Neurologic complications of cardiac surgery

INTRODUCTION — Neurologic complications are second only to heart failure as a cause of morbidity and mortality following cardiac surgery, and the presence of neurologic sequelae significantly increases the likelihood of requiring long-term care [1-4].

The neurologic complications of cardiac surgery in adults will be reviewed here. Methods to prevent these complications, issues related to coronary artery bypass grafting (CABG) in patients with known carotid artery disease, and an overview of all early complications following CABG are discussed separately. (See "Coronary artery bypass grafting in patients with cerebrovascular disease" and "Early noncardiac complications of coronary artery bypass graft surgery".)

Most neurologic problems following cardiac surgery can be divided into three categories (table 1) [2,5]:

●Stroke

●Neuropsychiatric abnormalities or encephalopathy

●Peripheral neuropathies

A retrospective report from the Society of Thoracic Surgeons (STS) National Cardiac Database of over 400,000 cardiac surgeries between 1996 and 1997 reported an overall incidence of a new neurologic event (stroke, transient ischemic attack, or unexplained coma lasting more than 24 hours) of 3.3 percent [6]. A prospective study evaluated 2108 patients undergoing coronary artery bypass graft surgery (CABG) at 24 hospitals in the United States between 1991 and 1993 [2]. Overall, 6.1 percent suffered a cerebral complication, roughly equally divided between stroke and encephalopathy. Increased age was a dominant risk factor, and adverse events were associated with increased mortality, longer hospitalization, and a higher rate of discharge to chronic care facilities compared with those without neurologic sequelae (figure 1).

CEREBROVASCULAR DISEASE

Incidence — The incidence of stroke related to cardiac operations ranges from 0.4 to 14 percent in different series, depending upon patient populations and specific procedures [1,2,7-14]. The type and number of procedures that are performed during cardiac surgery may have an impact on stroke incidence:

●Coronary artery bypass graft surgery (CABG) – Reported perioperative stroke rates most commonly fall between 0.8 and 5.2 percent [11,12,14-26]. In a review of a Society of Thoracic Surgeons (STS) database of patients in 2017, 160,160 patients underwent isolated CABG with a reported permanent stroke incidence of 1.4 percent [14].

●Valve repair – Based on the STS database in the calendar year 2017, isolated open aortic valve repairs (AVR) of 25,940 patients and open mitral valve repairs (MVR) of 12,388 patients had permanent stroke risk of 1.3 and 2.3 percent, respectively [14].

●Combined open AVR and CABG procedures – In the same STS database, 15,971 patients underwent combined procedures, which were associated with a 1.9 percent permanent stroke incidence; combined open MVR and CABG procedures were associated with a 3.1 percent stroke incidence [14].

Regardless of the cardiac procedure, the permanent stroke incidence appears to be declining over time in large national databases [14,23-26].

The process for assessing stroke incidence may be a critical factor. In trials, where stroke is diagnosed by neurologists, stroke rates are substantially higher than those reported in surgeon-reported databases like the STS. In the Determining Neurologic Outcomes from Valve Operations (DeNOVO) trial, which included patients undergoing open aortic valve surgeries where neurologists performed neurologic examinations before and after surgery, the stroke incidence was double that of those identified in the same cohort in the STS database [27]. Of the 196 patients in the study, the clinical stroke incidence was 17 percent, whereas only 7 percent of these patients were reported as having a clinical stroke in the STS database.

In addition, the burden of cerebral ischemic injury may be underestimated if only the incidence of clinical stroke is considered. Neuroimaging studies suggest that there is an even higher incidence of so-called silent cerebral ischemia after cardiac surgery [27-29]. In the aforementioned DeNOVO trial, of the 109 patients who did not have a clinical stroke, magnetic resonance imaging (MRI) detected 59 additional silent infarctions [27].

