Clinical manifestations

Presentation and course

Neurologic complications of heart transplantation.

The main neurologic complications of cardiac transplantation are:

• Ischemic stroke • Seizure • Opportunistic infections • Post-transplantation lymphoproliferative disorders

Cerebrovascular events occurring soon after transplantation resemble those seen after valvular or coronary revascularization surgery. The risk factors include hypotension, cardioembolism, seizures, metabolic disturbances, and medication side effects. In a series of 322 patients, preexisting valvular heart disease was associated with ischemic stroke, whereas diabetes and renal failure were associated with seizures (122).

Circulatory support during cardiac transplantation is maintained by left ventricular assist device or total artificial heart. These devices are thrombogenic and may cause cerebral embolism (159).

Other embolic sources may persist after cardiac transplant, particularly in the aortic arch and the cervical vessels. Neurologic complications occurring weeks to months later include opportunistic infections, side effects of immunosuppressants, and post-transplantation lymphoproliferative disorders.

Neurologic complications of coronary artery bypass graft surgery.

	• Central nervous system complications
		- Ischemic stroke and transient ischemic attack - Depressed level of alertness - Altered awareness - Seizure - Asymptomatic embolism
	• Peripheral nervous system complications
		- Brachial plexus injury - Phrenic nerve injury

Ischemic stroke. Experience at a single center including 45,432 patients who underwent coronary artery bypass graft revealed that perioperative stroke rate has declined over the last 30 years. Ischemic stroke occurs postoperatively more often than intraoperatively. Intraoperative stroke rates were lowest with off-pump coronary artery bypass graft (0.14%) and on-pump beating-heart coronary artery bypass graft (0%) and highest with on-pump coronary artery bypass graft with hypothermic circulatory arrest (5.3%) (157).

Brain MRI detects more ischemic lesions than head CT. The most common neuroimaging pattern is of multiple embolic punctate infarcts in the cerebral and cerebellar cortex, including in patients with encephalopathy without focal neurologic findings (172).

Microembolic perioperative stroke occurs with predilection in the border zone areas, but also in subcortical regions resembling lacunar lesions (69; 96; 91).

Involvement of the border zone between the middle and posterior cerebral arteries produce visual field defects and transcortical sensory aphasia. Lesions located between the middle and anterior cerebral arteries result in proximal-dominant extremities weakness (man in the barrel syndrome) and transcortical motor aphasia. Infarctions between superficial and deep branches of middle cerebral arteries may present as mutism and rigidity. Watershed infarct of cerebellum can present as dizziness or vertigo, and ataxia (68; 62).

Border zone infarction in the spinal cord occurs often at the midthoracic level between the spinal arteries and artery of Adamkiewicz. This may result in acute urinary retention, flaccid paraparesis, and sensory loss with level (62).

Of interest is the timing of surgery in patients with both carotid artery stenosis and coronary artery disease. In simultaneous surgery, the rate of stroke was higher but mortality was similar compared to surgery-carotid artery stenosis, followed by coronary artery bypass graft. However, there were no significant differences in 1-year mortality, and rates of myocardial infarction and transient ischemic attacks (29).

Depressed level of alertness. Decreased consciousness suggests diffuse encephalopathy. Somnolence occurs when the patient is fully attentive to surroundings while aroused, but alertness decreases rapidly to a sleep-like state. Brief arousal associated with decreased awareness or interest in surroundings is called obtundation. Decreased alertness associated with grimacing or withdrawal from painful stimuli characterizes stupor.

The persistence of sleep-wake cycle characterizes the vegetative state. Chewing, teeth grinding, or scratching movements can occur. Persistent vegetative state lasts more than four weeks, and it becomes permanent after three months.

The comatose patient is not arousable, but the brainstem reflexes are preserved.

Brain death is characterized by absence of alertness to external stimuli and of all brainstem reflexes.

Nonconvulsive status epilepticus must be excluded in patients with altered alertness. Electrolyte imbalance, acute liver or renal dysfunctions, hypoxia, or hypotension should also be considered. Stroke does not usually affect alertness unless it involves the reticular activating system. Large hemorrhagic or ischemic strokes can decrease alertness by increasing intracranial pressure.

Altered awareness. Confusion is characterized by normal alertness, but altered awareness, orientation, or memory. Restlessness or agitation are typical for delirium. Together, they cause cognitive impairment.

Approximately half of the coronary artery bypass graft procedures result in early cognitive impairment. Improvement occurs within six months, but in one third of cases, the deficits persist up to 12 months (112). Two thirds of patients have complete resolution of neurologic deficits during long-term follow-up (88). Delayed cognitive impairment may occur after five years (113).

Prior stroke is a risk factor for delirium after coronary artery bypass surgery (134). MRI imaging in patients with cognitive impairment can reveal multiple small brain lesions (84).

Cognitive impairment may be related to cerebral microembolism, hypoperfusion, edema, inflammation, or metabolic derangements.

The levels of tau protein and neurofilament light protein, specific markers for neuronal injury, may help predict the risk of delirium (136; 27). Tau protein level improves with the clinical status of the patient (07).

Seizure. Incidence of seizures after cardiopulmonary bypass is about 3% (51). Etiology is multifactorial, including stroke, air embolism, atherosclerotic debris, release of inflammatory mediators during surgery, and exposure to potentially epileptogenic medications such as antifibrinolytics and cephalosporins.

