Clinical manifestations

Presentation and course

The term cerebral anoxia indicates any form of inadequate oxygen delivery to the brain, including hypoxemia and ischemia. Generalized brain anoxia is most commonly a consequence of systemic circulatory arrest caused by cardiac arrhythmia. Unlike other organs, such as the kidney and heart, which can tolerate ischemic periods of up to 30 minutes, the brain cannot tolerate more than a few minutes of anoxia. The degree of anoxic brain injury ranges from very mild to very severe or even fatal due to elevated intracranial pressure.

Brief episodes of cerebral anoxia are usually well tolerated, with patients escaping any irreversible deficits. Yet, an amnestic syndrome may follow transient periods of global ischemia. Patients may experience a severe antegrade amnesia and variable retrograde memory loss (131). Individuals with anoxic-ischemic coma of more than six hours duration, but with unremarkable MRI and CT imaging, have demonstrated persistent poor learning and recall of paired associations when compared with age- and IQ-matched controls (87). After apparently recovering from the immediate effects of an anoxic insult, rare patients can also lapse back into unconsciousness. Such delayed severe neurologic deterioration occurs in approximately one to two individuals per 1000 arrests and is not predictable by the type of insult, duration of anoxia, or any other identifiable variable (117). Furthermore, secondary neurologic injury after the initial anoxic insult oftentimes leads to neurologic worsening after initial brief awakening or recovery.

Severe or prolonged periods of hypotension can result in global or focal cerebral lesions. Individuals usually remain in coma for 12 hours or more and on awakening experience deficits including partial or complete cortical blindness, bibrachial paresis, and quadriparesis. Cortical blindness is rarely permanent in nature and is a result of ischemia to either occipital pole, which is located in an arterial border zone. Bilateral infarction of the cerebral motor cortex in the border zone between the anterior and middle cerebral arteries appears to be responsible for the syndrome of bibrachial paresis, sparing the face and lower extremities following cardiac arrest.

Movement disorders, such as myoclonic jerks and cerebellar ataxia, may also follow episodes of cardiac arrest. The "action myoclonus" syndrome can occur following global cerebral ischemia. The myoclonic jerks frequently are stimulus-activated by light, sound, or initiation of movement and can incapacitate individuals with their daily living activities. Therapy with various agents, such as serotoninergic agents, clonazepam, and valproic acid has been used with some success (118). Both levomepromazine and intrathecal baclofen have been used to control posthypoxic movement disorders and hyperkinetic storms, with promising and prompt responses (135). Cerebellar ataxia, involving the trunk or extremities, is an infrequent postanoxic syndrome and may be secondary to the selective vulnerability of Purkinje cells to anoxia. Although rare, several movement disorders may arise as a consequence of hypoxic injury including dystonia, akinetic-rigid syndromes, tremors, and chorea, implicating various parts of the central nervous system, including the basal ganglia, thalamus, midbrain, and cerebellum (145). Extrapyramidal tract dysfunction following anoxia from cardiac arrest can yield a clinical syndrome identical to parkinsonism. Parkinsonian features may represent only a small part of widespread cerebral injury, but in other instances, the clinical presentation of rigidity, akinesia, and tremor may represent the only neurologic disability.

Delirium is also a common feature after cardiac arrest. Most patients may have delirious symptoms during their first weeks after resuscitation. The most common type of delirium among this population is hypoactive delirium (92). It is very important to identify patients with higher risk of delirium (advanced age, intensive smoking, daily alcohol intake, pre-existent cognitive impairment, and preceding periods of sedation, coma, or mechanical ventilation) because it is associated with longer ICU stay, hospital admission, or discharge to rehabilitation center or nursing home, in addition to poor neurologic recovery (62).

The most devastated group of survivors following cardiac arrest suffer widespread destruction of the cerebral cortex and progress to either a persistent vegetative state or death from neurologic complications.

In general, the spinal cord is considered to be more resistant to ischemic insults than rostral sections of the central nervous system. Rarely, however, cases of isolated spinal cord infarction can occur without evidence of cerebral injury. Neurons in the lumbar or lumbosacral regions of the spinal cord appear to possess a greater vulnerability during hypotensive events (31). Syndromes of spinal cord ischemia following transient hypotension are characterized by flaccid paralysis of the lower extremities, urinary retention, and a sensory level in the thoracic region with the anterior spinal-thalamic tracts usually more affected than the posterior columns.

Prognosis and complications

Although hospital mortality after cardiac arrest in the United States has decreased from 66% in 2006 to 56% in 2012 (112), the current overall survival rate is approximately 10.4% (48). Neurologic injury remains the leading cause of death in this patient population and accounts for approximately two thirds of all-cause mortality (70).

