Multiple advances in perioperative techniques, especially the maturation of cardiopulmonary bypass, have allowed the practice of cardiac surgery to develop rapidly. Despite these advances that have enabled contemporary cardiac surgery to address a diverse range of cardiac pathologies, neurological injury, ranging from delirium to stroke, has persisted as a relatively common and important complication after cardiac surgery. As a result, extensive research has explored the etiologies of neurologic injury after cardiac surgery to identify interventions that may reduce its incidence. Although the strategies for neuroprotection have some similarities across the types of cardiac surgery, many differences remain as a result of varied etiologies for neurological injury. For the sake of clarity, this chapter will approach neurologic injury and neuroprotective strategies by the type of cardiac procedure in the following four sections: cardiac surgery with cardiopulmonary bypass; off-pump coronary artery bypass grafting; transcatheter aortic valve replacement; and thoracic aortic surgery.
Neurological Injury After Cardiac Surgery with Cardiopulmonary Bypass
Incidence and Significance
Multiple clinical trials have demonstrated that the incidence of new infarcts on diffusion-weighted magnetic resonance imaging following cardiac surgery with cardiopulmonary bypass (CPB) ranges from 25% to 40%. Although the vast majority of these lesions are not clinically significant, the incidence of clinical stroke after cardiac surgery with CPB has persisted in the 5%–10% range, depending on the diagnostic approach. Stroke in this setting remains an important complication because it significantly increases perioperative mortality and morbidity. Postoperative cognitive dysfunction including delirium is even more common than stroke after cardiac surgery with CPB. This neurological complication has also been associated with significant reductions in independence and quality of life.
Multiple clinical trials have reported patient and procedure-specific risk factors for the development of neurologic injury after cardiac surgery with CPB. These risk factors include advanced age, diabetes, preexisting cerebrovascular disease, severe atherosclerotic vascular disease, previous cardiac surgery, prolonged CPB time, and perioperative hypotension. The stroke risk is also typically significantly increased in the setting of combined cardiac procedures on CPB as follows: coronary artery bypass grafting (CABG) < valve repair/replacement < CABG and valve repair/replacement < multiple valve interventions. The risk factors for neurologic injury in this setting are listed in Table 22.1 .
Cerebral embolization is a common etiology for neurological injury after cardiac surgery with CPB. In a clinical evaluation of an intra-aortic filter device deployed during cardiac surgery with CPB ( N = 2297), detailed histopathological analysis confirmed filter-trapped debris in 98% of study patients. The debris consisted of fibrous atheroma in 79%, platelets and fibrin in 44%, red blood cell thrombus in 8%, adventitial tissue in 3%, and miscellaneous debris in 2% of cases. Patients with neurological injury after cardiac surgery with CPB will often have evidence of multiple cerebral lesions, consistent with cerebral embolization.
Embolic material can be further classified into three categories: macroemboli, microemboli, and gaseous emboli ( Table 22.2 ). Macroemboli are frequently derived from atheromatous plaque, valvular calcium, or thrombus that become dislodged during surgical manipulation to result in a focal stroke when they are clinically significant. Microemboli are typically derived from small lipid particles and platelet-fibrin aggregates and may be responsible in part for delirium after cardiac surgery with CPB. Gaseous emboli can enter the CPB circuit during vascular cannulation and anastomosis, vacuum-assisted venous drainage, as well as during open cardiac chamber procedures. The amount of embolized gaseous microemboli very probably correlates with the degree of subsequent neurological injury, including both delirium and stroke. Larger gaseous emboli may result in discrete stroke.
Hypotension, Hypoperfusion, and Cerebral Autoregulation
Periods of hypotension are relatively common in cardiac surgery compared with other surgical procedures. Most patients with intact cerebral autoregulation can tolerate moderate swings in mean arterial pressure and still maintain adequate cerebral blood flow. Despite this compensation, however, cerebral autoregulation may be impaired in the setting of multiple factors including advanced age, cerebrovascular disease, cardiopulmonary bypass, sepsis, and temperature. As a consequence, patients with impaired cerebral autoregulation may be at greater risk of cerebral hypoperfusion and ischemia during periods of hypotension. Hypoperfusion injuries are relatively common in patients who develop strokes after cardiac surgery with CPB, with a reported incidence of 20%–40% in computed tomographic imaging and 40%–60% in magnetic resonance imaging. Recent trials have demonstrated that the duration and degree of hypotension during cardiac surgery with CPB significantly predicted the risks of stroke ( P < 0.05) as well as major morbidity and mortality ( P < 0.05). A recent trial ( N = 197) randomized adult cardiac surgical patients to one of two mean arterial pressure groups during CPB: higher blood pressure defined as a mean of 70–80 mmHg and lower blood pressure defined as a mean of 40–50 mmHg. The defined primary outcome was the total volume of new cerebral lesions detected on diffusion-weighted magnetic resonance imaging in the first postoperative week. The primary outcome was similar between groups (median difference estimate 0; 95% confidence interval -25 to 0.028; P = 0.99). There were new cerebral lesions evident in over 50% of patients in each group, a finding that is consistent with the existing literature. Taken together, these clinical trials suggest that, in addition to optimal blood pressure management and in consideration of the individual cerebral autoregulation reserve, further multimodal trials are indicated to identify best practice measures for neuroprotection in cardiac surgical procedures with CPB.