Asymptomatic embolism. Asymptomatic embolism occurs frequently during coronary artery bypass and endovascular procedures (54). These lesions may not be truly asymptomatic because they may cause memory loss or dementia.

Brachial plexus injury and mononeuropathies. Brachial plexus injury and mononeuropathy may occur after compression or stretching during surgery. The lower trunk of brachial plexus is more vulnerable than the upper trunk. Injury to the phrenic nerves with unilateral or bilateral diaphragmatic paralysis may result from direct manipulation of the nerve, ischemia, or cold cardioplegia (39; 76).

Neurologic complications of left heart catheterization.

	• Central nervous system complications
		- Ischemic stroke and transient ischemic attack - Altered awareness (delirium) - Depressed level of alertness (transient unresponsiveness) - Transient global amnesia - Seizure - Contrast reaction - Migraine
	• Peripheral nervous system complications
		- Lumbosacral plexopathy - Femoral nerve injury

CNS complications after left heart catheterization are rare. Ischemic stroke is the most common (55%), followed by transient ischemic attack (26%), contrast reaction (15%), and neurologic symptoms of unknown etiology (4%). The most common symptom is visual disturbances including diplopia and visual loss. Hemiparesis or aphasia, facial weakness with or without aphasia/dysarthria, and paresis in face and arm were also described. Transient unresponsiveness, transient global amnesia, seizures, delirium, and migraine may follow left heart catheterization.

Ischemic stroke caused by large vessel occlusion has worse outcome compared with small vessel occlusion (139). Cerebral embolism usually originates from the atheromatous aortic arch or the clots forming at the catheter tip.

Femoral nerve injury. Femoral artery puncture can rarely cause pseudoaneurysm, hematoma, arteriovenous fistula, and femoral artery occlusion. Femoral nerve compression (incidence of 0.2%) severity ranges from mild transient sensory neuropathy to disabling paralysis. The average delay from catheterization to recognition of symptoms is 37 hours; severe pain precedes the diagnosis of neuropathy in about half of patients. Although initially disabling, neuropathy usually resolves completely; surgery is only recommended if there are coexisting complications (79).

Lumbosacral plexopathy. Retroperitoneal hematomas can cause lumbosacral plexopathy, with femoral and obturator nerves being primarily involved. The size of the hematoma does not correlate with the severity of the sensory or motor deficits.

Neurologic complications of other cardiac procedures.

Aortic aneurysm repair. Aneurysms of the aortic arch are often treated by surgery. In a series of 284 patients undergoing aortic aneurysm repair, death or stroke occurred in 6.6% (154). Transient neurologic dysfunction occurred in 9.2%. Two patients developed brachial plexus injury from axillary artery cannulation. In another report of 103 patients undergoing total aortic arch replacement, about 9% developed stroke, and approximately 7% had transient neurologic deficits. Four percent also had paraparesis or paraplegia (164).

Hypothermic circulatory arrest and selective cerebral perfusion are used to minimize these complications. In a series of 39 consecutive patients undergoing aortic arch replacement surgery with selective brain perfusion by axillary artery cannulation, only one patient had a permanent postoperative neurologic deficit, and two patients had temporary deficits (145).

Endovascular repair of the aortic arch with branched endografts is an alternative to open surgical repair. Out of 147 (99.3%) procedures that were successful, disabling stroke has occurred in six (4.1%) patients, and nondisabling stroke in eight (5.4%) patients suffered (156).

The risk factors for spinal cord ischemic during aortic aneurysm repair are age greater than 65 years; history of neurologic disease, hyperlipidemia, diabetes, coronary artery disease, heart failure, and renal insufficiency; less than six months since last aortic repair; chronic anticoagulant use; preoperational hemoglobin less than 9; coagulopathy; hypotension (mean arterial pressure < 70 mm Hg); longer operations (> 100 min); aneurysms longer than 5 cm; and anatomic location of aneurysm caudal to T-11 (11).

Aortic dissection repair. Aortic dissection classification aids in standardizing the management. The American cardiothoracic surgeon DeBakey classified aortic dissection in type I when both ascending and descending segments are involved, type II involves the ascending aorta, and type III involves the descending aorta after the origin of subclavian artery. Stanford type A dissection includes the segment before the left subclavian artery (LSA) and Stanford type B the segment beyond the left subclavian artery. Aortic dissection surgery is lifesaving but can be very complex and associated with severe neurologic symptoms.

Aortic dissection DeBakey type I, which involves both the ascending and descending aorta, remains a severe condition whose repair may result in stroke in 8% and cerebral edema in 1% of patients (41).

In a metaanalysis of 3000 patients who underwent frozen elephant trunk technique for type A aortic dissection repair, the pooled rates of adverse events were 4.7% for spinal cord stroke, 7.6% for cerebral stroke, and 8.8% for perioperative mortality. The LSA may occasionally be sacrificed causing exertion-induced weakness. This may be treated with carotid-subclavian bypass (50).

Retrograde type A acute aortic dissection may benefit from endovascular repair. A retrospective study of 19 patients from one medical center who underwent endovascular repair found no in-hospital or neurologic complication (85).

Type A aortic dissection surgical repair may also cause hemorrhagic stroke, especially if preoperative antithrombotic medication is used. Hemorrhagic stroke was associated with higher complexity of surgery and longer cardiopulmonary bypass and aortic clamping time compared to ischemic stroke. Inpatient mortality, degree of disability, and 5-year mortality were higher in the hemorrhage group (92).