The timing of most perioperative strokes is variously reported. Approximately 30 to 40 percent of strokes occur intraoperatively, according to large prospective reviews [7,23]. Most strokes occur in the first one to two days after surgery and are relatively uncommon after the first week [7,18,23,30]. However, the increased risk of stroke persists for up to two years after cardiac surgery [31].

Risk factors — Patient-specific risk factors for perioperative stroke that have been identified in some studies include [1,2,4,6-12,15-17,23,31-45]:

●Prior stroke or transient ischemic attack

●Significant atherosclerosis of the proximal or ascending aorta, carotid, and/or intracranial cerebral arteries

●Advanced age

●Diabetes

●Renal failure

●Low cardiac output syndrome

●Peripheral artery disease

●Hypertension

●Pre- or postoperative atrial fibrillation

●Female sex

●Recent myocardial infarction or unstable angina

●Moderate or severe left ventricular dysfunction

The history of a clinical stroke is a consistent risk factor for perioperative stroke among studies, with an associated increased odds ratio of 3.2 to 4.5 [2,42]. One study also found that the presence of asymptomatic stroke detected by MRI preoperatively was also a risk factor for perioperative stroke [46]. The time interval between the prior stroke and surgery did not influence perioperative stroke risk, according to one study [44].

Aortic atherosclerosis is also an important risk factor, with an associated odds ratio of 4.5 [2]. In a review of 921 consecutive patients who underwent cardiac surgery, the incidence of stroke was 8.7 percent in those with aortic atherosclerosis, compared with 1.8 percent in those without this abnormality [37]. Among patients with aortic atherosclerosis, the risk of stroke is greatest in those with aortic arch atheromas that are mobile or protrude ≥5 mm into the aortic lumen [38,47]. Large atheromas are more likely to occur at the site of new aortic injury during surgical clamping and cannulation [39]. Large or mobile atheromas are also risk factors for stroke in patients not undergoing surgery. (See "Thromboembolism from aortic plaque".)

The data associating carotid disease with perioperative stroke in patients are more limited. This is discussed in detail separately. (See "Coronary artery bypass grafting in patients with cerebrovascular disease".)

The presence of any one risk factor for stroke is not considered a contraindication for cardiac surgery [42]. A number of risk models have been developed to predict stroke after cardiac surgery (usually CABG) using preoperative risk factors. While these models vary substantively, they emphasize the additive effect that individual risk factors have on perioperative stroke risk [15-18].

Operative factors are also important [2,4]. Closed chamber procedures, such as CABG, have a lower risk than open chamber procedures, such as valve replacement [48,49]. This difference was demonstrated in a multicenter prospective trial of 273 patients in which the rate of cerebral complications (predominantly stroke) was 6 percent in patients undergoing CABG compared with 16 percent in patients undergoing intracardiac surgery [49]. Combined or complex surgical procedures, including those that are prolonged and involve manipulation of the ascending aorta, increase the risk of neurologic complications [4,6]. Increasing the temperature of cardiopulmonary bypass (CPB) may also augment the risk of stroke [9].

Mechanisms of cerebral infarction — Different mechanisms are thought to be responsible for early onset or intraoperative ischemic stroke versus delayed ischemic stroke [30,50].

Intraoperative stroke — Cerebral hypoperfusion and atheroembolization are understood to be the major mechanisms underlying intraoperative stroke:

●Cerebral hypoperfusion can result from intraoperative hypotension and/or diminished cardiac output. Furthermore, it is hypothesized that decreased blood flow during surgery results in diminished washout of embolic material from blood vessels in the brain, particularly in borderzone (watershed) areas along the boundaries of major vascular territories, predisposing to ischemia in these regions [51]. These mechanisms may combine to increase the cumulative risk of injury in cardiac surgery and explain the presence of postoperative strokes identified as having multiple mechanisms [18].