Preexisting medical conditions, such as coronary artery disease, hypertension, and diabetes mellitus, may influence an individual’s outcome from cardiac arrest, but an increased burden of pre-existing medical comorbidity has only been inconsistently associated with unfavorable outcome. In a study of 314 patients treated with targeted temperature management (TTM) after their arrest, the pre-existing medical comorbidity as determined by the Charlson Comorbidity Index, was not associated with neurologic outcomes (102). In the era of TTM, the most commonly identified prognostic indicators in comatose survivors of cardiac arrest include younger patient age, initial rhythm and cardiogenic etiology, receipt of bystander cardiopulmonary resuscitation (CPR), short time from collapse to advanced cardiac life support initiation, and time from arrest to return of spontaneous circulation (ROSC). Navab and colleagues found bystander CPR to be the most effective predicting factor for the success rate of return of spontaneous circulation and survival to hospital discharge (SHD) (101). Because hemoglobin is the main determinant of tissue oxygen delivery, anemia can compromise cerebral oxygenation after an acute brain injury. Low hemoglobin value (Hb less than 9.9 g/dL) and Hb oxygen venous saturation less than 60% are significantly associated with unfavorable neurologic outcome in adult patients resuscitated from cardiac arrest (158). Additionally, higher hemoglobin level and absence of diabetes were found to be independently associated with good cerebral performance category at six months (107). On the contrary, higher blood glucose levels at admission and during the first 36 hours and higher glycemic variability were associated with poor neurologic outcome and death in cardiac arrest patients treated with TTM (08). Interestingly, higher and continuous requirement for neuromuscular blockade during TTM was associated with improved neurologic outcome and decreased in-hospital mortality (71). Not surprisingly, the presence of impaired functional cardiac performance with associated structural heart disease can result in an unfavorable outcome (100). Individuals with the apoE 3/3 genotype appear to experience a more favorable outcome following cardiopulmonary resuscitation (136). Pressor requirement in the intensive care unit after arrest and resuscitation was another factor significantly linked to prognosis; of 16% overall survivors in one study, the group requiring pressors was only half as likely to survive to discharge and to be discharged home than patients not taking pressors (141). A significant association of mortality rates after cardiac arrest has also been found for lactate clearance, but this does not pertain to neurologic outcome (91). In a study of 352 patients of whom 249 had lactate measured three hours post-arrest period, patients who required vasopressors had higher mortality, with a stepwise increase in mortality associated with increasing lactate. A combination of lactate and vasopressors in the immediate post-arrest period might be helpful to predict mortality (22). Furthermore, severe vitamin D deficiency with 25(OH)D less than 10ng/ML was found to be strongly associated with unfavorable neurologic outcome and mortality in sudden cardiac arrest patients (13).

Of all tools used for prognostication of neurologic outcome after cardiac arrest, the neurologic examination remains the centerpiece. In general, delay in recovery of neurologic function is associated with a worse prognosis. Individuals with the best chance of recovery have preserved brainstem function following the initial insult. Early onset of incomprehensible speech, orienting spontaneous eye movements, or the ability to follow commands are considered indicative of a good prognosis (75). Traditionally, the most reliable aspects of the examination for predicting poor outcome have been absent pupillary and corneal reflexes as well as extensor posturing or no movement to noxious stimulation (75). However, a poor motor response (extensor or no movement to noxious stimulation) has been drawn into question, both with and without the use of therapeutic hypothermia, because of unacceptable false-positive rates (121; 39). The motor response is also confounded by use of neuromuscular blockade. Similarly, absent pupillary response to light and absent corneal reflexes at 72 hours after arrest or rewarming constitute a poor sign if absent but are flawed by low sensitivity (128). There is an increasing number of reports of patients with delayed awakening with good outcome, despite traditional poor prognostic features (39; 42).

Somatosensory evoked potentials (SSEP) determine the functional integrity of the spinal cord posterior columns, brainstem medial lemniscus, thalamus, and frontoparietal sensorimotor cortex. Bilateral absence of cortical peaks on SSEPs has been regarded as one of the most reliable tools for early prediction of nonawakening in nontraumatic comatose patients with hypoxic brain injury (09; 35; 121; 125) and is recommended by national and international guidelines (155; 127). A reduction in N20 amplitude greater than 53% predicts poor outcome at 72 hours post cardiac arrest (134).

Patients who maintain normal responses throughout their illness maintain a good prognosis but may have permanent neurologic sequelae. Adding somatosensory evoked potential results to the context of prior knowledge (demographic and clinical information) changed the predicted probability of withdrawal of life-sustaining therapy and survival to discharge in comatose post-cardiac arrest patients (99). SSEPs, however, can be confounded: hypothermia (core and limb low temperatures) has been shown to prolong latencies of cortical potentials in an intact neural pathway by decreasing conduction velocities and increasing synaptic transmission delay because of impaired neurotransmitter release (69). SSEPs may also be affected by sedatives (143); hence, their reliability has been questioned, specifically after therapeutic hypothermia (04; 30). In a small cohort, though, the positive predictive value of SSEP in predicting nonawakening from hypoxic-ischemic encephalopathy in patients treated with therapeutic hypothermia was 100% (84). In an effort to avoid some of the challenges discriminating between low-amplitude and absence of N20 SSPE, measurement of P25 and P30 (positive peaks with latencies between 25 and 30 ms following N20) may help to distinguish odds of poor outcome with higher sensitivity (65).

Brainstem auditory evoked potentials correlate with brainstem dysfunction during coma. Simultaneous latency increase of all components is consistent with progressive ischemia of the posterior fossa and a decrease in cerebral perfusion pressure. At least one of the brainstem auditory evoked potentials is absent bilaterally in half of survivors after cardiac arrest, but use for prediction of poor neurologic outcome remains limited as brainstem auditory evoked potentials can be altered by hypothermia, anesthetics, and barbiturates (26).