Type B aortic dissection, beyond the left subclavian artery, can lead to life-threatening complications, from rupture, to stroke and paraplegia. There is insufficient information regarding the effectiveness and safety of open surgical repair compared to endovascular approach (72). During thoracic endovascular aortic repair (TEVAR) of type B aortic dissection, the origin of the left subclavian artery is covered in approximately half the cases. The main complications are arm claudication, stroke, and paraplegia (170).

A metaanalysis of left subclavian artery coverage revascularization did not reduce the risk of stroke or mortality (59). However, those who had revascularization before or during TEVAR were less likely to develop neurologic complications (109).

Spinal cord ischemia was significantly higher when the length of graft exceeded 15 cm or extended beyond T8 (126).

Although unrepaired acute aortic dissection is usually fatal, some unoperated patients survive to the chronic phase. Out of 205 patients operated after a median duration of dissection of approximately seven months, 3% had a stroke and 7% had a fatal outcome (174).

Placement of a lumbar drain preoperatively is indicated for prevention of spinal cord ischemia during TEVAR (47). A study of prophylactic lumbar drain that included 309 procedures performed in 268 patients for 307 endovascular aortic repairs found a complication of 8.1% (4.2 major and 3.9 minor). The complications included paraplegia, intracranial hemorrhage, meningitis, arachnoiditis, CSF leak, bloody tap, and a retained catheter tip (124).

Cardiac valve surgery.

	• Ischemic stroke or transient ischemic attack
	• Seizure
	• Vasovagal syncope

Heart valve repair and replacement have been performed for several decades. The incidence of neurologic complications after heart valve surgery is less documented in the literature compared to the complications after coronary artery bypass surgery.

Balloon valvuloplasty. Balloon valvuloplasty of the aortic and mitral valves is effective and safe. Of 674 patients enrolled in the National Heart Lung and Blood institute (NHLBI) registry who underwent percutaneous aortic balloon valvuloplasty, 2% had strokes, 5% had seizures, 2% had vasovagal syncope, and 0.6% developed transient neurologic deficits in the peri-procedural period (114).

Aortic valve replacement. Severe symptomatic aortic stenosis is treated by surgical aortic valve replacement (SAVR). In a single center retrospective analysis of 132 patients who underwent aortic valve replacement, there were no recorded strokes (31).

Transcatheter aortic valve replacement (TAVR) is less invasive and used for symptomatic aortic stenosis with high surgical risk.

In a study of 1660 patients with severe, symptomatic aortic stenosis of intermediate surgical risk, the rate of stroke was lower with TAVR than with SAVR. Moreover, those with strokes after TAVR recovered earlier. However, at 12 months, stroke rates and quality of life were similar in both study groups (42). Another study of 1204 pairs of patients treated for severe aortic stenosis found that SAVR was associated with more frequent early major strokes than TAVR (75).

In a metanalysis and systematic review, TAVR in patients with low and intermediate risk, had a similar rate of neurologic complications and mortality, a lower risk of acute myocardial infarction, and acute kidney failure, but a higher risk of arterial complications (165). Similarly, in another metanalysis and systematic review of patients at low risk, transcatheter aortic valve replacement and surgical aortic valve replacement had similar neurologic complications (135). In a study of DWI lesions seen on MRI after aortic valve replacement in 38% of patients., the preoperative T2 white matter lesions volume is related to the number of perioperative DWI lesions (153).

Mitral valve repair or replacement. Repair or replacement of the nonfunctional mitral valve is often indicated. In a database of 1250 patients, the incidence of stroke after mitral valve surgery was 4.2% to 5% (52; 160). Mitral valve repair offers excellent short- and long-term outcomes. In elderly, valve repair was less often, but not statistically significant, associated with neurologic complications (40).

Balloon mitral valvuloplasty is an alternative to open surgery for mitral stenosis. In a study of 45 patients, only one patient had a postprocedural stroke (162). Of 201 consecutive patients undergoing this procedure, three developed “thromboembolism.” It is not clear whether these patients developed strokes (173).

Combined procedures. Combined procedures such as double-valve replacements or valve replacement with coronary artery bypass graft increase the risk of neurologic complications. In a review of 2008 patients undergoing valve surgery, the overall risk of stroke was 2.2%. However, with combined procedures, the risk was 5.4%. The risk factors were calcified ascending aortic atherosclerosis, advanced age, diabetes, and ejection fraction less than 30%. Stroke increased hospital mortality from 4.6% to 24% (45).

In a study, the subcortical lesions seen on brain MRI after valve replacement surgery were associated with impaired memory, attention, and information processing speed. All these cognitive impairments resolved after four months (84).

In high-risk patients, endovascular valve replacement is preferred. However, this approach is associated with more frequent neurologic complications.

New, often multiple, ischemic lesions were observed on brain MRI of 27 (84%) patients (73). In another study of high-risk patients, although the 1-year mortality rate was similar in the endovascular and surgical group, major stroke was more frequent in the endovascular group (5.1% vs. 2.4%) (148).

Patent foramen ovale closure. Recurrent neurologic events may result from paradoxical embolism through patent foramen ovale. A clinical trial in patients with stroke attributable to patent foramen ovale presence that enrolled 980 patients between 18 to 60 years of age found that patent foramen ovale closure was not superior to medical treatment alone in intention-to-treat analysis but was superior in the prespecified per-protocol and as-treated analyses. Procedure and device related complications included groin and retroperitoneal hematomas, nerve injury, and atrial fibrillation.