While mean arterial pressures of 50 to 70 mmHg are common and typically well tolerated by patients during CPB, higher mean arterial blood pressure may decrease both cardiac and neurologic complications [52]. This and other techniques to avoid intraoperative cerebral hypoperfusion are discussed in detail separately. (See "Management of cardiopulmonary bypass", section on 'Mean arterial pressure' and "Management of special populations during cardiac surgery with cardiopulmonary bypass", section on 'Cerebrovascular disease'.)

●Arterial emboli can cause transient or permanent occlusion of cerebral vessels and produce cerebral ischemia. This appears to be the predominant cause of intraoperative stroke, as is supported by the finding that most patients with stroke have multiple cerebral infarcts in different arterial territories on neuroimaging studies [30,50].

Three major types of emboli occur during cardiac operations: thromboemboli, atheroemboli, and air emboli [53]. Thrombi or atheromatous debris can be released from complex aortic plaques during clamping and unclamping of the ascending aorta, during the construction of the proximal CABG anastomoses in the ascending aorta, while excising severely calcified and diseased cardiac valves, or by turbulent high-velocity blood flow from the aortic cannula within a diseased aorta. Transcranial Doppler (TCD) monitoring of the middle cerebral artery blood flow detects arterial microemboli entering the cerebral circulation during cardiac surgery; these are most frequent during placement and removal of the aortic cross-clamp and the aortic cannula [54,55]. (See "Initiation of cardiopulmonary bypass", section on 'Aortic cannulation'.)

Gaseous emboli enter the arterial circulation via open cardiac chambers, vascular cannulation sites, or arterial anastomoses. These may be the most common type of embolic material seen during CPB; however, the impact that these have on neurologic injury and outcome is unclear [56]. The size of the gaseous emboli may be important, with larger emboli-causing strokes and smaller emboli resulting in endothelial injury in the brain's vasculature. (See "Thromboembolism from aortic plaque", section on 'Cardiovascular procedures' and "Embolism from atherosclerotic plaque: Atheroembolism (cholesterol crystal embolism)" and "Air embolism".)

Postoperative stroke — Stroke that is noted after an initial uneventful neurologic recovery from surgery is most likely to be due to cardiogenic embolism [50]. Atrial fibrillation, a common postoperative complication, is believed to contribute to a significant proportion of delayed postoperative strokes [18,57]. Underlying heart disease also likely plays an important role. As in other clinical settings, the CHADS₂ score helps stratify the risk for postoperative stroke [31]. (See "Atrial fibrillation in adults: Use of oral anticoagulants".)

Intracranial hemorrhage — Primary intracranial hemorrhage in patients is a rare complication of cardiac surgery; most intracranial hemorrhage in this setting occurs as a complication of ischemic infarction.

Although 5 percent of patients with perioperative ischemic strokes have some hemorrhagic transformation [58], the incidence of sizable hemorrhages with mass effect is much smaller. The risk factors for significant hemorrhage include large infarcts, anticoagulation, or the presence of an infarct that is already hemorrhagic [59]. In any patient with prior intracranial hemorrhage undergoing a cardiac operation, it is imperative to perform a neurologic examination as soon as possible postoperatively to determine whether there is clinical evidence of increasing hemorrhage that might require neurosurgical intervention.

Patients with endocarditis often have ongoing neurologic injury, which puts them at greater risk of intracranial hemorrhage during anticoagulation for CPB [60]. This risk may be reduced by delaying the operation for two to three weeks while antibiotic therapy is initiated; however, hemodynamic instability may preclude delay [60]. Another method to reduce bleeding risk is to perform cardiac surgery with a heparin-bonded CPB circuit, which permits reduced systemic heparinization [61,62].

Ophthalmologic complications — Arterial emboli can cause symptomatic ophthalmologic complications. Intraoperative fluorescein angiography demonstrated transient microvascular occlusion in all 21 patients studied in one report [63]. In a larger series of 312 patients undergoing CABG, 26 percent experienced neuroophthalmologic sequelae, including 17 percent with retinal infarctions [64]. Visual loss due to ischemic optic neuropathy is uncommon, occurring in approximately 0.1 percent of patients undergoing CPB [65,66].