Although both absent SSEP responses and exam findings of absent pupillary or corneal reflexes at 72 hours after cardiac arrest are regarded as reliable predictors of poor outcome and, hence, are included in current guidelines, these measures have a low sensitivity for detection of an eventual poor outcome: only about 20% of patients without chances of recovery can be identified (21). Cumulative evidence of EEG findings after cardiac arrest and its value for prognostication has been recognized. Furthermore, a multimodal approach based on bilateral absent/absent-pathological amplitude cortical SSEP patterns, gray matter/white matter (GM/WM) ratio <1.21 on brain CT, and isoelectric/burst-suppression EEG patterns helps to accurately predict poor neurologic outcome at six months within the first 24 hours after cardiac arrest (157).

Traditionally, presence of myoclonic status epilepticus, suggestive of diffuse neocortical damage, was felt to be a strong predictor for poor outcome (155), but multiple reports have surfaced of patients with good outcomes despite myoclonic status epilepticus, with or without treatment with therapeutic hypothermia (04). Three distinct clinical semeiologies of postanoxic myoclonic status were identified: type 1 = distal, asynchronous, variable; type 2 = axial or axial and distal, asynchronous, variable; and type 3 = axial, synchronous, stereotyped (96). Their implications on prognostication are subject of ongoing research. Studies revealed that the combined variables of EEG continuity and absence of anoxic changes on MRI were associated with coma recovery at hospital discharge in patient with postanoxic myoclonus (07). The overall most promising feature on EEG for predicting poor outcome is the absence of reactivity to external stimulation (121). Interestingly, when adding EMG to assess reactivity of EEG seems to reduce false negative predictions for identifying patients with favorable outcome. (12). If continuous EEG activity was restored within eight hours, it was associated with good outcome. The maximum sensitivity for EEG to predict outcome is in the first 24 hours after cardiac arrest. Continuous EEG patterns at 12 hours after event are associated with good recovery (123). Patients with a poor neurologic outcome showed slower or no recovery of physiological EEG rhythms. Furthermore, EEG findings indicative of poor outcome include generalized suppression, burst suppression, or generalized periodic complexes (20). However, treating rhythmic or periodic EEG activity with antiseizure medications did not significantly improve neurologic outcome at 3 months in comatose survivors of cardiac arrest compared to standard care alone (122). Studies revealed that frontal EEG monitoring using raw EEG patterns and spectrograms (color density spectral arrays) can be used to predict poor neurologic outcomes in postcardiac arrest patients. Irregular high-voltage and high-frequency beta waves detected by frontal EEG are significantly strong predictors of poor prognosis (14).

Although continuity, frequency, and discrete seizures were unaffected by temperature and did not show variance with temperature phases in a study assessing EEG findings in relation to temperature management, rewarming was associated with emergence of interictal epileptiform discharges (64). Increased interictal epileptiform discharges may suggest poorer prognosis (64). Valproate, levetiracetam, and low body temperature were each associated with increased probability of transitioning from epileptiform states to non-epileptiform states in patients with continuous EEG background activity, whereas in patients with discontinuous EEG background activity, no agent was linked to an increased probability of transitioning from epileptiform states (139). The impact of anticonvulsants on the epileptiform discharges depends on the presence of continuous cortical background activity. Another, novel EEG-based method for prognostication after cardiac arrest may include model-based quantitative EEG analysis (state space analysis) (32). Given that continuous EEG is labor-intensive, simplified EEG tracings obtained by bispectral index device might be a good alternative to assist in outcome prognostication (33).

Conventional imaging studies, including CT and MRI of the brain, first and foremost assist in assessment of morphologic abnormalities. Although CT is overall quite insensitive for outcome prediction, good 1-month outcomes have been found with early (less than six hours) higher bilateral mean ASPECTS score (72). Gray-white matter ratio and quantitative regional abnormality on early CT did not have significant implications. MRI of the brain can be used to assess findings of anoxic injury not overt on CT scan, classically visible as reduced diffusion in highly active metabolic areas of the brain. Reduced diffusion three to five days post ROSC can predict poor outcome (128), and postanoxic diffusion changes using quantitative brain MRI may aid in predicting persistent coma and poor functional outcomes (119; 98). Both PET and MRS also have been used to follow cerebral metabolic function in individuals suffering from cardiac arrest (140). More advanced imaging techniques such as fMRI and diffusion tensor imaging show promising results, but further evaluation is warranted (61).

It is known that large muscle mass is associated with a good premorbid condition (81). Skeletal muscles have inhibitory effects on neuronal cell death in postcardiac arrest survivors via different mechanisms: 1) antioxidant property of oxytocin, 2) anti-inflammatory action of myokine, and 3) antiapoptotic activity of creatine (73). Because lean body mass (LBM), which is the weight of the entire body excluding fat component, reflects an individual’s muscle mass better than body mass index (BMI), high lean body mass (greater than 48.98 Kg) is associated with better neurologic outcomes in cardiac arrest survivors who had undergone therapeutic temperature management (74).