Procedure-related or device-related serious adverse events occurred in 4.2% of the patients in closure group (26).

A metaanalysis including five randomized trials totaling 1829 patent foramen ovale closure and 1611 medical therapy patients found that patent foramen ovale closure prevented cryptogenic ischemic stroke and neurovascular and mortality events but increased the risk of atrial fibrillation 4.6 times (171).

Radiofrequency ablation for atrial fibrillation.

Atrial fibrillation is an important cause of stroke. Rhythm can be controlled by radiofrequency ablation (RFA) or long-term antiarrhythmic drugs (82). Although atrial fibrillation is more common in elderly, radiofrequency ablation is less effective in this population. Those at medium to high risk should continue oral anticoagulant after the procedure (152). The transseptal puncture during ablation creates an interatrial shunt, which favors paradoxical embolism. Moreover, endothelial damage may activate coagulation.

Radiofrequency ablation was associated with 1.02% neurologic complications in a study of 93,801 procedures performed in the U.S. between 2000 and 2010 (36). Analysis of 3360 consecutive cases of radiofrequency ablation revealed 0.5% peri-interventional thromboembolic events within one month, of which 45% were strokes. Most events occurred within 48 hours. Peripheral vascular disease and previous stroke were risk factors (86).

A rare, potentially fatal complication of radiofrequency ablation, which requires early diagnosis and treatment, is atrioesophageal fistula (AEF). Retrospective analysis of 28 cases of atrioesophageal fistula found that complications developed within one month from the procedure. They included fever, neurologic symptoms, and hematemesis. The neurologic symptoms are confusion, seizures, meningitis, stroke, and post-prandial transient ischemic attack. Blood tests, including cultures, were consistent with sepsis. Lumbar puncture may show pleocytosis, elevated protein, lactate, and bacteria. The fistula may be aggravated by insertion of nasogastric tube, endoscopy, or transesophageal echocardiography (151). CT of chest is the preferred diagnostic test and esophageal repair may be lifesaving (30).

Watchman left atrial appendage closure for atrial fibrillation. The Watchman left atrial closure device was noninferior to warfarin for prevention of stroke due to nonvalvular atrial fibrillation (67). In a systematic review of thrombosis associated with 2118 devices (Watchman, Amplatzer cardiac plug, and Amulet), the incidence of thrombosis was 3.9%. The median time to diagnosis was 1.5 months (89). However, in a small pilot study of 21 patients who underwent brain MRI 48 hours before and 24 hours after insertion of Watchman, Amplatzer, or Amulet devices, the rate of silent cerebral infarcts was not significantly increased (87). Recurrent neurologic symptoms may be caused by late device dislodgement, which requires device extraction (106).

Prognosis and complications

Postoperatively, prediction of early neurologic complications may accelerate diagnosis as well as early intervention (04). Less than one third of perioperative stroke cause severe disability at the time of discharge (144; 100; 91).

Several biomarkers were studied. for prediction of perioperative neurologic complications. Preoperative serum NMDA receptor antibody levels but not CRP correlates with neurologic complications in a study of 557 of high-risk patients who underwent CABG or valve replacement (15). Elevated preoperative serum C-reactive protein level correlated to higher mortality after coronary artery bypass surgery (74).

Preoperative S100-β level did not correlate with perioperative neurologic complications (15). However, persistent postoperative elevation of S100-β, a calcium regulating astroglial protein released by brain injury, correlated with cardiopulmonary bypass time and decrease in mini-mental state examination (44). Another study of 35 patients who underwent abdominal aorta repair failed to confirm a correlation between plasma or CSF S100-β level and cardiopulmonary bypass (133).

Analysis of 92 neurologic-related proteins in the serum and CSF of 23 patients, before and after aorta surgery, aimed to assess the risk of neurologic injury. Spinal cord injury was associated with IL-6, GFAP, CSPG4, delirium with TR4 and EZH2, and hallucinations with NF1 (93).

However, cognitive dysfunction after cardiac surgery does not correlate with global cortical beta-amyloid deposition (83).

The patients’ response to surgery is only partially explained by their characteristics or procedure’s variables. Knowledge of genetic variability may improve short- and long-term outcome prediction. Neurologic complications range from subtle cognitive deficits (up to 69%) to stroke and coma (up to 6%). For example, APOE genotype participates in modulating the inflammatory response, atherosclerosis, and autoregulation. Genetic variants in platelet membrane glycoproteins modulate thrombosis, and along with the inflammatory mediators, the propensity for perioperative stroke. The intensity of individual neuroendocrine and metabolic response to surgical trauma, like stress hormones, antidiuretic hormone, or the renin-angiotensin system, depends on the severity of trauma or sepsis. Insulin resistance is related to perioperative adverse events after cardiac surgery (125).

APOE4 genotype may be related to adverse cerebral events (158). Moreover, APOE4 allele is associated with worse cognitive recovery at five years after surgery (09). Genetic variants of the C-reactive protein and interleukin-6 genes are associated with stroke after cardiac surgery (56).