A supranuclear gaze palsy is another described complication of cardiac surgery, particularly aortic valve replacement and repair of an ascending aortic dissection [67-70]. This is probably under-recognized as the patient's complaints are often somewhat vague, and demonstration of the neurologic deficit requires specific examination of vertical and horizontal saccades. Spastic dysarthria, emotional lability, and gait disorder are frequent accompanying abnormalities. The presumed mechanism is ischemic injury to the midbrain, or elsewhere in the brainstem, but this has not been specifically documented.

Diagnosis and treatment — In patients stable enough to undergo brain MRI, diffusion-weighted imaging (DWI) is very sensitive and accurate for the diagnosis of acute ischemic events [18]. In addition to confirming the presence of a cerebral infarction, it is useful to identify additional, otherwise unsuspected areas of infarction and may hint at the underlying mechanism, that is, embolism versus watershed. (See "Neuroimaging of acute stroke".)

It is also important to confirm the timing of the stroke. A stroke that occurs postoperatively suggests the potential for recurrent events due to ongoing risk factors, such as atrial fibrillation, cerebrovascular disease, intramural thrombus in the heart, or inadequate anticoagulation for a mechanical valve prosthesis. An intraoperative stroke is less likely to recur.

Potential interventions for the treatment of stroke after cardiac surgery are discussed in detail separately but include:

●The management of blood pressure in acute stroke has not been well studied. In other settings (not postoperative), blood pressure lowering is not recommended unless in excess of 220 mmHg systolic or 120 mmHg diastolic [71], or unless the patient is suffering from, or at risk of, aortic dissection, acute myocardial infarction, heart failure, or other adverse effects. However, after cardiac surgery, surgeons may recommend somewhat tighter blood pressure control. Hypotension should be avoided and treated if it occurs using supine posture and an intravenous fluid bolus. If these measures are insufficient, a pressor such as phenylephrine at low dose may be useful. Although there is no support in the literature for blood pressure augmentation in normotensive patients, it is our general practice to treat these patients with judicious use of volume expansion to improve brain perfusion. (See "Initial assessment and management of acute stroke", section on 'Blood pressure management'.)

●Patients with hypoxia should receive supplemental oxygen to bring saturations to levels ≥92 percent [71]. (See "Initial assessment and management of acute stroke", section on 'Airway, breathing, and circulation'.)

●Hyperglycemia after stroke has been associated with worse outcomes; treatment may improve outcome. (See "Initial assessment and management of acute stroke", section on 'Hyperglycemia'.)

●Fever has also been associated with worse outcomes in the acute stroke setting; antipyretics are generally recommended to lower fever in patients after stroke. (See "Initial assessment and management of acute stroke", section on 'Fever'.)

●Use of intravenous tissue plasminogen activator is associated with significant risk of bleeding in the postoperative period and is contraindicated in patients after cardiac surgery. Other reperfusion techniques may be considered in centers with requisite expertise and for specific cases where therapy is instituted within a 24-hour time window [72]. This may be particularly relevant for patients who have undergone cardiac surgery and have suffered a perioperative stroke [73]. (See "Approach to reperfusion therapy for acute ischemic stroke".)

●Anticoagulation has not been shown to be effective in acute stroke treatment, regardless of etiology. In patients with acute stroke following cardiac surgery, it carries the additional risks of systemic complications such as hemopericardium. (See "Early antithrombotic treatment of acute ischemic stroke and transient ischemic attack", section on 'Limited role of early anticoagulation'.)

●The use of antiplatelet therapy following CABG has been shown to be safe and effective in decreasing neurologic complications [74], and has also been shown to improve outcomes following acute stroke [75]. Aspirin or clopidogrel is recommended for all patients following cardiac surgery except those with absolute contraindications. (See "Coronary artery bypass surgery: Perioperative medical management", section on 'Aspirin' and "Coronary artery bypass surgery: Perioperative medical management", section on 'Platelet P2Y12 receptor blocker therapy'.)