There is a growing body of biochemical outcome prognosticators (45). Peripheral blood markers neuron-specific enolase (NSE) and protein S100beta, have been studied extensively for their use in prognostication of functional outcome after cardiac arrest (43). An increase in serum concentrations of these destruction proteins indicate ongoing neuronal destruction. NSE has been associated with poor outcome if its level is greater than 33μg/L at days one to three, but the lack of a standardized assay and the fact that these levels are not applicable to patients treated with TTM have limited its usefulness (45). Newer data indicate that determination of relative neuron-specific enolase values over time (obviating the need for an absolute cut off) may be reflective of ongoing neuronal injury of greater than 1.0 (19). Knowing the fact that both age and time to return of spontaneous circulation have an impact on neuron-specific enolase value, neuron-specific enolase at 48 hours after out-of-hospital cardiac arrest was used to predict 12-month-prongnosis in young patients and in patients with a long time from collapse to return of spontaneous circulation (154). Protein S100beta, a calcium-binding protein and an indicator of glial cell damage, lacks specificity when it comes to predicting neurologic outcome after cardiac arrest. Both markers may play an important role once standardization of measurement techniques is achieved. Procalcitonin measurement was used as a prognostic blood biomarker for outcomes in postcardiac arrest patients. Elevated procalcitonin level at zero to 48 hours was associated with poor outcomes (138). The stress-response cytokine growth differentiation factor-15 (GDF-15) is elevated after various kind of tissue injury including brain hypoxemia. It was used to predict poor survival and neurologic outcome in postcardiac arrest patients. (120). Neurofilament light (NfL) concentration is noted to be elevated in postanoxic patients, hence, neurofilament light concentration level was also used as a prognostic marker after cardiac arrest (29).

Other biomarkers under investigation, including co-peptin, secretoneurin, glial-fibrillary acidic protein, ubiquitin C-terminal hydrolase L1, myelin basic protein, and tau all are limited by sensitivity or specificity as well as variation in both time course of release and absolute concentration (45). Glycosylated ceruloplasmin and haptoglobin as well as other glycoproteins may play critical roles in neuroprotection and are evaluated as possible sensitive prognostic markers for patients treated with therapeutic hypothermia (27).

Prognostication of higher cognitive function and neuropsychological impairment after cardiac arrest is difficult, especially in the immediate post-arrest period. Patients can suffer various degrees of memory impairment, especially if coma duration is at least six hours or longer (87). Diffuse, sudden, ischemic-hypoxic injury caused by cardiac arrest does not seem to preferentially damage memory systems, though subacute or stepwise hypoxic or excitotoxic injury may cause isolated hippocampal injury and amnesia (77). A common pattern of impairment in the postacute phase after cardiac arrest is a combination of memory, subtle motor, and variable executive deficits. "Cognitive" event-related brain potentials (eg, P300 and mismatch negativity), when present, reflect the functional integrity of higher-level information processing and, therefore, the likelihood of capacity for cognition. In a small study of 12 patients with cognitive impairment after cardiac arrest, a dysexecutive syndrome was noted in all patients, but the overlap of behavioral frontal lobe syndrome, memory deficit, and extrapyramidal syndrome varied (116). Subjective estimates of patient’s condition after arrest is related to cognitive and functional outcome (58).

Other predictors include integrated physiologic response; in an analysis of patients with out-of-hospital cardiac arrest who underwent targeted temperature management, increased heart rate in the 48-hour rewarming phase was an independent predictor of favorable neurologic outcome (56).

In summary, the following are listed as poor predictors of neurologic outcomes after cardiac arrest in most studies (129):

On the other hand, the following findings are listed as good predictors of neurologic outcomes in a systematic review (130):

Clinical vignette

A 72-year-old, right-handed male completed cardiac bypass surgery and complained of loss of motor power in the lower extremities. The patient’s past medical history was significant for coronary artery disease, hypertension, and a 15-minute period of hypotension during the recent cardiac bypass surgery. The neurologic examination was significant for flaccid paralysis of the lower extremities, urinary retention, and loss of pinprick sensation bilaterally below the level of T10. The presentation was consistent with acute thoracic spinal cord ischemia secondary to transient systemic hypotension. The patient was stabilized with proper control of his chronic hypertension and any necessary correction of blood chemistries. Intermittent catheterization was employed for urinary retention. Over the course of the next 72 hours, the patient was fortunate to regain most of his original lower extremity motor power, sensory function, and autonomic function with proper supportive care.

Biological basis

Etiology and pathogenesis

During cardiac arrest, the brain is exposed initially to two primary insults: (1) the loss of oxygen and (2) the loss of glucose. The brain utilizes oxygen to metabolize glucose. It cannot store oxygen and survives only for minutes after its oxygen supply is reduced below critical levels. In acute anoxia, consciousness is lost within 15 seconds. Pyramidal cells in the CA1 sector of the hippocampus, Purkinje cells of the cerebellum, and pyramidal cells of the third and fifth layers of the cerebral cortex are vulnerable to even moderate degrees of anoxia. Widespread necrosis of the cortex with the brainstem intact produces a vegetative state. More profound anoxia affecting the cortex, basal ganglia, and brainstem results in coma and subsequent death.

Under physiologic conditions, glucose is the brain's only substrate and crosses the blood-brain barrier by facilitated transport. Each minute, the normal brain uses about 5.5 mg (31 µmol) of glucose per 100 g tissue. If there is hypoglycemia, defined in adults as a blood glucose concentration of less than 40 mg/dL, signs and symptoms of encephalopathy result secondary to cerebral cortex or brainstem dysfunction. The cerebral cortex is more vulnerable to the effects of hypoglycemia, whereas the brainstem and basal ganglia exhibit less histologic damage during periods of reduced serum glucose. Although periods of hypoglycemia may precipitate or confound anoxic coma, serum glucose must be maintained in a finely controlled range even after cardiac arrest to prevent further neurologic disability. Similar to the detrimental effects of hypoglycemia, periods of hyperglycemia (greater than 180 mg/dL) following global cerebral ischemia have been shown to worsen neurologic outcome (10).