Biological basis

The main causes of neurologic complications of cardiac procedures are:

	• Embolism of blood clots, atherosclerotic debris, or originating in the devices used to maintain circulation.
	• Activation of coagulation on contact between blood and cardiopulmonary bypass circuit or left ventricular assist device.
	• Alteration of the cerebral excitability by medication
	• Compression of peripheral nervous fibers
	• Infection
	• Immunosuppression

Etiology and pathogenesis

Ischemic stroke or transient ischemic attack after cardiac procedures. Cerebral disorders after cardiac surgery were attributed to embolization and intraoperative alterations in blood pressure (49). In a retrospective study of 16,184 patients undergoing cardiac surgery, hypertension, diabetes, prior stroke, peripheral vascular disease, preoperative infection, and prolonged cardiopulmonary bypass were independent risk factors for stroke (21). Emboli can be divided into macroemboli, greater than or equal to 200 microns, and microemboli (14). Embolization can occur during cannulation/decannulation of the aorta, as well as during application or removal of the cross clamp.

Emboli may consist of air from the heart or open aorta, calcium from the aorta or mitral and aortic valves, and blood clot from the left atrium or ventricle. Large emboli may occlude large cerebral arteries producing symptoms corresponding to their distribution from stroke to coma (132; 16).

Pathological studies reveal dilation of small capillaries and arterioles often at bifurcations. The number of fatty microemboli increases with duration of the cardiopulmonary bypass (20).

Other microemboli. Other microemboli consist of gas bubbles from oxygenators, fat from cardiotomy cell clumps, and silicone particles from the bypass pump. The smallest emboli scatter widely and are more likely to cause generalized neuropsychological alterations than hypoxia (62). Acute brain swelling on MRI FLAIR sequence shortly after surgery appeared to involve primarily the cortex. This correlates with the pathological finding of cortex and deep gray matter arteriolar and capillary dilatations (61). These vascular dilatations after cardiopulmonary bypass are thought to be caused by embolism (105).

Activation of coagulation by the cardiopulmonary bypass circuit leads to platelets and coagulation factors consumption (02; 80). The release of tissue factor is main trigger of the coagulation during cardiopulmonary bypass (38).

Cognitive impairment after cardiac procedures. The mechanism of the biphasic cognitive decline after coronary artery bypass graft surgery is unclear. Some early short-term cognitive changes may be multifactorial and reversible (140). Perioperative ischemic stroke is not always related to cognitive decline (32). Apolipoprotein E4 genotype is also not related to the post-operative cognitive decline (150; 05).

The perioperative burden of hypertension or hypotension affects both short- and long-term outcome in cardiac interventions including stroke, delirium, and death (94).

Cerebral autoregulation in a noninjured brain is diminished perioperatively and may lead to cognitive dysfunction, worse clinical outcome, and even death (95).

Epidemiology

The main complications of cardiac procedures are:

	• Stroke
	• Encephalopathy
	• Cognitive impairment
	• Headache
	• Seizures
	• Infections

Percutaneous coronary intervention (PCI). Almost two thirds of coronary revascularization procedures are percutaneous coronary intervention, an increasingly popular approach. The incidence of stroke was 0.08% to 0.4% (70; 97). However, many perioperative embolic events are clinically silent. Using diffusion-weighted MRI, 15% of individuals undergoing cardiac catheterization had asymptomatic ischemic cerebral lesions (23).

Coronary artery bypass graft (CABG). In a retrospective study of 2985 patients who underwent coronary artery bypass graft, the incidence of stroke was 1.6%, similar between conventional coronary artery bypass graft and off-pump coronary artery bypass graft. The mechanism was large embolism in 76% and watershed in 15% of infarcts. Previous stroke, extensive aortic calcification, female gender, and congestive heart failure were predictors of stroke. The in-hospital mortality and long-term survival rate were affected by perioperative stroke (46). In prospective studies, the incidence of stroke after CABG was 1.5% to 5.2% (101; 142; 22).

Out of 4140 patients who underwent isolated coronary artery bypass graft, 2.5% had early (< 24h) and 0.9% late (> 24h to discharge) stroke. Early stroke was predicted by age, obesity, smoking, recent myocardial infarction, femoral vascular disease, reoperation for bleeding, postoperative sepsis and/or endocarditis, and respiratory failure. Delayed stroke risk factors were female gender, white race, renal, and respiratory failure. As opposed to late stroke, early stroke was an independent predictor of in-hospital and long-term mortality (161).

Combined surgery. In a single center study of 4335 patients, in coronary artery bypass graft with or without aortic valve replacement, stroke was detected in 1.8%. The combination of coronary artery bypass graft with carotid endarterectomy increased the risk of perioperative stroke significantly from 0% to 15.1% (90).

PCI versus CABG. Pooled data analysis of 11 randomized clinical trials including 11,518 patients found a reduced risk of stroke at 30 days and 5 years after the procedure. Patients who had a stroke within 30 days after the procedure had a significantly higher mortality compared to those who did not have a stroke (63).

Heart transplantation. Experience from a single center including 200 patients revealed that 23% had neurologic complications: ischemic stroke, seizures, diffuse encephalopathy, headache, and cerebral infections. Preoperative mechanical circulatory support requirement was the most important predictor of poor neurologic outcome (175). The risk factors for ischemic stroke after cardiac transplant are history of stroke and dilated cardiomyopathy (71; 12).

Prevention

Prevention of perioperative neurologic complications includes:

	• Preoperative evaluation should include history of stroke, large vessel atherosclerosis, valvular disease, and atrial fibrillation.
	• Treatment of hyperglycemia, hypotension, and hyperthermia is recommended.
	• Dexmedetomidine administered perioperatively does not prevent perioperative delirium.
	• Cerebral protection devices may be currently tested.