The Inflammatory Response
The conduct of CPB triggers a profound systemic inflammatory response because blood interacts with the hardware components of the CPB circuit. This inflammatory response includes activation of the coagulation system, fibrinolytic system, complement system, leukocytes, endothelial cells, and platelets. This acute phase reaction may lead to tissue injury, depending on factors such age, genotype, and functional reserve. This systemic inflammatory response may also have neurological effects by causing glial cell injury, cerebral edema, and altering the integrity of the blood–brain barrier. Clinical trials have suggested a role for the systemic inflammatory response in neurologic injury after cardiac surgery with CPB. Future trials will probably investigate the modulation of this systemic inflammatory response in a multimodal approach to neuroprotection in this clinical setting.
Neuronal necrosis can result from cellular injury, including apoptosis. Neuronal apoptosis is mediated by gene activation that results in a protein cascade as the mechanism for cell death. These neuronal pathways depend on individual genomic predisposition for cerebral injury. Multiple genes have been implicated in influencing the risk of neurologic injury after cardiac surgery, including the apolipoproteins, interleukin polymorphisms, and platelet glycoprotein alleles. Future trials will undoubtedly further explore neurologic risk in cardiac surgery with CPB as a function of genotype with the goal of identifying therapeutic targets for neuronal protection in the perioperative period and beyond.
Patients undergoing cardiac surgery with CPB are typically cooled to a level of hypothermia depending on the procedure, surgeon preference, institutional norms, and patient risk factors. As cerebral temperature is lowered, cerebral metabolic rate decreases as does an ischemic tolerance with hypothermia. Hypothermia alters gene regulation of multiple inflammatory pathways to inhibit apoptosis and inflammation, as well as suppress free radical formation.
In contrast, perioperative hyperthermia has been clearly shown to be detrimental to neurological outcomes after adult cardiac surgery with CPB. This variable will be discussed in further detail in the next section.
Modifications to the Cardiopulmonary Circuit
Contemporary cardiopulmonary circuit designs have been gradually optimized to reduce the risks of the embolic generation and subsequent cerebral vasculature, given that the circuit hardware is a major source of microbubble formation and embolization. Risk factors for microbubble formation include vacuum-assisted venous drainage, roller versus centrifugal pump exposure, bubble versus membrane oxygenator application, hard-shell versus soft-shell venous reservoir, and low venous reservoir levels. Furthermore, in-line arterial filters and venous reservoir filters have also been integrated in contemporary perfusion practice for removal of microbubbles and other embolic particles from the CPB circuit. Although effective, these filters do not remove all emboli, despite reduced pore size. The benefits of reduced pore size must be balanced with the increased risk of hemolysis and damage to other blood components. In addition to air filters, the presence of a dynamic bubble trap in between the arterial filter and aortic cannula has been shown to reduce the incidence of microbubbles passing into the arterial circulation.
In addition to microbubbles and macroemboli, blood suctioned directly from the surgical field via cardiotomy suction is known to contain high levels of lipid particles. These lipid emboli have been associated with capillary occlusion and cerebral ischemia. In an effort to reduce the effects of lipid embolization due to blood returned from the cardiotomy suction, expert consensus has considered processing the blood through a cell-saver device before returning it to the CPB circuit. Two randomized controlled trials have explored this approach, with conflicting results, although both trials demonstrated increased transfusion in the cell-saver group. In a recent randomized trial ( N = 30) testing the benefits of a lipid microemboli filter integrated into the CPB circuit, the intervention group had a significant reduction in lipid microemboli and serum markers for neurological injury. Future trials will target ongoing improvement of clinical outcomes from a multimodal filter strategy to minimize embolic load and protect against neurological injury.
Aortic Cannulation Strategy
In addition to CPB components, investigators have also investigated the approach to aortic cannulation for CPB to reduce further the risk of cerebral embolization. Multiple clinical trials have examined the effects of aortic cannula design with respect to features such as length and tip design to reduce aortic wall sheer stress and minimize cerebral atheroembolism. The goals in this design process have been to reduce aortic jet velocity, to control direction of the jet, and to decrease turbulence with features such as a funnel and curved tip, including side apertures for dispersal of flow. Furthermore, cannulation of the distal aortic arch has also been established as a cannulation strategy to reduce the burden of cerebral emboli.
Another area of innovation in embolic reduction has been the addition of embolic protection to the design of the arterial cannula itself. There are currently two aortic cannulae that have embolic protection devices incorporated into their design, namely the CardioGard backward suction cannula (CardioGard Medical, Israel) and the Embol-X intra-aortic filtration device (Edwards Lifesciences, Irvine, California). The CardioGard cannula ( Fig. 22.1 ) is a curved-tip, 24-French dual-lumen aortic cannula. The first lumen is a standard forward-flow lumen for CPB arterial outflow and the second lumen is placed to suction, venting back into the venous reservoir of the CPB circuit. This returned blood is then filtered by the venous reservoir filter and the arterial filter. In vitro and animal studies demonstrated a reduction in cerebral embolic load when compared with standard aortic cannulae. A multicenter, randomized controlled trial of 66 patients demonstrated a significant decrease in the new lesion volume in the CardioGard group as measured by diffusion-weighted magnetic resonance imaging.