Although no pharmacological interventions are currently available to provide neuroprotection after cardiac arrest, several pathways are being investigated based on the available understanding of secondary cell damage after anoxic injury and cardiac arrest. Following a global cerebral ischemic insult, the cellular environment in the central nervous system initially is influenced by the compromise of cerebral blood flow and metabolism. Experimental models have illustrated that ischemic generation of the free radical nitric oxide can be a "trigger" for the subsequent induction of neuronal injury, such as in cortical neurons (38) and hippocampal neurons (88; 86). The cellular pathways generated by free radicals, such as nitric oxide, that result in neuronal injury vary, including neuronal endonucleases (146; 80), intracellular acidification (57; 148), and cysteine proteases (142; 79; 86). As a result of its close link to the molecular pathways that lead to both neuronal injury and vascular injury, nitric oxide functions not only as a potential therapeutic target, but also as a valuable investigational agent. Current work focuses on subsequent "down-stream" pathways, such as protein kinase B (Akt) and microglial activation (59; 60), that appear to play a central role during the prevention of cellular injury and inflammation (16; 17). Another potential neuroprotective measures following cardiac arrest include inhibition of serum and glucocorticoid-regulated kinase 1 (SGK1), which protects against neuroinflammation, cardiac arrest-induced hypoperfusion, neuronal cell death, and neurologic deficits (73). There is currently a randomized, multicenter, clinical trial investigating the antiinflammatory and neuroprotective effect of prehospital administration a high-dose glucocorticoid following out-of-hospital cardiac arrest to mitigate the progression of post-cardiac arrest syndrome (106).

Programmed cell death is also believed to be one of the contributing factors to neuronal and vascular injury. Cellular programmed cell death can proceed through two dynamic, but distinct pathways that involve both DNA fragmentation and the loss of membrane asymmetry with the exposure of membrane phosphatidylserine residues (147; 88). The internucleosomal cleavage of genomic DNA into fragments may be a late event during programmed cell death and ultimately commits a cell to its demise. In contrast, the redistribution of membrane phosphatidylserine residues can be an early event during programmed cell death that usually precedes DNA fragmentation. Work that employs the ability to follow the progressive externalization of membrane phosphatidylserine residues in adherent monolayer living cells over time has provided evidence that cellular programmed cell death also may be reversible in nature (147; 88). This exposure of phosphatidylserine residues may serve to "tag" injured cells for phagocytosis or promote thrombosis in vascular cells (132; 15; 16; 17). In addition, phosphatidylserine exposure may facilitate injury mechanisms, such as attempted cell cycle induction, that lead to a cell's demise (78; 82). Postcardiac arrest patients are noted to have normal range of plasma glutamine and glutamate levels. One trial to reduce glutamine and glutamate plasma concentration by using continuous venovenous hemodiafiltration failed to statistically significantly lower their level. (103).

Other signal transduction pathways can also influence neuronal survival. Trophic factors and metabotropic glutamate receptors are increasingly being investigated as therapeutic regiments to prevent or reverse neuronal injury. The mechanisms employed by trophic factors to achieve neuroprotection can be diverse and are not well understood. Yet, growth factors from diverse sources, such as erythropoietin, have been shown to prevent neurodegeneration in hippocampal cultures during a variety of insults such as glutamate toxicity, hypoglycemia, and nitric oxide toxicity (15; 16; 17).

Several cloned metabotropic receptor subtypes have been linked to the modulation of neuronal survival (18). They function through signal transduction pathways such as cyclic adenosine monophosphate, protein kinase C, inositol phosphate, ion channel flux, and phospholipase D. Activation of the metabotropic receptors can reduce glutamate toxicity, protect synaptic transmission during periods of hypoxia, and increase neuronal survival during nitric oxide exposure (88; 11). Investigations have elucidated the role of the metabotropic glutamate system during cysteine protease activation (79), neuronal endonuclease and pH modulation (147), and intracellular calcium flux (85). Other work has expanded the cytoprotective role of the metabotropic glutamate receptor to the vascular system (79).