●Statin therapy is usually recommended in patients following CABG for secondary prevention of cardiovascular as well as cerebrovascular events. (See "Coronary artery bypass surgery: Perioperative medical management", section on 'Statins'.)

●Other therapies, such as cerebral protection or hyperbaric oxygen, may be of value in the appropriate circumstances [76,77]. Hyperbaric oxygenation is of special value in otherwise stable patients with suspected air emboli. (See "Hyperbaric oxygen therapy".)

Outcome after stroke — Patients who have a perioperative stroke have a poorer outcome compared with those without a stroke [12,18]. As an example, 35,733 consecutive patients were followed after CABG; 1.6 percent had a perioperative stroke. Patients with stroke had 1-, 5-, and 10-year survival rates of 83, 59, and 27 percent, respectively, representing a threefold increased risk of death compared with patients without stroke [78]. Similar findings were noted in other reports [10,18,79].

The biggest differences in mortality are seen early and stabilize over time [11,78]. In-hospital mortality rates for stroke after CABG range from 14 to 30 percent [11,12,16] and appear to be higher in patients with intraoperative versus postoperative stroke (41 versus 13 percent) [7].

Patients who have had a stroke are also likely to have functional deficits limiting independence and to be discharged to locations other than home [18].

ENCEPHALOPATHY — The major neuropsychiatric abnormalities associated with cardiac surgery are neurocognitive dysfunction, seizures, and delirium. At times, coma is seen. Rather than distinct entities, these represent a spectrum of neurologic dysfunction with overlapping clinical features and pathogeneses. Premorbid or perioperative stroke and other vascular mechanisms may also play a role in some patients with postoperative encephalopathy.

Delirium — Delirium after cardiac surgery occurs in 3 to 32 percent of patients, particularly those with preexisting organic mental disorders (eg, stroke, dementia), significant prior alcohol consumption, advanced age, or intracranial cerebral artery disease [2,36,80-83]. The wide incidence range indicates the difficulty of attributing delirium specifically to the surgical procedure, rather than to opiates, anesthetics, and sedative agents administered during and immediately following surgery. Other causes of postoperative delirium in the cardiac surgery patient include renal failure, hepatic failure, and thyroid abnormalities. Strokes, particularly right parietal lesions, can also present as delirium.

Treatment of postoperative delirium requires a search for, and correction of, potentially reversible causes. The electroencephalography (EEG) is usually abnormal in postoperative delirium, but not in a primary psychiatric illness. (See "Acute toxic-metabolic encephalopathy in adults" and "Diagnosis of delirium and confusional states" and "Delirium and acute confusional states: Prevention, treatment, and prognosis".)

Delirium may persist for more than one week and overlaps with more persistent postoperative neurocognitive dysfunction (see 'Cognitive impairment' below). In one prospective series of 225 patients after cardiac surgery, those who experienced postoperative delirium were more likely than those who did not to have persistent cognitive decline over baseline (40 versus 24 percent, p = 0.001 at six months; and 31 versus 21 percent, p = 0.055 at 12 months) [84]. Another prospective study found that delirium after CABG was a risk factor for late mortality, particularly in younger patients (<65 years old), in whom the associated hazard ratio was 2.4 [85]. (See "Delirium and acute confusional states: Prevention, treatment, and prognosis", section on 'Outcomes'.)

Seizures — Approximately 0.5 to 3.5 percent of patients experience seizures after coronary artery bypass graft surgery (CABG) [2,36]. Causes include hypoxemia, metabolic disturbances (eg, hyponatremia, hypoglycemia), drug toxicity (eg, lidocaine, procainamide), and structural brain injury such as stroke.

EEG should be considered in patients who are unresponsive 18 to 24 hours after surgery to detect possible nonconvulsive seizure activity. (See "Convulsive status epilepticus in adults: Classification, clinical features, and diagnosis".)