The Embol-X intraaortic filtration device ( Fig. 22.2 ) is a 24-French aortic cannula with a built-in side port through which the filtration device can be deployed and recaptured. In early studies, the device was deployed before the removal of the aortic cross-clamp and left in place until the patient was weaned from CPB. Early results demonstrated safety and facility of the procedure, as well as the effectiveness of the device in capturing embolic debris. Several trials have examined the Embol-X device in standard cannulation in open heart surgery with somewhat conflicting results. One trial has suggested that deployment of this device prior to aortic clamping, as opposed to only after removal, may improve embolic capture.
A recent multicenter, randomized controlled trial assigned 383 patients to the Embol-X device (133 patients), the CardioGard backward suction cannula (118 patients), or standard aortic cannula (132 patients) during surgical aortic valve replacement. In this landmark trial, significant embolic debris was found in 99% of the Embol-X filters after retrieval and 75% of the CardioGard suction cannulae. The investigators found no significant difference between groups with regards to clinical endpoints such as clinically apparent stroke and stroke volume as quantified by diffusion-weighted magnetic resonance imaging. There was no difference in mortality, length of hospital stay, and hospital readmission rate. In secondary outcome analysis, embolic protection was significantly associated with reduced risks of perioperative delirium and acute kidney injury. Overall, this important trial did not find major significant neuroprotective benefit from embolic protection. Future adequately powered trials will probably focus on high-risk patients in the search of further neuroprotective benefit from refinements in aortic cannula design.
Reducing Gaseous Emboli with Carbon Dioxide
Cardiac surgical patients during CPB are exposed to a high risk of gaseous emboli because of factors such as open chamber procedures, vascular cannulation, and assisted vacuum venous drainage. Once gas emboli reach the cerebral vasculature, they may occlude small vessels causing ischemia, or activate an inflammatory cascade further exacerbating injury. It has been hypothesized that by flooding the surgical field with highly blood-soluble carbon dioxide, most of the gas introduced into the patient will dissolve, reducing the risk of embolization. Although multiple trials have tested his hypothesis by demonstrating reduced intracardiac gas, they have collectively not proven that there is clinical neuroprotection from this intervention. A recent high-quality meta-analysis demonstrated no significant difference in the incidence of clinical stroke (1.0% versus 1.2%, P = 0.62) or neurocognitive decline (12% versus 21%, P = 0.35) with carbon dioxide exposure, despite a reduction in gaseous microemboli as detected by transesophageal echocardiography. Although there is evidence of reduced gaseous embolic load from flooding the surgical field with carbon dioxide, this has not translated to a reduction in stroke or neurocognitive impairment. A large randomized control trial is currently in progress to examine the effects of carbon dioxide application during coronary artery bypass grafting. The results of this trial may determine the future niche for this intervention in cardiac surgery with and without CPB.
Assessment of Aortic Atheroma
The presence of atheroma in the ascending aorta is a significant risk factor of neurological injury in cardiac surgery as it is a major source of cerebral emboli. Aortic events such as cannulation, clamping, graft anastomosis, and manipulation may dislodge atheromatous debris and precipitate significant cerebral atheroembolism. Furthermore, the “sandblasting” effects of blood flow through the aortic cannula can embolize atheromatous plaques, as outlined earlier.
Traditionally, direct digital palpation of the aorta by the surgeon guided final selection of the aortic cannulation site. In contemporary practice, aortic imaging has proven superior to digital palpation for assessment of ascending aortic atheroma. The two imaging techniques to assess aortic atheroma are transesophageal echocardiography and epiaortic ultrasound. The major limitation of transesophageal imaging is that it cannot reliably image the distal ascending aorta and proximal arch because of the interposition of the tracheobronchial tree at this mediastinal level. This “blind spot” of transesophageal imaging has been overcome with the A-view technique. This modified technique involves the introduction of a balloon catheter into the left main bronchus, which is then filled with sterile saline, allowing visualization of the distal ascending aorta from the esophagus ( Fig. 22.3 ).
When epiaortic imaging has been applied to identify atheroma, it is more likely to prompt a change in aortic cannulation site or strategy compared with palpation alone. Large trials have demonstrated that epiaortic imaging was significantly associated with a reduction in stroke risk after adult cardiac surgery with CPB. The cumulative evidence suggests that epiaortic scanning provides the best method overall for grading ascending aortic atheromatous disease to guide aortic cannulation and to reduce consequent stroke risk.