Thorough preoperative evaluation of risk factors, including prior stroke, carotid stenosis, and aortic atherosclerotic plaques may help prevent complications after coronary artery bypass graft surgery (25).

Preoperative statin has been shown to reduce mortality of coronary artery bypass graft surgery but not the risk of stroke (118). In a retrospective study, beta-blockers and statins reduced the number of perioperative strokes (17).

Control of atrial fibrillation, intraoperative quantitation of aortic atherosclerosis by ultrasound, and treatment of hyperthermia, hyperglycemia, and hypotension are also recommended (55; 149; 102).

Cerebral embolism after TAVR did not differ significantly between the groups receiving dual antiplatelet and anticoagulation with edoxaban (119).

In a small clinical trial, perioperative administration of dexmedetomidine lowered the incidence of delirium in patients 60 and older (168). However, a systematic review found that after excluding the trials at high risk of bias, dexmedetomidine is no longer associated with decreased delirium (120).

Neurologic complication risk factors associated with general cardiac surgery. Several studies have attempted to identify the subset of patients at high risk preoperatively.

	A. Factors predicting postoperative neurologic complications (131)
		1. Preoperative factors
			a. Carotid stenosis greater than 50% b. Prior heart surgery c. Prior stroke
		2. Intraoperative factors
			a. Valve surgery
	B. Factors predicting higher postoperative mortality (91)
		1. Preoperative factors
			a. Carotid stenosis greater than 50% b. Prior heart surgery c. Peripheral vascular disease d. Hypercholesterolemia
		2. Intraoperative factors
			a. Longer cardiopulmonary bypass
	C. Factors predicting perioperative stroke (101)
		1. Preoperative factors
			a. Age b. Previous stroke c. Presence of carotid bruit d. History of hypertension e. History of diabetes mellitus
		2. Intraoperative factors
			a. Longer cardiopulmonary bypass
	D. Factors predicting postoperative delirium (21)
		1. Preoperative factors
			a. Prior cerebrovascular disease b. Peripheral vascular disease c. Atrial fibrillation d. Diabetes mellitus e. Ejection fraction less than 30% f. Urgent operation
		2. Intraoperative factors
			a. Need for massive blood transfusion b. Operation time greater than 3 hours
	E. Factors predicting postoperative neurologic complications (141)
		1. Preoperative factors
			a. Hypertension b. Diabetes c. Prior stroke d. Smoking e. Aortic atherosclerosis
		2. Intraoperative factors
			a. Severe hypotension b. Manipulation of the atherosclerotic aorta c. Prolonged bypass time
		3. Postoperative factors
			a. Occurrence of atrial fibrillation
	F. Factors predicting postoperative seizure (98)
		1. Preoperative factors
			a. Renal dysfunction b. Advanced age c. Prior cardiac surgery d. Peripheral vascular disease e. Cardiac arrest
		2. Intraoperative factors
			a. Open heart surgery b. Prolonged cardiopulmonary bypass time (less than 120 minutes) c. Exposure to high dose tranexamic acid (less than 100 mg/kg)

Age. Advanced age is a major predictor for stroke and cognitive dysfunction following cardiac surgery (53). In a study of 2000 patients who underwent cardiopulmonary bypass, the neurologic deficits occurred 10 times 75 years compared to before 65 years of age. Mortality rate was nine times higher (35.7% vs. 4.0%) in patients with perioperative neurologic deficits (104). Fortunately, new surgical techniques have improved the outcome in elderly patients (130).

Sex. Women have an increased risk of perioperative neurologic complications and mortality (65).

Previous stroke. History of stroke predicts perioperative complications or stroke. Awakening from anesthesia is prolonged, and the likelihood for confusion or reintubation is higher. The risk of focal neurologic deficits, new strokes, and reappearance or worsening of previous stroke symptoms is increased (128; 66).

Open heart surgery. Intracardiac air embolism may complicate open heart surgery. Leaflet and annulus calcium deposits embolization during cardiac valve replacement high intensity transient signals (HITS). The number of high intensity transient signals correlates with neuropsychological outcome (127).

Ascending aorta atheroma. Presence of ascending aorta atheroma predicts neurologic events, stroke, prolonged hospitalization, and mortality after surgery (34; 35; 08; 129). Embolism is reduced by avoiding aortic manipulation during surgery (166).

Off-pump surgery lowers the incidence of stroke compared to the on-pump group (0.4% vs. 3.9%) (121). The off-pump, no-touch surgical technique avoids aortic manipulation and the effects of cardiopulmonary bypass and (129; 24). Intraoperative ultrasound of the proximal aorta helps to avoid the area of plaque before clamp placement and prevent perioperative stroke (104; 116).

Cardiopulmonary bypass. Cardiopulmonary bypass provides circulatory support during cardiac surgery. However, the bypass circuit is thrombogenic. Coating the bypass circuits with phosphoryl-choline or heparin seems to reduce the thromboembolic complications (37; 64). Longer duration of cardiopulmonary bypass increases the perioperative neurologic complications (104; 91).

In a small study, surgery performed without cardiopulmonary bypass, the off-pump technique, has reduced overall morbidity and mortality compared with on-pump surgery (03). However, another study, enrolling 2203 patients who underwent coronary artery bypass graft, did not find a significant difference in the composite outcome, stroke, or neuropsychological decline between the two groups at 30 days (146).