Coma — Coma infrequently complicates cardiac surgery and results from diffuse global anoxic injury, strokes (large hemispheric, brainstem, or multifocal infarcts), intracerebral hemorrhages, seizures, or severe metabolic derangements.

Coma as a result of anoxic or structural injury carries an extremely poor prognosis. Patients with severe anoxic injury and coma may also exhibit postanoxic myoclonus and seizures.

Cognitive impairment — Disturbances in memory, executive function, motor speed, attention, and other cognitive functions can be detected in 3 to 79 percent of patients in the first several weeks after CABG when formal neuropsychiatric testing is performed [2,4,86-91]. The severity can range from subtle deficits detectable only by sophisticated neuropsychiatric testing to clinically overt delirium as described above.

The wide variation in the reported incidence of postoperative cognitive problems is probably related to several factors, including variability in CABG procedures, constraints reducing participation in cognitive studies, different methodologies used to assess neurocognitive dysfunction, variations in the follow-up time interval, and a paucity of studies with control groups [4,18,91,92].

The importance of control groups has been illustrated in two studies. One compared neurocognitive function at baseline, 3 months, and 12 months after CABG using two different control groups: those with coronary disease but no surgery and those who were heart healthy [93]. At baseline, patients with heart disease had lower cognitive scores than the heart-healthy patients. At baseline and at follow-up, the cognitive test performance of CABG patients did not differ from control groups with heart disease.

Mechanisms and risk factors — While this syndrome has been called "postperfusion syndrome," "postpump syndrome," and "pump-head," these terms imply a mechanism of action that is speculative and that may not be a factor in some patients with neurocognitive dysfunction after cardiac surgery. The incidence of neuropsychiatric deficits in patients undergoing cardiac operations exceeds that of corresponding patients undergoing operations for peripheral vascular disease, suggesting that some specific features of cardiac operations, such as cerebral microemboli, may cause neuropsychiatric disturbances [55,83,86].

Microemboli appear to be generated by manipulation of the heart and aorta, particularly during aortic cannulation and clamping and cardiotomy suctioning. Clinical investigations provide mixed support for a role of microemboli in post-CABG encephalopathy. Transcranial Doppler (TCD) monitoring studies and neuropathologic studies find that platelet microemboli and ischemic lesions are common during and after cardiac surgery [94]. In one series of 40 patients undergoing intracardiac surgery, comparison of preoperative and postoperative magnetic resonance imaging (MRI) scans and neuropsychologic testing found that postoperative decline in cognitive function was associated with the presence and severity of new ischemic lesions [95]. However, other studies have not confirmed this association [29]. Similarly, studies have not shown a causal association between the presence or number of microemboli detected during surgery and the postoperative cognitive impairment or pathologic change [83,96].

Intraoperative hypotension and prolonged oxygen desaturations may also contribute to cerebral injury and subsequent neurocognitive decline [83]. In one small study a drop in mean arterial blood pressure of 27 mmHg or greater (compared with preoperative baseline) was associated with a decrease in the Mini-Mental State Examination score of 1.4 points [97].

Other identified risk factors for cognitive impairment after cardiac surgery may be identified preoperatively and overlap with those for stroke, further implicating an underlying vascular mechanism [2,4,16,18,46]:

●Hypertension

●Carotid disease

●Advanced age

●Previous stroke (clinical event or detected on MRI)

●Underlying pulmonary disease

Temperature is a modulator of experimental cerebral injury, and postoperative hyperthermia may be related to the development of cognitive dysfunction. In one study of 300 patients, the maximum postoperative temperature was associated with a greater amount of dysfunction at six weeks [98]. This suggests that interventions to avoid postoperative hyperthermia may be beneficial for improving cerebral outcome after CABG.

Other mechanisms implicated in post-CABG cognitive decline are more speculative and include systemic inflammation and exposure to general anesthesia [83].