Blood Gas Management During Cardiopulmonary Bypass
There has been significant debate regarding the optimal pH management strategy during cardiopulmonary bypass. In the adult population, alpha-stat management is most often utilized. In alpha-stat management, the hypocarbia and alkalosis created by hypothermia is not corrected and the pH and carbon dioxide in the blood is measured at a temperature of 37°C. Alpha-stat maintains electrochemical neutrality and preserves the coupling between cerebral metabolic rate and cerebral blood flow (both decrease with cooling). In pH-stat management, carbon dioxide is added to the CPB circuit to account for the increase in solubility as cooling occurs, in an effort to maintain a normal pH and PaCO 2 . The increased levels of CO 2 lead to increased cerebral blood flow and oxygen delivery, but also increase the risk of cerebral embolization. A landmark randomized controlled trial ( N = 316) demonstrated a significant cognitive benefit to alpha-stat management in adult patients undergoing CPB for more than 90 min. The current evidence suggests that in cardiac surgery with deep hypothermic circulatory arrest alpha-stat management may be advantageous in adult cardiac surgery. In pediatric cardiac surgery with CPB, pH-stat management may be preferred for more efficient cooling of the brain and better neuroprotection.
Temperature Management: Hypothermia and Hyperthermia
In cardiac surgery with cardiopulmonary bypass, patients are often cooled to a certain hypothermia level, as outlined earlier. The benefits of hypothermic CPB have not been consistently demonstrated across multiple studies. Despite large meta-analyses (cumulative N > 10,000: normothermic versus hypothermic CPB), there was no significant difference in neurological outcomes in both adult and pediatric practice. It remains challenging with the current evidence to rule out hypothermic CPB as being beneficial because of the excessive heterogeneity across clinical studies.
In contrast, perioperative hyperthermia has been clearly shown to be detrimental to neurological outcome in cardiac surgery with CPB. In a clinical trial evaluating the clinical effects of temperature management during CPB and cardioplegia, the investigators found an increased stroke risk in the setting of warm CPB with warm cardioplegia. Perioperative hyperthermia has also been significantly associated with a higher risk of cognitive dysfunction after cardiac surgery with CPB. Multiple clinical trials have shown a reduction in neurological injury after CPB with a slower rewarming strategy during CPB. A comprehensive temperature management strategy during and after cardiac surgery with CPB is an important neuroprotective strategy in perioperative practice.
Blood Pressure Management During Cardiopulmonary Bypass
Cerebral ischemia can result not only from the obstruction of blood flow caused by embolic events, but also from inadequate perfusion pressure despite cerebral autoregulation. Higher blood pressure goals may be required in the settings of chronic hypertension and cerebrovascular disease to ensure adequate cerebral perfusion pressure during CPB. In a single-center randomized trial ( N = 248), a high mean arterial pressure (80–100 mmHg) compared with low mean arterial pressure (50–60 mmHg) during CPB significantly decreased the incidence of combined cardiac and neurological complications (4.8% versus 12.9%: P = 0.026). A second single-center randomized trial ( N = 92) demonstrated that a lower mean arterial pressure (60–70 mmHg) compared with a higher mean arterial pressure (80–90 mmHg) during normothermic CPB was significantly associated with increased risks of delirium ( P = 0.017) and cognitive dysfunction ( P = 0.012). However, a recent large, multicenter trial ( N = 197) found no significant difference in neurological injury with respect to clinical or radiographic stroke whether the mean arterial pressure was higher (70–80 mmHg) or lower (40–50 mmHg) during CPB. These conflicting trials suggest that it may not be appropriate to standardize a target mean arterial pressure during CPB but rather to consider the determination of an individual patient’s cerebral autoregulation thresholds. This determination would then facilitate individualized maintenance of perfusion pressure above the lower limit of cerebral autoregulation during CPB to ensure adequate cerebral perfusion and optimal neuroporotection.
Cerebral oximetry with near-infrared spectroscopy has displayed significant promise for estimating cerebral autoregulation and perfusion. Clinical trials have identified that low cerebral oximetry values significantly predict the risk of delirium after CPB. A recent meta-analysis ( N = 2057: 15 randomized controlled trials) found that cerebral oximetry-guided blood pressure management was associated with a significant reduction in cognitive dysfunction (risk ratio 0.54; 95% confidence interval 0.33–0.90; P = 0.02). Although this meta-analysis has suggested a neuroprotective benefit with cerebral oximetry, this result should be regarded as mostly hypothesis-generating, given the high levels of heterogeneity across the included trials. Future large, randomized trials with standardized monitoring and management protocols will probably evaluate the value of cerebral oximetry in the management of cerebral autoregulation and perfusion during adult cardiac surgery with CPB.
Pulsatile Perfusion During Cardiopulmonary Bypass
Recent clinical trials have suggested that pulsatile as compared with the typical non-pulsatile perfusion during CPB is more physiologic with enhanced preservation of the microcirculation and attenuation the systemic inflammatory response. Clinical trials thus far have not consistently demonstrated neuroprotection because of pulsatile perfusion during CPB. With such conflicting data, there is no clear neuroprotective benefit of pulsatile flow during CPB, suggesting that future trials should address this gap.