Although a comprehensive review of all ongoing research in neuroprotection would exceed the framework of this article, it is to be mentioned that there are multiple angles for potential neuroprotective agents. To name a few: dichloroacetate, a pyruvate dehydrogenase kinase inhibitor, was found to have neuroprotective effects, likely linked to an increase in mitochondrial energy metabolism in the brain (151). Temporary inhibition of apoptosis was achieved with a P53-inhibito, pifithrin-μ (37). Administration of the anesthetic sevoflurane after cardiac arrest was found to improve mitochondrial function and, hence, provide neuroprotection (152). Metformin may also be promising; an animal model showed that metformin ameliorated cardiac-arrest induced neuronal degeneration and glial activation (161). Injection of adenosine 5’-monophsophate was found to induce hypothermia in mice by reducing membrane potential and Ca signaling of mitochondria in neuronal cells, which helps to induce a hypometabolic state that slows mitochondrial respiration, reduces oxygen demand, and delays hypoxia-reperfusion mitochondrial injury in the brain and other organs (67). Administration of drugs that act on Akt1 can mimic hypothermia effects without the need for physical cooling, hence, can overcome challenges to implementing CPR cooling clinically (76). Furthermore, enteral administration of ubiquinol (reduced coenzyme Q10) in postcardiac arrest patients has been studied by Holmberg and colleagues to see whether it can improve oxygen consumption and reduce neurologic biomarkers of injury. The study showed enteral ubiquinol increased plasma coenzyme Q10 in these patients as compared to placebo but with no difference in neuron specific enolase, S100B, lactate, cellular and global oxygen consumption, neurologic status, or in-hospital mortality (52). Further development and translation of these findings will hopefully serve to establish clinically relevant neuroprotective strategies in the future.

Epidemiology

Every year, between 300,000 and 400,000 people die suddenly and unexpectedly in the United States (89). In 2017, sudden cardiac deaths contributed to nearly 380,000 deaths in the United States (149). There are 292,000 adult in-hospital cardiac arrests and 15,200 pediatric in-hospital cardiac events in the United States each year (53). A comprehensive systematic review and metaanalysis in 2022 by Amacher and colleagues demonstrated that long-term survival after 10 years in patients surviving the initial hospital stay following out-of-hospital cardiac arrest was between 62% and 64% (02).

Prevention

Less than 10% of cardiopulmonary resuscitation attempts (pre-hospital or in hospitals, exclusive of intensive care units) result in survival without central nervous system damage (124). Reduction in response times and improved resuscitative techniques can improve prognosis. One factor that appears to increase survival and to decrease morbidity is the rapid onset of cardiopulmonary resuscitation initiated by a bystander. In some cases, this can triple the chance of surviving a cardiac arrest outside of the hospital (49). Furthermore, cardiopulmonary resuscitation provided for four minutes in a compression-ventilation ration of 100:2 can achieve improved neurologic outcome (126), and resuscitation algorithms have been adjusted to reflect this finding. Efforts to facilitate early defibrillation include the use of automatic defibrillators that employ electrocardiographic diagnosis and defibrillation. Such devices have proven to be efficacious both inside (28) and outside (63) of the hospital environment. Studies have promoted the virtues of automatic defibrillators on airlines (110) and in casinos (144). Mechanical compression devices, introduced in efforts to improve quality of CPR, do not seem to infer improvements in neurologic outcomes (150). Studies showed that emergency department arrests resemble cases of in-hospital cardiac arrest (IHCA), for which both arrests occur in well-resourced healthcare settings where resuscitations can be done effectively and promptly more than out-of-hospital cardiac arrest (OHCA) (95). Team-based CPR training for emergency medical service providers was shown to improve the prehospital return of spontaneous circulation rates of outside hospital cardiac arrest (OHCA) patients (111). Interestingly, survival rate after in-hospital cardiac arrest is associated with time and location of arrest. It is more than 20% if the arrest occurs between 7 am and 11 pm but only 15% if the arrest occurs between 11 pm to 9 am. Also, survival rate is about 9% at nighttime in unmonitored settings compared to 37% in operating room/postanesthetic care unit during the daytime (115). Surprisingly, a French study on 30,672 adult, nontraumatic outside hospital cardiac arrest patients showed that patients with early advanced life support (E-ALS) were less likely to have a good neurologic outcome at one month post-arrest. One explanation to their unexpected result was the total duration of resuscitation performed, which may interrupted prematurely in cases of early advanced life support (94).

Differential diagnosis

The differential diagnosis for anoxic or hypoxic brain injury is broad. The main consideration is to corroborate the appropriate history. The most important categories to consider for the encephalopathy include anoxic, toxic-metabolic, vascular, traumatic, infectious, inflammatory, or epileptic. Although history and circumstances are oftentimes overt in setting of a documented cardiac arrest, cardiac arrest itself can be caused by a variety of underlying etiologies, with cardiogenic causes usually being the main culprit in cardiac arrest with shockable rhythms. Overall, more than 70% of cardiac arrests are caused by acute myocardial infarction or massive pulmonary embolism.

Diagnostic workup

When assessing the postcardiac arrest patient, two main considerations drive diagnostic workup: (1) the potential need for treatment of complications from anoxic encephalopathy, such as status epilepticus or elevated intracranial pressure, and (2) the quest for prognostication. Main diagnostic tests for the assessment of need for treatment are continuous EEG and brain imaging, largely CT scan of the head.