On-pump CABG does not appear to be associated with a higher risk for cognitive or neuropsychologic impairment compared with off-pump CABG according to a meta-analysis and subsequently performed clinical trial. (See "Off-pump and minimally invasive direct coronary artery bypass graft surgery: Clinical use", section on 'Neurologic dysfunction'.)

Clinical course — Neurocognitive deficits usually resolve gradually. Most patients return to their preoperative baseline by 3 to 12 months after surgery [93]. In one study, long-term recovery was more likely in patients with less severe impairments after surgery, as well as in those with more education and greater activities of daily living at six weeks [99].

However, postoperative encephalopathy is associated with poor short-term outcomes. In one series, the length of stay was 14 days for those with encephalopathy, versus an average of eight days for patients without this complication [55]. These patients also had a three times greater in-hospital mortality rate and were less likely to be discharged to home.

Late cognitive decline — Some believe that post-CABG patients are at higher risk for long-term cognitive decline. In one study, serial neurocognitive testing was performed in 261 patients after CABG [89]. A decline in neurocognitive function was defined as a drop of ≥1 standard deviation (SD), representing a decline of approximately 20 percent. The incidence of neurocognitive decline at discharge, six weeks, six months, and five years was 53, 36, 24, and 42 percent, respectively. The main predictor of cognitive decline at five years was neurocognitive deficits at discharge.

However, accumulating evidence suggests that the cause of late decline in cognitive function occurring one to five years after CABG is progression of underlying cerebrovascular disease or vascular dementia rather than CABG or cardiopulmonary bypass (CPB) [4,91]. A number of longitudinal studies have compared cognitive changes among post-CABG patients with a control group of patients with coronary disease with no surgery, revealing similar rates of cognitive decline [93,100-103]. A case-control study comparing 557 patients with incident dementia with nondemented controls found similar rates of CABG history in both groups [104]. Some studies suggest that cardiac disease (particularly low cardiac output and left ventricular diastolic dysfunction) in the absence of cardiac surgery is associated with cognitive impairments [105].

PERIPHERAL NEUROPATHY

Upper extremity — Upper extremity peripheral nerve injury follows cardiac operations in 2 to 15 percent of patients [8,48,106-108]. Putative mechanisms of injury include brachial plexus traction, brachial plexus compression between the clavicle and the first rib during sternal retraction, and nerve injury during internal mammary artery dissection, hypothermia, and hemodynamic changes during cardiopulmonary bypass (CPB). In one case series, hypertension, tobacco use, and diabetes were risk factors for nerve injuries [108].

Peripheral neuropathy typically presents with numbness, weakness, pain, diminished reflexes, and discoordination of one upper extremity. The presence of pain suggests a peripheral rather than central injury, while confusion, cranial nerve involvement, or hemiparesis suggests a central injury.

Sensory symptoms are most frequent in the fourth and fifth fingers. Although this distribution is expected with ulnar nerve injury at the elbow, similar symptoms are associated with injury to the lower (medial) brachial plexus elements. Overall, brachial plexus injury is more common than isolated ulnar nerve deficits after cardiac surgery.

Weakness in muscles supplied by the lower cervical nerve roots but not the ulnar nerve, such as the extensor indicis and abductor pollicis brevis, indicates a plexopathy. Following intraaortic balloon pump insertion, femoral nerve injury may occur due to local trauma, vascular occlusion, pseudoaneurysm formation, or emboli.

In most neuropathies related to cardiac surgery, symptoms improve or resolve within three weeks, suggesting a neurapraxic injury caused by disruption of the myelin sheath. Given this good prognosis, conservative treatment, consisting of physical therapy to improve strength and flexibility in the affected muscles, is indicated.

Axonal injury is much less common and is associated with neurologic symptoms lasting for an extended period. Thus, failure to improve within three weeks following cardiac surgery should prompt electromyography and nerve conduction studies to confirm the diagnosis, delineate the site of injury and the extent of nerve disruption, and provide an accurate prognosis for recovery.