Because of the systemic inflammatory response to CPB and its possible role in cerebral injury, the administration of corticosteroids has been examined for neuroprotective benefit. Multiple clinical trials presented results with conflicting evidence. In high-quality meta-analysis, there was no discernible neuroprotective benefit from steroid exposure. Two recent, large, multicenter randomized controlled, double-blinded trials, the Dexamethasone for Cardiac Surgery (DECS) and Steroids in Cardiac Surgery (SIRS) trials, examined the neuroprotective effects of steroids in cardiac surgery with CPB. The DECS trial randomized 4494 patients to receive 1 mg/kg of dexamethasone or placebo. The investigators found no difference in the composite primary study endpoint that included stroke: the risk of stroke was also equivalent between groups. The SIRS trial randomized 7507 patients to receive 250 mg methylprednisolone at induction and another 250 mg at initiation of CPB. The investigators found no significant difference in stroke rates between study groups. A substudy of the DECS trial also demonstrated no difference in neurocognitive dysfunction in the first year after adult cardiac surgery with CPB both at 1 month (relative risk 1.87; 95% confidence interval 0.90–3.88; P = 0.09) and at 12 months (relative risk 1.98; 95% confidence interval 0.61–6.40; P = 0.24). A recent, large, meta-analysis ( N > 7000: 56 randomized clinical trials) also did not demonstrate any neuroprotective benefits from steroid exposure during cardiac surgery with CPB.
In addition to steroids, lidocaine has also been examined for neuroprotection because of its sodium-channel blocking and potential anti-inflammatory properties. Initial clinical trials demonstrated inconsistent neuroprotective effects of lidocaine. Subsequent meta-analyses have suggested a dose-dependent protection from cognitive dysfunction after cardiac surgery with CPB, although there was significant heterogeneity across the included trials.
Magnesium has also been studied for its neuroprotective effects, including its antiexcitotoxic agent and anti-inflammatory properties. Although it is safe, magnesium does not appear to be neuroprotective in recent clinical trials. Although ketamine may reduce delirium after cardiac surgery, this result has not been consistently evident in clinical trials.
Multiple agents have been tested for neuroprotection, including propofol, calcium channel blockers, beta blockers, antioxidants, erythropoietin, piracetam, the complement inhibitor pexelizumab, the serine protease inhibitor aprotinin, and dexmedetomidine. Despite all this research, significant neuroprotective benefits have yet to be demonstrated.
Neurologic Injury After Off-pump Coronary Artery Bypass Grafting
Incidence and Significance
The incidence of clinical stroke after off-pump CABG is typically 1%–2%. Because many causes of neurologic injury during cardiac surgery are related to CPB, off-pump CABG provides a cardiac surgical option that obviates the requirements for CPB, including aortic cannulation, aortic clamping, and further aortic manipulation. Despite this, off-pump CABG is still a major surgical procedure with concomitant tissue damage, hemodynamic swings, inflammation, and manipulation of the ascending aorta to the extent required for proximal graft anastomoses, and therefore the risk of neurologic injury is still present.
Multiple high-quality multicenter randomized trials have compared the risks of neurologic injury between off-pump and on-pump CABG. Despite these excellent trials, they have collectively failed to demonstrate a clear neuroprotective benefit of off-pump CABG. A recent meta-analysis ( N > 16,900) reported an increased incidence of stroke in the on-pump CABG group, however a subgroup analysis of only high-quality randomized control trials showed no significant difference. A recent meta-analysis looking at studies with 5-year follow-up did not find a significant difference in the incidence of stroke in on-pump or off-pump CABG. In contrast, a separate large meta-analysis ( N = 123,137) revealed a statistically significant reduction in stroke associated with off-pump CABG. There may be certain high-risk subpopulations that have a greater benefit from off-pump CABG. Recent meta-analyses have shown a reduction in the stroke risk in high-risk patients because of a high-risk off-pump CABG.
Neurologic injury during off-pump CABG is most likely caused by the systemic inflammatory response, cerebral emboli, and/or cardiac manipulation, which can lead to hemodynamic instability.
Although off-pump CABG avoids CPB, multiple trials have demonstrated that it nevertheless elicits a significant systemic inflammatory response. This marked inflammatory response can play a role in organ dysfunction during the postoperative period, including delirium and stroke. The neuroprotective effects of off-pump CABG dissolve with conversion to on-pump CABG, regardless of etiology for conversion. The increased risk of neurological dysfunction in this setting may be because of the step-up in the systemic inflammatory response. Because the conversion rate is about 5%, the role of neuroprotective pharmacology has not been adequately studied in this setting.
Manipulation of the Heart and Aorta
Manipulation of the aorta and heart, especially lifting of the heart, leads to acute and severe decreases in preload and stroke volume, resulting in compromised cerebral perfusion that can precipitate emergency conversion to on-pump CABG. Multiple trials have reported significant drops in jugular bulb saturation during manipulation of the heart, which may be tied to the neurologic injury seen during off-pump CABG. There may be an expanded role for cerebral oximetry in this setting, given its ability to diagnose cerebral hypoperfusion, as outlined earlier.
By avoiding aortic cannulation and clamping, it is likely that there is a reduction in the risk of cerebral emboli during off-pump CABG. However, clinical trials clearly demonstrate that cerebral emboli still occur during off-pump CABG with a consequent risk of stroke. Techniques to minimize aortic manipulation are discussed in the next section as a way to reduce stroke risk.
