surgery.67 More recent data (the NICE-SUGAR study) demonstrated in a multicenter,
randomized trial that a glucose target of 81 to 108 mg/dL was associated with higher mortality compared with a
target of ≤180 mg/dL in adult critically ill patients. Importantly, these prior trials were conducted in the ICU.68 Data
in cardiac surgery patients have not supported intensive insulin therapy for improving patient neurological
outcomes. In a randomized trial of 400 patients undergoing cardiac surgery, the frequency of the composite outcome
of death, cardiac morbidity, stroke, or renal failure was higher for patients given an insulin infusion to maintain
intraoperative glucose between 80 and 100 mg/dL than for patients who received conventional treatment [insulin
given when glucose was >200 mg/dL (p=0.02)].69 A prospectively randomized trial found no difference in the
frequency of neurological complications 6 weeks or 6 months after CABG surgery between patients who received
intraoperative insulin when glucose was >100 mg/dL and those who received insulin when glucose was >200
mg/dL.70 In nonsurgical patients with stroke, maintenance of euglycemia with a glucose-insulin-potassium infusion
conferred no benefits on 90-day mortality or functional outcome compared with controls.71 Administration of insulin
when glucose is >140 mg/dL is recommended for nonsurgical patients with stroke.72

Drugs For Cerebral Protection: A great deal of investigative energy has been invested testing multiple drugs for
neuroprotection during cardiac surgery. For the most part, clinical trial results have been disappointing. Agents that
showed promising effects in vitro via multiple mechanisms but failed to provide clinical protection against ischemic
cerebral injury include barbiturates, propofol, Ca2+ channel blockers, NMDA receptor antagonists, anti-
inflammatory agents, anti-oxidants, GABA receptor blockers, 17β-estradiol, piracetam, and others.5,73-76 Volatile
anesthetics and xenon have laboratory-proven cerebral protective effects. A systematic literature review,
nonetheless, concluded that evidence was insufficient to choose one anesthetic over another for the purposes of
cerebral protection.77 Magnesium was investigated as a cerebral protectant in a randomized, blinded, placebo-
controlled trial in 350 patients undergoing cardiac surgery.78 At a dose that increased serum levels to 1½ to 2 times
normal, Mg2+ use was associated with improved cognitive performance 24 to 96 hours after surgery compared with
placebo, but these benefits were not present 3 months after cardiac surgery. Experimentally, ketamine protects
against ischemic neuronal injury by blocking NMDA receptors and by attenuating inflammation.79 In a small
randomized pilot study (n=52), the frequency of POCD 1 week after cardiac surgery was lower in patients who
received ketamine than in those who received placebo (33% vs. 81%, p<0.001).75 Likewise, the frequency of
delirium was lower for patients given ketamine during surgery than for those given placebo.80 Other promising early
data have been reported with erythropoietin (EPO). Thirty-two patients undergoing CABG surgery were randomized
to receive one of two doses of EPO or placebo starting the day before surgery.81 Two months after surgery, there
was a trend toward a lower frequency of POCD in patients who received EPO (p=0.085). In contrast, EPO failed to
provide cerebral protective benefits and was associated with higher mortality and thromboembolic events compared
with placebo in a large trial of patients with acute ischemic stroke.82 Early ongoing investigations of novel
compounds that act on molecular, genomic, and proteomic targets have had exciting results.83 In a placebo-
controlled, double-blinded trial of 242 patients undergoing cardiac surgery with CPB, investigators found no
difference in the frequency of POCD between patients prospectively randomized to receive lidocaine and those who

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

 302
Page 6

received placebo (45.5% vs. 45.7%, respectively, p=0.97). 74 Diabetic patients who received lidocaine had a higher
frequency of POCD than those who received placebo (p=0.004). Hyperbaric O2 was evaluated as a potential
neuroprotective strategy in a study of 64 patients undergoing surgery with CPB.62 This therapy reduces
inflammation and ischemic infarct volume in animals. Patients who received hyperbaric O2 in three sessions 24, 12,
and 4 hours before CPB had lower levels of inflammatory marker and a lower frequency of POCD 4 months after
surgery compared with controls who received pressurized atmospheric air. Another exciting approach for cerebral
protection in the preliminary phase of clinical investigations is remote limb transient ischemia. In a study of 33
patients with subarachnoid hemorrhage, limb preconditioning consisting of three 5-minute inflations of a blood
pressure cuff every 24 to 48 hours for 14 days was found to be safe and well tolerated.84

Table 1. Evidence-based strategies for cerebral protection during cardiac surgery.

Strategies Supported by Clinical Investigations

Epiaortic ultrasound for detection of atherosclerosis of the ascending aorta
Avoidance of hyperthermia during CPB re-warming
Strategies With Reasonable Level of Clinical Evidence
The use of a membrane oxygenator and an arterial line filter (≤40 µm) during CPB
α-stat pH management during CPB
Single cross-clamp technique for proximal CABG anastamosis patients at risk for atheroembolism
NIRS monitoring especially in high-risk patients
Arterial blood pressure kept >70 mmHg during CPB in high-risk patients
Strategies That Are Acceptable and Considered Reasonable Treatment by Most Experts
Serum glucose should be kept <140 mg/dL with an infusion of insulin
Consider processing cardiotomy suction aspirate with a cell-saver device
Transfusion of packed red blood cells should be considered in high-risk patients when hemoglobin is ≤7 g/dL or
higher depending on other patient-specific considerations.

