30.1 Introduction
Cardiac arrest (CA) is a sudden loss of heart function due to either a primary cardiac event such as myocardial event or secondary effect on the heart from neurological, respiratory, or metabolic causes. For out-of-hospital cardiac arrests (OHCA) patients, <10% survive to hospital discharge compared to 20–25% of in-hospital cardiac arrest (IHCA) patients that survive to hospital discharge [1, 2]. Eighty to Ninety percent of post-cardiac arrest survivors emerge with coma and severe long-term neurological deficits [3].
At the epicenter of the high mortality and morbidity from CA, is the post-cardiac arrest syndrome, which includes anoxic brain injury, myocardial dysfunction, and a systemic ischemia and reperfusion syndrome. Of these, perception of injury to the brain is the most common cause of death in two-thirds of patients with OHCA and a quarter of patients with IHCA [4]. The brain is prone to injury due to lack of significant intrinsic energy and nutrient stores, therefore highly dependent on a constant supply of oxygen and nutrients. Multiple factors play a role in the extent and pattern of brain injury post-cardiac arrest including the initial ischemic cascade, the reperfusion injury after return of spontaneous circulation (ROSC), the delay ischemia due to the no reflow phenomenon, and post-resuscitation variable such as pyrexia and hypoglycemia. We should also note that only 10% of these patients with brain injury progress to brain death [5].
This chapter will review the clinical manifestation of brain injury after a cardiac arrest, the pathophysiology of brain injury in this setting, the management post-CA, and neuroprognostication of patients post-CA.
30.2 Clinical Manifestations of Brain Injury After Cardiac Arrest
Certain neuronal populations in the brain are particularly susceptible to ischemia and tend to be prone to injury from hypoxia. Cortical pyramidal neurons, cerebellar Purkinje cells, and the CA-1 neurons of the hippocampus are the most vulnerable [6]. Areas of the brain that are phylogenetically more ancient such as the brainstem, the thalamus, and the hypothalamus tend to be more resistant to injury compared to the cortex [7, 8]. The brainstem in particular is resistant to hypoxic injury, as manifested by the preservation of cranial nerve and sensory motor reflexes. Bilateral cortical or thalamocortical complex injuries result in dysfunction in arousal or consciousness [9]. Other areas affected by hypoxia include the basal ganglia and the cerebellum, which are responsible for movement disorders and the dyscoordination that is observed post-arrest.
30.2.1 Disorders of Arousal and Consciousness
Disorders of arousal are the most common presentations of hypoxic-ischemic injury after cardiac arrest. The ascending arousal system largely originates from a series of well-defined cell groups in the brainstem and is composed of two major pathways [10]. The first pathway ascends from the upper brainstem nuclei to the thalamus and activates the thalamic-relay nuclei and the thalamic reticular nucleus, with subsequent transmission to the cortex. The second pathway originates in the upper brainstem and caudal hypothalamus and bypasses the thalamus, and activates neurons in the lateral hypothalamus, basal forebrain, and throughout the cerebral cortex. Lesions anywhere along either of these arousal systems can cause a decrease in arousal.
If the pattern of injury involves the cortex but spares the brainstem and thalamus, one can observe a more vegetative state characterized with arousal and preservation of sleep-wake cycles, but a lack of awareness of self and purposeful response to the environment. However, involvement of these deeper structures including the thalamus, hypothalamus, and brainstem can result in a complete coma with no arousal and awareness.
30.2.2 Seizures
Damage to the cortex from hypoxic-ischemic injury can result in neuronal injury leading to seizures, which can present clinically as convulsive or non-convulsive. Incidence of seizure has increase over the last decade in part because of the increasing use of continuous EEG leading to better detection. One study that particularly looked at patients undergoing therapeutic hypothermia after cardiac arrest suggested that up to one in three patients had seizures, often non-convulsive, and this occurred despite the use of sedative medications [11]. In the targeted temperature management study (TTM), post-hoc analysis revealed a seizure incidence of 29% [12]. Seizures are potentially treatable complication that must not be used as a negative predictor until intractability to aggressive therapy is proven [13].
30.2.3 Post-Hypoxic Myoclonus
Myoclonus refers to brief, involuntary, and shock-like muscle contractions or lapses in muscle tone. Post-hypoxic myoclonus (PHM) can be distinguished into two fairly distinct clinical syndromes, acute post-hypoxic myoclonus that is generally self-limiting after a few days, and chronic post-hypoxic myoclonus, or Lance–Adams syndrome.
Acute post-hypoxic myoclonus is the most common form of myoclonus after cardiac arrest and often occurs within hours of hypoxic injury. It is poorly localized to the cortex, hippocampus, cerebellar Purkinje cell layer, the reticular thalamic nucleus, and the reticular formation of the medulla, based on animal and clinical studies [14]. Acute post-hypoxic myoclonus typically warrants an EEG evaluation to evaluate for seizures or myoclonic status epilepticus. There are no published guidelines of treatment of acute PHM, but if epileptiform discharges are present, clinicians can consider treatment with antiepileptic drugs.
Chronic post-hypoxic myoclonus, also known as Lance–Adams syndrome, develops days or weeks after the hypoxic episode and tends to persist. It is typically a multifocal action or intention myoclonus, which means that it is exacerbated during muscle activation or with intention in the setting of preserved consciousness. Some antiseizure medications have been used successful to control this syndrome.