Strategies for Neuroprotection
Although the systemic inflammatory response of CPB may be avoided by off-pump techniques, a certain level of inflammation still ensues. Furthermore, manipulation of the heart to perform distal anastomoses is unavoidable. However, surgical techniques exist to reduce the manipulation of the ascending aorta, consequently reducing hemodynamic swings and cerebral embolization.
Surgical techniques to perform proximal anastomoses during off-pump CABG include using a partial side-biting clamp, clampless technique using a stabilizing device, or a complete aortic “no-touch” technique utilizing only the internal mammary artery with Y-grafts off the mammary artery. A large meta-analysis ( N = 25,163) compared a complete aortic “no-touch” technique with a proximal anastomotic device with or without a partial side clamp and found a significantly lower risk of cerebrovascular accident in the “no-touch” group. This was demonstrated again in a recently published, large meta-analysis of 37,720 patients comparing anaortic off-pump CABG (no touch) and off-pump CABG with the remainder of the CABG techniques. This high-quality analysis found that the anaortic approach was most effective in reducing the risk of postoperative stroke at 30 days.
Neurological Injury After Transcatheter Aortic Valve Replacement
Incidence and Significance
Since the first transcatheter aortic valve replacement (TAVR) was performed in 2002, this procedure has evolved into a mainstream therapeutic option to manage severe aortic stenosis. The indications for TAVR have gradually expanded from patients with excessive operative risk to high-risk and now to intermediate-risk patients with severe aortic stenosis. As the indications for TAVR are presently expanding to low-risk patients with severe aortic stenosis, it remains a priority to evaluate the clinical outcomes of TAVR against surgical AVR, which is the current gold standard for this population.
Despite the clinical progress, neurologic injury after TAVR remains a serious complication of TAVR. In the landmark trials for balloon-expandable TAVR, the reported stroke incidences at 30 days in the inoperable, high-risk, and intermediate-risk cohorts were 6.7%, 5.5%, and 6.4% respectively. The stroke risk, however, has steadily fallen to below 5% in recent years. This steady reduction in stroke risk is probably attributable to multiple factors, including technological advances in the newer generation devices, increased operator experience, and inclusion of lower-risk patients.
Etiologies of Neurological Injury after TAVR
In studying cerebrovascular events after TAVR, a temporal pattern has been established, with approximately one-half of events occurring within 48 hours of TAVR (early phase), an increased incidence within 30 days (delayed phase), and another increase at 1 year (late phase). This temporal distribution suggests varied causes of stroke, including procedural factors in the early phase and patient-related and postoperative treatment-related factors in the late phase. The etiologies of stroke after TAVR include procedural and nonprocedural factors ( Table 22.4 ).
|Procedural factors:||Catheter manipulation within an atheromatous thoracic aorta |
Hardware manipulation across a calcific aortic valve
Balloon inflation events
|Procedural factors:||Air embolism |
|Procedural factors:||Watershed ischemia caused by hypoperfusion |
Acute hypertension after valve deployment
|Nonprocedural factors||Female gender |
Chronic kidney disease
Atheromatous arterial disease
Because of the calcific nature of the stenotic aortic valve and coexisting atherosclerotic disease of the proximal thoracic aorta, hardware manipulation during TAVR of guidewires, delivery catheters, and balloon valvuloplasty results in cerebral embolization in almost all cases with a risk of ischemic stroke. The embolic material has been identified as including thrombus, calcium, tissue fragments, and foreign material. Multiple clinical trials have also demonstrated with intraprocedural transcranial Doppler ultrasound that microembolization occurs throughout the procedure, most frequently during valve positioning, deployment, repositioning, and balloon valvuloplasty. These data suggest that limiting embolization events may reduce the risk of stroke after TAVR.
Atrial fibrillation, especially new-onset atrial fibrillation, also increases the risk of stroke associated with TAVR. A meta-analysis ( N = 14078: 26 clinical trials) reported from the TAVR population the incidences of chronic and new-onset atrial fibrillation to be 33.4% and 17.5% respectively. Furthermore, multiple trials have also demonstrated that atrial fibrillation significantly increased the risk of stroke after TAVR because of cerebral embolization.
Nonembolic sources of cerebral injury during TAVR are primarily caused by hypotension with falls in cerebral perfusion. Severe hypotension can occur as a result of induction of general anesthesia, bleeding, valve positioning, and rapid ventricular pacing during valve intervention. Cardiovascular collapse may also accompany rare but serious complications in TAVR such as acute aortic regurgitation after valve deployment, coronary occlusion, aortic rupture, and acute aortic dissection. Fluctuations in cerebral perfusion pressure can severely decrease blood flow in the vulnerable cerebral watershed areas causing ischemic insult, but also decrease washout of embolized material, exacerbating their effects. In a study of 214 strokes after TAVR, computed tomography demonstrated both embolic and perfusion-related injury patterns. In contrast to cerebral hypoperfusion, acute hypertension following acute relief of aortic stenosis with valve deployment in TAVR can lead to cerebral hyperemia and hemorrhagic stroke.