For prognostication, national guidelines are available (155); however, those do not include the commonly used modalities of EEG and neuroimaging as they lack validation (47). However, traditional guidelines for neurologic prognostication have proven unreliable in modern studies of patients with cardiac arrest in the era of therapeutic hypothermia and targeted temperature management; hence, an algorithmic approach is suggested (40). In the first 24 hours, assessment includes clinical examination and continuous EEG as well as baseline imaging if indicated. The EEG is useful in assessing cortical dysfunction and identifying the presence of epileptic activity and can be assessed for reactivity and evolution of patterns in the prognosis section of this article. Between 24 and 48 hours, neuron-specific enolase may be measured in addition to continuous assessment of clinical examination and EEG. At 48 to 72 hours after arrest or after rewarming from temperature management, additional assessment with repeat imaging and SSEP can be undertaken. The usual imaging modality to start with is a CT scan of the brain, which will show evolving ischemia, hemorrhages, or cerebral edema and imminent herniation. If the CT is unremarkable, an MRI of the brain, performed three to five days post-arrest or after rewarming, can assist in determining the presence of less overt complications following cardiac arrest, namely findings of anoxic injury. MRI imaging using diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) mapping are becoming more sensitive tools to detect progressive neuronal injury and estimate clinical outcome in post cardiac arrest patients (108). Overall, DWI changes in cortical gray matter structures, deep gray nuclei, white matter, brainstem, and the cerebellum within 72 to 96 hours after ROSC have significant prediction of poor neurologic outcome (03).

Overall, a multimodal approach with a combination of clinical, biochemical, and electrophysiological investigations is recommended in order to predict neurologic outcome after cardiac arrest reliably (160). Furthermore, delay of prognostication to determine which patients might make a delayed, meaningful recovery and avoid self-fulfilling prophesy is recommended (40).

Management

On the establishment of adequate ventilation and perfusion, blood should be obtained for determination of serum glucose and routine chemistries. Bedside stat glucose determinations should identify hypoglycemia. Identification of hyperglycemic states also is important because elevated serum glucose may promote ischemic damage in cases of anoxic coma (10).

The hemodynamics of the patient should be closely controlled. Hypertension may be secondary to Cushing reflex with increased intracranial pressure or a result of brainstem ischemia. Bradycardia associated with an elevated blood pressure suggests brainstem compression or raised intracranial pressure. Reversible causes of transtentorial herniation, such as subdural hematoma, should be immediately considered before (repeat) cardiovascular collapse ensues. Hypotension may be indicative of myocardial infarction, hemorrhagic shock, sepsis, or sedative-hypnotic drug effects. A double-blind randomized trial looked at aggressive hemodynamic strategies by targeting mean arterial blood pressure of 77 mmHg compared to 63 mmHg in patients who had been resuscitated from an out-of-hospital cardiac arrest of presumed cardiac cause and did not result in significantly different percentages of patients dying or having severe disability or coma (66). The same trial also demonstrated that targeting a restrictive oxygen target of partial pressure of arterial oxygen (PaO2) of 9 to 10 kPa (68 to 75 mmHg) resulted in a similar incidence of death or severe disability or coma when it was compared with a liberal oxygen target of PaO2 of 13 to 14 kPa (98 to 105 mmHg) (137). When treated with extracorporeal cardiopulmonary resuscitation (ECPR), low-flow time seems to be associated with favorable neurologic outcomes, and ECPR should be performed within 58 minutes of the low-flow time according to one series (109). Although infectious complications are common among survivals of cardiac arrest, administering prophylactic antibiotics were not associated with increased survival, survival with good neurologic outcome, critical care length of stay, or incidence of pneumonia (24).

In addition to managing the acute medical problems, focus on systems and best organization of care is also important. Data from Australia indicate that shorter time in the emergency department until transfer to destination ward was associated with improved outcomes (113).

Status epilepticus following cardiac arrest can result in progressive anoxic brain damage and requires immediate attention. Following airway stabilization, generalized convulsions can initially be treated with intravenous benzodiazepines. This is to be followed by intravenous loading with a first-line antiepileptic medication, such as phenytoin, valproic acid, phenobarbital, levetiracetam, or lacosamide. If status epilepticus continues and becomes refractory, general anesthesia with propofol or midazolam is required. In the case of generalized convulsions that are not consistent with status epilepticus but considered epileptic seizures, an antiepileptic medication in an appropriate daily dose (dependent on body size) should be maintained, especially in individuals with EEG, CT, or MRI evidence of a persistent epileptic focus (hemorrhage, neoplasm, large ischemic infarct, abscess, etc.). However, distinct epileptiform patterns with evolving seizures are rare, and other rhythmic activity, such as generalized periodic discharges or rhythmic delta activity, is more common (50). The overall usefulness of treatment of electrographic status epilepticus after cardiac arrest is unclear (51). Similarly, aggressive treatment of myoclonus is not shown to alter outcomes as irreversible damage is likely to be present in patients with myoclonus (127). However, some experimental animal studies have shown a beneficial effect of high-dose valproic acid in both seizure prevention and survival (107).

Measurement of the patient's core temperature is a vital component of the initial evaluation. Hypothermic patients with temperatures below 34°C (93.2°F) should be warmed slowly to a body temperature above 36°C (96.8°F). Because hypothermia below 80°F results in coma, resuscitative measures are indicated in all hypothermic patients even if vital signs are absent. Hypothermic patients have recovered following cardiac arrest, presumably because of the protective effects of low body temperature and depressed cerebral oxygen requirements. In addition, hypothermia has been shown to reduce neuronal death in the hippocampus and caudate-putamen in animal models with forebrain ischemia (153). Sawamoto and associates studied the outcome from severe accidental hypothermia with cardiac arrest resuscitated with extracorporeal cardiopulmonary resuscitation, and they found that patients with hypothermic cardiac arrest due to nonasphyxial hypothermia have improved neurologic outcomes when treated with ECPR compared to patients with asphyxial hypothermic cardiac arrest (133). Masaki and colleagues report a case of life-threatening deep hypothermia from which the patient was successfully revived after more than four hours of cardiopulmonary arrest (90). Conventional rewarming methods were used, and the patient recovered fully without any neurologic deficits. Four hours of circulatory arrest appears to be the limit for patients with deep hypothermia to survive with little or no cerebral impairment.