Phrenic nerve — Phrenic nerve injury occurs with cooling of the heart by iced slush. The reported frequency with which it occurs has varied widely from 1 to 30 percent, depending in part upon whether it is diagnosed by radiographic evidence of diaphragmatic dysfunction, which is not always due to phrenic nerve dysfunction [109], or phrenic nerve conduction studies [107,109-111].

Most patients have unilateral phrenic nerve injury, typically on the same side as the internal mammary graft with coronary artery bypass graft surgery (CABG) [110,111]. Patients with bilateral phrenic nerve dysfunction generally require prolonged mechanical ventilation [111]. Most affected patients recover fully within one year, but recovery may be delayed for two years or more [110,112].

The frequency of phrenic nerve injury has declined dramatically with changes in surgical technique. These include a reduction in the use of iced slush and an increase in the use of foam insulation [111].

Intercostal nerve — Harvesting of the internal thoracic artery may be associated with injury to the anterior intercostal nerves. This nerve injury can present with numbness, tenderness, light-touch evoked pain, or constant burning pain over the sternum and left anterolateral chest wall. In one study of 37 patients, for example, 81 percent reported protracted postoperative symptoms; although this usually subsided by four months, 15 percent had symptoms that persisted for 5 to 28 months after surgery [113].

OTHERS — A dyskinetic movement disorder, so-called post-pump chorea may complicate cardiac surgery in a small number of children with congenital heart disease. This has also been described in a single case series of five adult patients who underwent pulmonary endarterectomy with coronary artery bypass graft surgery (CABG) and hypothermic circulatory arrest [114]. Symptoms presented between one and three days postoperatively and subsequently remitted over several weeks to months. Two patients had persistent mild symptoms. (See "Hyperkinetic movement disorders in children", section on 'Post-pump chorea'.)

SUMMARY AND RECOMMENDATIONS — Neurologic complications of cardiac surgery are not rare; there is an estimated 6 percent incidence of combined central nervous system complications (stroke, encephalopathy).

●Cerebrovascular disease – Most strokes that occur in the setting of cardiac surgery are clinically apparent within 24 to 48 hours. The incidence is uncertain; surgical databases may substantially underestimate the incidence of stroke when compared with data collected in rigorous clinical or imaging-based studies. (See 'Incidence' above.)

Valvular repairs and combined coronary artery bypass graft surgery (CABG) and valvular repairs, particularly those involving the mitral valve, appear to have a significantly higher risk of stroke than CABG alone.

Mechanisms of intraoperative and postoperative ischemic stroke include cerebral hypoperfusion, artery-to-artery embolism, and cardiogenic embolism. (See 'Mechanisms of cerebral infarction' above.)

Traditional atherosclerosis risk factors, prior cerebrovascular disease, and atherosclerotic disease affecting the coronary or peripheral arteries are among the risk factors for perioperative ischemic stroke. (See 'Risk factors' above.)

Management of perioperative ischemic stroke in the setting of cardiac surgery is similar to that in other settings except that use of intravenous tissue plasminogen activator is contraindicated in patients after cardiac surgery. (See 'Diagnosis and treatment' above and "Initial assessment and management of acute stroke".)

●Encephalopathy – More diffuse cerebral involvement after cardiac surgery may manifest as a delirium, seizures, and sometimes even coma. The mechanisms underlying this complication are likely multifactorial. A minority of patients will have persistent neurocognitive dysfunction. Evaluation in these patients focuses on excluding a cerebrovascular event and reversible toxic and metabolic conditions. (See 'Encephalopathy' above.)

●Peripheral nerve complications – Peripheral nerve injury is an uncommon complication. Specific syndromes include:

•Brachial plexopathy with motor-sensory deficits in the upper extremity

•Phrenic neuropathy with prolonged ventilator dependence

•Intercostal neuropathy with pain and dysesthesia in a localized distribution in the lateral chest wall

Most patients improve, although recovery can sometimes take months to occur. (See 'Peripheral neuropathy' above.)