Preexisting factors such as chronic hypertension, history of stroke, carotid vascular disease, diabetes, chronic atrial fibrillation, and smoking may exacerbate the risk of stroke after TAVR. Furthermore, a large systematic review ( N = 72,318) identified the following significant predictors for stroke after TAVR: female gender, chronic kidney disease, new-onset atrial fibrillation, and enrollment within the first half of a center’s experience with TAVR. Another clinical trial highlighted the following significant procedural predictors for stroke after TAVR: total procedural time, total time the delivery catheter was in vivo, rapid ventricular pacing, and valve repositioning. This evidence underlines the importance of patient- and procedure-specific factors in the pathogenesis of stroke after TAVR ( Table 22.4 ). These collective risk factors for stroke after TAVR can be managed in a systematic perioperative fashion to minimize the stroke risk, as outlined in the following section ( Table 22.5 ).
Preoperative neuroprotection strategies center around the optimization of baseline comorbidities, including hypertension, hyperlipidemia, diabetes, smoking, and atrial fibrillation, given their roles in stroke risk. Furthermore, detailed preoperative evaluation typically includes high-quality imaging that can facilitate the assessment of stroke risk factors such as aortic atheroma, aortic valve calcification severity, left atrial size (risk of atrial fibrillation), carotid vascular disease, and baseline cerebrovascular disease.
Heart team experience
A major factor in reducing intraoperative stroke risk during TAVR is the experience of the heart team performing the procedure. Multiple clinical trials have demonstrated that stroke risk is reduced as centers gain experience because of factors such as reduced manipulation of hardware within the aorta and shorter procedural times. The maturation of alternative access routes for TAVR has offered the heart team viable options to avoid severely diseased aortic segments to minimize cerebral atheroembolism and subsequent stroke.
In the first decade of TAVR, general anesthesia was the anesthetic of choice. As the TAVR procedure has matured and heart teams have gained considerable experience, monitored anesthesia care has become a common anesthetic approach. The advantages of general anesthesia include a secure airway, immobile patient, quick conversion to open surgery, and the ready suitability for transesophageal echocardiography. The advantages of monitored anesthetic care include shorter procedural times, reduced intraoperative inotrope and vasopressor requirements, lower hospital costs, as well as shorter lengths of stay both in the intensive care unit and hospital. Furthermore, this technique enables an earlier, more accurate neurological examination and perioperative surveillance. It is likely that monitored anesthetic care will evolve into the preferred anesthetic approach for transfemoral TAVR, given the steady advancements in valve design and heart team experience.
The convincing evidence for cerebral embolization during TAVR has prompted the development of devices to capture this debris before it reaches the cerebral circulation. To date, three such devices have reached clinical trials ( Fig. 22.4 ). The Embrella Embolic Deflector (Edwards Lifesciences, Irvine, California) consists of two porous polyurethane petals within a nitinol frame that cover the right brachiocephalic trunk and left common carotid artery during TAVR deployment ( Fig. 22.4 ). The device is deployed through the right radial artery. The filters allow blood to flow through while particles larger than 100 μm are deflected away from the protected vessels. Two pilot clinical trials have demonstrated reductions in both embolic events on transcranial Doppler and new embolic lesion volume on magnetic resonance imaging. Despite this embolic protection with the Embrella device, there was no reduction in clinical stroke rate in these initial clinical trials.
The Sentinel Cerebral Protection System (Claret Medical, Santa Rosa, California) is the second generation of a device previously known as the Claret Montage Dual Filter System. The refined design included an improved delivery system and two independent filters that cover the right brachiocephalic trunk and left common carotid artery ( Fig. 22.4 ). A pilot randomized clinical trial ( N = 63) demonstrated the safety and feasibility of this device, with a significant reduction in both new cerebral lesions and new neurocognitive deficits. A recent, larger randomized controlled trial ( N = 363) demonstrated a favorable trend both in clinical stroke and neurocognitive function when using this device.
A recent prospective propensity-matched clinical trial ( N = 802;280 received the Sentinel device) demonstrated a significant reduction in clinical stroke associated with embolic protection from the Sentinel device (odds ratio 0.29; 95% confidence interval 0.1–0.93; P = 0.03). The composite endpoint of all-cause stroke and mortality was also significantly reduced in the embolic protection group (odds ratio 0.30; 95% confidence interval 0.12–0.77; P = 0.01). The main limitation of this clinical trial was that lack of randomization, although propensity score matching accounted in part for possible confounders.
The TriGuard Embolic Deflection System (Keystone Heart, Caesarea, Israel), deployed from the femoral artery, consists of a mesh filter that covers all three major arch vessels ( Fig. 22.4 ). Like the Embrella deflection system, this TriGuard deflects debris into the descending aorta and away from the cerebral circulation. Unlike the previous two embolic protection devices, the TriGuard protects all regions of the brain. A pilot randomized clinical trial demonstrated the safety of the device and showed a non-significant decrease in new cerebral lesions and neurologic deficits compared to the control arm. Although nonsignificant, this favorable trend has laid the groundwork for a larger clinical trial that is currently underway.