In an effort to minimize ongoing neurologic injury after resuscitation and ROSC, the use of targeted temperature management (TTM) in comatose survivors of cardiac arrest has become the standard of care after two landmark trials in 2002 that showed remarkable benefit of therapeutic hypothermia (06; 55). A TTM study group showed that mild therapeutic hypothermia to 36°C is comparable to 33°C in terms of neurologic outcomes (104). Interestingly, the duration of targeted temperature management was reported to have effects on cognitive outcome at six months in post-outside hospital cardiac arrest survivors. Targeted temperature management for 48 hours was associated with reduced memory retrieval deficits and lower relative risk of cognitive impairment compared to standard targeted temperature management for 24 hours (34). However, there was no significant difference in six-month survival rate upon shifting targeted temperature management from 33°C to 36°C in out-of-hospital cardiac arrest patients (01). Therapeutic temperature management at 33°C was associated with a lower risk of unfavorable neurologic outcome when the cerebral performance categories were analyzed (97). On the other hand, increasing age was associated with higher mortality, but not unfavorable neurologic outcomes at six months after cardiac arrest (114). A TTM2 trial in 2021 compared a target temperature of 33°C to normothermia (36.5°C to 37.7°C) in patients who are unconscious following out-of-hospital cardiac arrest and revealed no difference in all-cause mortality or poor functional outcome at 6 months (25). On the other hand, there was a higher rate of arrhythmia with hemodynamic instability in hypothermia group. Overall trends in aggressiveness of care after cardiac arrest also indicate not only that use of TTM has increased over the past years but also has use of coronary angiogram, PCI, and ECMO (112). Therapeutic hypothermia and targeted temperature management use was also more frequent when a neurology consultant was involved in patient assessment and care after cardiac arrest (46). In patients undergoing therapeutic hypothermia after cardiac resuscitation, awakening usually occurs within three days of the cardiac arrest and is not delayed compared with nonhypothermia patients (36).

Given that targeted temperature management and therapeutic hypothermia are resource-intensive, efforts are made to determine usefulness in individual cases. The CAST score, a scoring system using eight factors (initial rhythm, witnessed status and time until return of spontaneous circulation, pH, serum lactate, motor score according to the Glasgow Coma Scale (GCS), gray matter attenuation to white matter attenuation ratio (GWR), serum albumin, and hemoglobin) was developed to identify the post-cardiac arrest patients with a good potential for recovery prior to the initiation of induced therapeutic hypothermia. The score is awaiting prospective validation (105).

Some centers advocate aggressive treatment of raised intracranial pressure to significantly reduce mortality. Measurement of intracranial pressure can be performed through the use of parenchymal intracranial pressure monitors or through intraventricular pressure measurements. Intracranial pressure monitoring has been linked toward prognosis. Most patients with a maximum intracranial pressure increase of less than 30 mm Hg experience good recovery, whereas a pressure rise above 25 to 30 mm Hg represents a great risk for brain tamponade (44). Some therapy has focused on the use of hemicraniectomy, such as during large cerebral infarcts with impending herniation. Although in some studies mortality is reduced with this procedure, there is no apparent improvement in morbidity (54).

Special considerations

Pregnancy

Although cardiac arrest in pregnancy is rare, it is increasing in frequency with an incidence to be at least 1 out of 12,000 births (93). It requires resuscitative measures directed to the restoration of the maternal and fetal hemodynamics. Respiratory insufficiency and hypotension/hypoperfusion were common antecedents to maternal cardiac arrest (159). As with nongravid females, neurologic outcome is improved when the delay to initiate resuscitation is minimal (05). Cases of cardiac arrest during the third trimester of pregnancy may necessitate implementation of a more aggressive approach than standard resuscitative protocols. Cardiac arrest in advanced pregnancy may lead to perimortem cesarean delivery to save the mother and her infant (83). Cesarean section has been advocated for women in the latter part of pregnancy following cardiac arrest to avoid both maternal and fetal distress (23). Kozinski and colleagues reported that a regional system of care for out-of-hospital cardiac arrest (OHCA) survivors, when successfully implemented, leads to an improvement in neurologic status and to a reduction of in-hospital mortality in patients treated with mild therapeutic hypothermia, pending prospective trials (68). In addition, they achieved their results without any excess of complications. Wilberg and colleagues reported that in assessing intrapartum hypoxia, lactate in cord arterial blood at birth may replace base deficit as an acid-base outcome parameter at birth (156).

Causes of anesthesia-related cardiac arrest in pregnancy might include failed intubation or ventilation, aspiration of gastric contents, hemodynamic or respiratory complications related to high or total spinal anesthesia, or local anesthetic toxicity. Well-known causes of nonanesthesia-related cardiac arrest in pregnancy include hemorrhagic causes, acute coronary syndromes, rupture of aortic aneurysm, stroke, air embolism, amniotic fluid embolism, magnesium toxicity (especially in oliguric patients), pulmonary embolism, trauma, preeclampsia or eclampsia, status asthmaticus, and sepsis.