Perioperative mean arterial pressure. Cerebral blood flow remains constant to sustain the brain metabolism. The blood flow is defined as cerebral perfusion pressure divided by cerebrovascular resistance. Autoregulation maintains the cerebral blood flow relatively constant at cerebral perfusion pressure values between 50 to 130 mmHg.

Cerebral perfusion pressure is the difference between the mean arterial and intracranial pressure. As the intracranial pressure is constant, around 10 mmHg, it is suggested to maintain the mean arterial pressure between 60 to 140 mmHg during the procedure.

Initial studies suggested maintenance of mean arterial pressure at greater than 50 mmHg to prevent cerebral dysfunction. Subsequent studies did not confirm these findings (48; 104; 147).

Carotid artery stenosis. The incidence of hemodynamically significant carotid stenosis is 11% to 20% in patients undergoing cardiac operations (155).

In a retrospective review of 1022 coronary artery bypass surgery patients, the overall rate of perioperative stroke was 2.2%. The rate was increased with severity of extracranial carotid stenosis, from 4.7% in less than 70% stenosis, 5% in 70% to 99% stenosis to 8% in total occlusion (33).

In patients with asymptomatic carotid stenosis undergoing CABG, the risk of in-hospital stroke or death was similar regardless of the degree of carotid stenosis (138).

Summary. Common risk factors are advanced age, carotid artery stenosis, prior heart surgery, prior stroke, peripheral vascular disease, hypertension, diabetes mellitus, aortic atheroma, open heart surgery, and prolonged cardiopulmonary bypass time.

Risk factors associated with left heart catheterization.

Risk factors of CNS complication. Risk factors include female sex, peripheral vascular disease, left ventricular hypertrophy, severe coronary disease, poor left ventricular ejection fraction, concomitant peripheral vascular disease, performance of percutaneous coronary intervention, longer fluoroscopy times, and protruding aortic atheroma (163; 88; 139). Lack of aspirin prior to percutaneous coronary intervention has been associated with increased risk of in-hospital mortality and stroke (78).

Risk factors of peripheral nervous system complication. Thrombocytopenia, female sex, and excessive anticoagulation (aPTT > 150 seconds) are strong predictors of retroperitoneal hemorrhage. However, there is no strong correlation between the hematoma size and the severity of neurologic deficits. Most patients are treated by transfusion alone (an average of 4 units). Urgent surgery is needed if hypotension is refractory to fluid administration (79). Neurologic complications during intra-aortic balloon pump include peripheral nerve injuries and rarely paraplegia related to spinal cord infarction (10).

Differential diagnosis

Associated or underlying disorders

Perioperative neurologic dysfunction following cardiac catheterization and surgery are generally due to surgical and catheter manipulations. However, other nonsurgical factors such as drugs, infection, posttransplantation lymphoproliferative disorders, and metabolic derangements may also play a role, particularly in cardiac transplantation patients.

Diagnostic workup

The key tests used in patients who experience complications following cardiac procedures are:

	• Head CT
	• Brain MRI. There is increasing evidence of safety in patients with cardiac pacemakers and defibrillators.
	• Coagulation profile
	• Blood and fungal cultures
	• Lumbar puncture
	• EEG

CT is the first-line of imaging during acute stroke care. Diffusion-weighted MRI is more sensitive to ischemic lesions that are small or in unsuspected locations. However, MRI is not readily accessible in most emergency departments (23). Concern persists in patients with cardiac pacemakers and defibrillators (13). However, several studies have demonstrated the safety of cerebral MRI in these patients (117; 137; 107). Patients with pulmonary artery catheters that include pacing or thermistor wires should not undergo MRI (115). Coagulation intensity with heparin and coumadin is monitored with APTT and PT/INR, respectively. Fever is an indication for blood cultures and in some cases lumbar puncture.

Electrographic seizures occur infrequently after cardiac surgery and are generally associated with a good prognosis. Hence prophylactic cEEG monitoring is unlikely to be cost-effective in this population (51).

Management

	• Intravenous thrombolysis may be contraindicated in patients who underwent a recent surgical procedure.
	• Intra-arterial thrombolysis has been used successfully in strokes following cardiac catheterization.
	• Mechanical thrombectomy is an option if a large cerebral artery is occluded.
	• Perfusion studies may help selection of patients with large arterial occlusion up to 24 hours from onset of stroke.
	• Permissive hypertension should be used cautiously in unstable cardiac patients.
	• The risk of death or severe cardiovascular morbidity is high, but there is no significant difference between synchronous and staged carotid artery revascularization and coronary artery bypass graft surgery.

Management of peri-procedural stroke. Treatment of periprocedural stroke is similar to that of ischemic stroke. Urgent head CT is needed to exclude intracerebral hemorrhage. Intravenous t-PA can be given within 4.5 hours from last time normal unless contraindicated by coagulopathy or recent surgical procedure. Large vessel occlusion with salvageable penumbra, less likely to respond to t-PA, may be treated, up to 24 hours in selected cases, with endovascular thrombectomy. Cerebral perfusion imaging by CT or MRI may help selection mechanical thrombectomy of large vessel occlusion with significant mismatch between the penumbra and infarct core (99). In a retrospective database of 5022 consecutive patients, 870 underwent endovascular thrombectomy and had excellent outcomes (58).