Several meta-analyses have been undertaken with the goal of increasing the power of these smaller device studies to discover whether a clear benefit exists for embolic protection. A recent meta-analysis ( N = 1225) demonstrated a significant reduction in stroke risk in the first week after TAVR (relative risk ratio 0.56; 95% confidence interval 0.33–0.96; P < 0.05). A second meta-analysis demonstrated a significant reduction in new ischemic cerebral lesion burden. A third meta-analysis also demonstrated a significant reduction in clinical stroke at 30 days as a result of embolic protection during TAVR (odds ratio 0.55; 95% confidence interval 0.31–0.98; P = 0.04). Because these meta-analyses have combined data from smaller studies of different embolic protection devices, there is considerable heterogeneity in the data. Large, adequately powered randomized controlled trials are still required to explore in detail the neuroprotective effects of these devices in TAVR.
Although embolic protection devices have yet to demonstrate a definite neuroprotective advantage, they are being adopted in clinical practice, given the favorable safety and efficacy. Thus, although the application of these devices is not yet a standard of care in TAVR, their deployment is safe and reasonable.
Because new-onset atrial fibrillation increases the risk for stroke after TAVR, its early detection and protocol-driven management are essential. As a result of the thrombogenic nature of the TAVR procedure and prosthetic valve, optimal perioperative anticoagulation is necessary to prevent thromboembolism ( Table 22.5 ). Although intravenous heparin is preferred for adequate intraoperative anticoagulation, such levels of anticoagulation are unacceptable postoperatively because of the high risk of major bleeding. Dual antiplatelet therapy with aspirin and clopidogrel has been recommended to minimize the risks of valve thrombosis after TAVR. Recent meta-analysis has challenged this practice because of the excessive bleeding risk (relative risk 2.52; 95% confidence interval 1.62–3.92; P < 0.0001) with no difference in stroke risk, suggesting that monotherapy may be adequate. In TAVR patients who take coumadin for atrial fibrillation, dual antiplatelet therapy is also not recommended because of the excessive risk of bleeding. Future trials will identify the next antithrombotic strategy after TAVR, examining the options both for platelet blockade and the novel oral anticoagulants.
Neurologic Injury During Thoracic Aortic Surgery
Thoracic aortic disease is a major cause of mortality and morbidity, with aortic aneurysm and dissection as the most common pathologies. The aorta has five segments from its origin at the aortic valve to its termination at the iliac bifurcation in the abdomen ( Fig. 22.5 ). An aortic aneurysm is defined as having an aortic diameter greater than 150% of normal with all three layers of the aortic wall still intact. Aortic aneurysms are classified according to their size, shape, and extent. In the proximal thoracic aorta, an aneurysm may involve the aortic root, the ascending aorta, and/or the aortic arch ( Fig. 22.5 ). An aortic aneurysm distal to the arch may involve the descending thoracic aorta and/or the abdominal aorta ( Fig. 22.6 ). The timing of intervention for a given aneurysm depends on factors such as clinical presentation, location, size, and comorbidities. Aortic dissection is caused by an intimal tear that can extend to involve a variable extent of the aorta with both a true and a false lumen. This challenging disease process is classified according to the dissection extent in the DeBakey or Stanford systems ( Fig. 22.7 ). The timing of intervention for a given aortic dissection depends on factors such as clinical presentation, location, and extent. Thoracic aortic disease can present with cerebral and/or spinal stroke as a result of hypotension from aortic rupture, emboli from an aneurysm, and/or malperfusion from dissection.
The open surgical management of thoracic aortic disease entails resection of the diseased aortic segment with subsequent replacement with vascular tubular graft. If the disease is isolated to the ascending aorta, then this repair is achieved using traditional CPB with similar risks of neurologic injury outlined in the earlier section “Neurological Injury after Cardiac Surgery with Cardiopulmonary Bypass”. If, however, the aortic pathology involves the aortic arch and/or descending thoracic aorta, then surgical repair poses unique threats to the central nervous system.
When an aortic aneurysm or dissection affects the aortic arch, from which the cerebral arterial vasculature arises, the repair of this lesion requires the temporary cessation of blood flow to the brain. This circulatory arrest allows surgical exposure to facilitate reconstruction of the aortic arch. To mitigate its cerebral ischemic effects the period of circulatory arrest is typically performed with a degree of hypothermia with or without selective cerebral perfusion. Despite these advances, the stroke rate during aortic arch surgery remains between 4% and 10%, with higher risk during emergency settings such as acute type A dissection.
Just as aortic arch surgery requires specialized maneuvers to protect the brain, surgical repair of the thoracoabdominal aorta requires tailored management to protect the spinal cord because the spinal cord arterial network receives significant input from the thoracic aorta at multiple levels ( Fig. 22.8 ). In this surgical setting, spinal cord ischemia manifesting as paraparesis or paraplegia remains a serious complication that significantly increases mortality and morbidity after thoracoabdominal aortic interventions. Although the risks of spinal cord ischemia has decreased steadily with the multimodal advances in open and endovascular thoracic aortic repair, the incidence of spinal cord ischemia remains in the 2%–8% range, depending on the level of operative risk.