The existing data on pharmacologic neuroprotection during cardiac surgery must be considered in the context of
several limitations, including the investigation of drugs with marginally positive preclinical data, the reliance on
laboratory models that might not replicate clinical injury, often small study size, and limiting neurological
assessments to only one aspect of cerebral injury (e.g., POCD). It is unclear what metrics would be required from
the FDA for approval of a drug for the indication of cerebral protection during cardiac surgery (i.e., stroke vs.
POCD). Certainly confirmation of a drug’s safety and efficacy in separate, adequately powered, multicenter,
prospectively randomized, placebo controlled, double-blinded studies would be a prerequisite.
Conclusions
A comprehensive approach to cerebral protection that includes interventions to reduce cerebral embolism and ensure
cerebral O2 supply/demand balance may result in improved neurological outcomes.5 Strategies for improving
neurological outcomes from cardiac surgery are summarized in the Table.
 
References 11. Circulation 120:585-91, 2009.
1. N Engl J Med 335:1857-63, 1996. 12. J A m Coll Cardiol 49 1934-42, 2007.
2. Stroke 41:499-503, 2010. 13. Stroke 36 1854-8, 2005.
3. A nesth A nalg 107:21-8, 2008. 14. J Thorac Cardiovasc Surg 134:996-1005, 2007.
4. Circulation 103 2133-7, 2001. 15. Br J A naesth 91:656-61., 2003.
5. A nesth A nalg 103 21-37, 2006. 16. A nn Thorac Surg 59 1304-7, 1995.
6. N Engl J Med 344:395-402, 2001. 17. A nn Thorac Surg 65 1651-5, 1998.
7. A nn Neurol 63:581-90, 2008. 18. Stroke 25:1393-9, 1994.
8. A nn Thorac Surg 85:60-4, 2008. 19. J Cardiothorac V asc A nesth 10 113-8, 1996.
9. A nesth A nalg 112:1179-85, 2011. 20. J Neurol Neurosurg Psychiatry 65 573–6, 1998.
10. A rch Neurol 55:1475-82., 1998. 21. J Thorac Cardiovasc Surg 134:1143-9, 2007.
Copyright © 2011 American Society of Anesthesiologists. All rights reserved.
 302
Page 7

22. A nn Thorac Surg 58 1702-8, 1994. 71. Lancet Neurology 6:397-406, 2007.
23. Stroke 37:2306-11, 2006. 72. Stroke 38:1655-711, 2007.
24. Heart Surg Forum 3:198-202., 2000. 73. Stroke 38 2048-54, 2007.
25. A nn Emerg Med 36:483-506, 2000. 74. Stroke 40:880-7, 2009.
26. Circulation 118:2452-83, 2008. 75. Med Sci Monit 14:PI53-7, 2008.
27. J Thorac Cardiovasc Surg 125:797-808, 2003. 76. A nn Thorac Surg 87:820-5, 2009.
28. Eur Heart J 29:2601-16, 2008. 77. CNS Drugs 24:893-907, 2010.
29. J A m Med A ssoc 287:1405-12, 2002. 78. J Thorac Cardiovasc Surg 131:853-61, 2006.
30. N Engl J Med 361:1827-37, 2009. 79. J Cardiothorac V asc A nesth 24:131-42, 2010.
31. Circulation 121:1800-8, 2010. 80. A cta A naesthesiol Scand 53:864-72, 2009.
32. J A m Coll Cardiol 52:855-61, 2008. 81. Stroke 40:2769-75, 2009.
33. J Thorac Cardiovasc Surg 103 453-62, 1992. 82. Stroke 40:647-56, 2009.
34. Circulation 100:642-7, 1999. 83. Stroke 40:S111-4, 2009.
35. A nn Thorac Surg 75:1387-91 2003. 84. Stroke 42:1387-91, 2011.
36. A nn Thorac Surg 78:1579-85, 2004.
37. J Thorac Cardiovasc Surg 131:114-21 2006.
38. Heart Lung Circ 20:318-24, 2011.
39. A nn Thorac Surg 85:543-7, 2008.
40. Transfusion 44:1143-8, 2004.
41. Circulation 116:I-89-I-97, 2007.
42. Circulation 116 1888-95, 2007.
43. A SA A bstract, 2009.
44. A nesth A nalg 66 825-32, 1987.
45. Eur J Cardiothorac Surg, in press.
46. Stroke 38:2818-25, 2007.
47. Stroke 41:1951-6, 2010.
48. A nn Thorac Surg 80 1381-7, 2005.
49. Circulation 116:471-9, 2007.
50. A nesthesiology 107:577-84, 2007.
51. J A m Med A ssoc 304:1559-67, 2010.
52. Circulation 109:2755-60, 2004.
53. N Engl J Med 358:1229-39, 2008.
54. www.sts.org.
55. Cochrane Database Syst Rev:CD002138, 2001.
56. A nesth A nalg 94 4-10, 2002.
57. Stroke 33 537-41, 2002.
58. Circulation 104 I-85-I-91, 2001.
59. J Thorac Cardiovasc Surg 134:1443-52, 2007.
60. A nesth A nalg 110:321-8, 2010.
61. J Cardiothorac V asc A nesth 18:552-8, 2004.
62. J Thorac Cardiovasc Surg 130:1623-30, 2005.
63. A nesth A nalg 104:51-8, 2007.
64. Semin Cardiothorac V asc A nesth 4:274-81,
2007.
65. Neurology 59:669-74, 2002.
66. A rch Neurol 67:1123-30, 2010.
67. N Engl J Med 354:449-61., 2006.
68. N Engl J Med 360:1283-97, 2009.
69. A nn Intern Med 146 233-43, 2007.
70. J Thorac Cardiovasc Surg 130 1319, 2005.

Copyright © 2011 American Society of Anesthesiologists. All rights reserved.

Disclosure

Covidien, Astra-Zeneca, Funded Research