In acute period, permissive hypertension should be avoided in certain cardiac conditions like pulmonary edema, aortic dissection, and acute myocardial infarction, or when organ injury is suspected.

During cardiac catheterization. There is anecdotal evidence of successful intra-arterial thrombolysis for ischemic stroke during or after cardiac catheterization (81).

After coronary artery bypass surgery. Recent surgery is a contraindication to intravenous t-PA. The is anecdotal evidence of benefit from intra-arterial thrombolysis (77). Permissive hypertension may be used cautiously after surgery to prevent bleeding from an unhealed surgical site.

After cardiac transplantation. Patients waiting for cardiac transplantation require circulatory support from a left ventricular assist device. This device is thrombogenic and requires anticoagulation, a contraindication for intravenous tPA (108). Mechanical thrombectomy can be used for large vessel occlusion. Permissive hypertension should be cautiously used in patients with heart failure.

Management of potential stroke risk factors that are commonly found during preprocedural screening.

Patent foramen ovale. Patent foramen ovale allows an embolus to travel from the venous to the arterial systemic circulation. Larger right-to-left shunts are linked to higher rates of embolism. Patent foramen ovale closure reduces the risk of stroke in selected patients without other risk factors for stroke, which is not the case in most perioperative strokes (103).

Carotid artery stenosis. Carotid endarterectomy (CEA) may benefit carefully selected patients with symptomatic and asymptomatic carotid stenosis (60). As coronary artery disease and carotid stenosis may coexist, optimal timing of CEA in relation to coronary artery bypass graft surgery has been of interest. Carotid duplex ultrasound screening is reasonable before elective coronary artery bypass graft surgery in patients older than 65 years of age and in those with left main coronary stenosis, peripheral arterial disease, a history of cigarette smoking, a history of stroke or transient ischemic attack, or carotid bruit.

A systematic review of 97 studies of combined carotid endarterectomy and coronary artery bypass graft surgery did not find a significant difference between the synchronous and staged procedures. The risk of stroke or major cardiovascular morbidity was 10% to 12% (110). In a metaanalysis of 12 studies including coronary artery bypass graft patients with asymptomatic unilateral carotid stenosis, the lowest risk of stroke or death was observed in synchronous carotid endarterectomy plus off-pump coronary artery bypass graft surgery group (2.2%), whereas the risk of the other groups was 7% to 8% (43).

Carotid artery stenting (CAS) is an alternative to carotid endarterectomy (19). A metaanalysis of 11 studies examined the overall operative risk of cardiovascular events in patients undergoing staged carotid artery stenting plus coronary artery bypass graft surgery. In a cohort of predominantly asymptomatic unilateral carotid disease patients, the 30-day risk of death/any stroke was 9.1% (111).

Additionally, carotid revascularization by endarterectomy or stenting with embolic protection before or concurrent with myocardial revascularization surgery is reasonable in symptomatic patients with stenosis more than 80%. However, safety and efficacy of carotid revascularization in patients with asymptomatic carotid stenosis, even if severe, are not well established (18).

General management of cardiac transplanted patients. Usually, opportunistic CNS infections occur after three weeks from surgery. Post-transplantation lymphoproliferative disorder may be confined to the nervous system or disseminated. The response to treatment is poor (123).

Some antiepileptic drugs may reduce immunosuppressant levels in post-transplantation patients (28). Newer agents are not subject to this interaction.

Outcomes

Out of 100 patients who underwent heart valve procedures, 24 had focal deficits, of which 92% recovered completely at five months (147). Vocal cord dysfunction due to direct trauma or recurrent laryngeal nerve injury can lead to partial laryngeal obstruction and may require reintubation (143).

Cognitive decline after cardiac surgery may affect 7% to 49% of patients at three months and up to 33% after one year (167). Of concern was also late decline in cognitive function due to cardiac procedures. However, this is most likely related to progression of preexisting vascular disease, other neurodegenerative disease, or new neurologic events (102).

Analysis of the Veteran’s Affairs Quality Improvement Program database including 1422 patients found that the medial survival time of patients with stroke after coronary artery bypass graft was 6.7 years. The mortality of patients with stroke compared to those without a stroke within the first 30 days postoperatively was 79% versus 96%, 58 versus 83%, and 36% versus 63% at 1, 5, and 10 years, respectively. Predictors of 1-year mortality were renal failure, prolonged ventilation, coma, and reoperation for bleeding (169).

Special considerations

Anesthesia

The Prevention of Delirium and Complications Associated with Surgical Treatments (PODCAST), an international, multicenter, randomized, controlled trial is ongoing to study effects of subanesthetic dose of intraoperative ketamine on postoperative delirium and other neurologic and psychiatric outcomes in cardiac and noncardiac surgeries. Smaller randomized controlled trials have shown significant reduction in postoperative delirium from 31% to 3%; beneficial effects have been attributed to ketamine’s antiinflammatory and antiexcitotoxic actions (06).

Studies have shown the correlation between low cortical oxygen saturation and cognitive dysfunction, and prolonged hospitalization and perioperative cerebrovascular accidents in patients undergoing cardiopulmonary bypass. A randomized trial has demonstrated higher intraoperative cortical oxygen saturation in the patients undergoing anesthesia with sevoflurane compared to total intravenous anesthesia with midazolam and fentanyl during cardiopulmonary bypass (57).