30.3 Pathophysiology of Brain Injury After Cardiac Arrest
Injury to the brain after cardiac arrest occurs in two distinct windows. Primary injury to the brain occurs during the hypoxia itself, whereas secondary injury occurs after return of spontaneous circulation. An appreciation of the mechanisms of injury provides the intensivist insight into the types of interventions that are available and future therapeutic options that are being explored to mitigate the extent of injury.
30.3.1 Primary Injury
Brain tissue is very sensitive to lack of oxygenation, and even brief ischemic periods of a few minutes can trigger a complex sequence of events leading to cell death. These mechanisms have been identified through study in various mammalian models. During hypoxia itself, the lack of oxygen in brain tissue causes cells to switch over to anaerobic respiration and this subsequently leads to a decrease in adenosine triphosphate (ATP) production. With diminished ATP supply, the sodium-potassium-ATPase pump on the cellular membrane no longer functions [15]. This results in accumulation of sodium ions inside cells, and subsequent cytotoxic edema. The lack of ATP also causes the cell to switch over to anaerobic respiration, resulting in cerebral lactate accumulation and intracellular acidosis [16]. The ischemia causes calcium influx through the cell, and release of glutamate, which serves an excitatory neurotransmitter. Glutamate binds to post-synaptic NMDA and AMPA receptors, resulting in downstream activation of degradative lipases and proteases that ultimately take neurons and other cells down a path to cellular death [17]. The immediate goal is to minimize the primary injury to the brain and systemic organs by providing prompt high quality cardiopulmonary resuscitation.
30.3.2 Secondary Injury After ROSC
The injury to the brain that occurs during hypoxia itself sets the stage for a maladaptive response of cells to reperfusion. This results in secondary injury that continues to cause cell death even after oxygen and blood flow have been re-established. The precise mechanisms of injury that occur during this phase are still under investigation, but multiple pathways of injury have been established.
Hypoxia during primary injury can be thought of as creating the pre-requisites for a maladaptive response. Terminal sequelae that ultimately result in death during secondary injury include direct cell death from formation of reactive oxygen species; cytotoxic edema and vasogenic edema that cause further ischemia and result in herniation; and finally, disorders of cerebral blood flow autoregulation that cause further hypoxic injury.
The hypoxia that occurs in primary injury causes significant mitochondrial dysfunction. After return of spontaneous circulation (ROSC), reactive oxygen species form in the mitochondria that then activate degradative enzymes, and damage the cellular structure, eventually leading to cell death. Furthermore, hypoxia during the initial lack of perfusion activates microglia that then initiate a pro-inflammatory cytokine release. There is also an associated increased migration of peripheral macrophages, monocytes, and neutrophils. Reperfusion results in the persistence and exacerbation of a cerebral inflammatory cascade, which results in damage to the endothelium. Endothelial dysfunction in turn leads to the formation of diffuse microthrombi in the cerebral vasculature [18]. Along with impaired vasodilation, there is an increased resistance to flow which decreases cerebral perfusion. The increased resistance to flow and the endothelial dysfunction also contribute to vasogenic edema, which in turn causes mass effect and further decreased flow to the area [19]. This increased edema also contributes to intracranial hypertension and resultant decrease in cerebral perfusion, eventually leading to transtentorial herniation and brain death.
There are various physiological factors that can further exacerbate the process of secondary injury. Hyperthermia can increase blood–brain barrier permeability, worsening cerebral edema, intracranial pressure (ICP), and ischemia [20]. Hyperthermia also increases glutamate production, which can cause further activation of degradative and harmful cellular enzymes. Significant anemia has been shown in animal studies to exacerbate secondary injury but data in human studies is limited. Both hypocapnia and hypercapnia can cause further injury. Hypocapnia can cause vasoconstriction, decreased cerebral blood flow, increased oxygen extraction, and result in ischemia. Hypercapnia can increase cerebral blood flow, which can cause hyperemia, increased intracranial pressure, excite-toxicity, and increased cerebral oxygen demand. While the optimal partial pressure of carbon dioxide in post-cardiac arrest patients is not known, the recommendation is to maintain carbon dioxide within the normal range.
30.4 Management Post-Cardiac Arrest
Post-cardiac arrest care is tailored toward identifying the cause of the arrest in individual patient along with management of ischemic reperfusion injury of multiple organ failure. This management focuses on cardiovascular care, targeted temperature management, respiratory care, and other neuroprotective measures in the ICU.
30.4.1 Cardiovascular Care
Per recent AHA guidelines, emergent coronary angiography should be performed in OHCA patients with suspected cardiac etiology of arrest and ST elevation on ECG, and reasonable in those patients without ST elevation. During the post-resuscitation, it is reasonable to avoid and immediately correct hypotension, defined as systolic blood pressure <90 mmHg or MAP <65 mmHg [21].
30.4.2 Therapeutic Hypothermia (TH) vs. Targeted Temperature Management (TTM)
Several studies have evaluated the effect of hypothermia on outcome post-cardiac arrest with different temperature goal, 32–34 °C in HACA study, 33 °C in the Bernard study, and recently a comparison of 33 °C vs. 36 °C in the Nielsen TTM study [22]. The 2015 AHA guidelines recommend that comatose adult patients with return of spontaneous circulation after cardiac arrest have a targeted temperature between 32 °C and 36 °C for both shockable and nonshockable rhythm and both OHCA and IHCA for at least 24 h with level of evidence B and C [21, 23]. However, the recent American Academy of Neurology practice guideline summary on post-cardiac arrest care has specific recommendation based on the scientific evidence and study designs. In OHCA patients with shockable rhythm, TH 32–34 °C for 24 h has a level A evidence for improving functional neurological outcome compared with TTM of 36 °C, which has a level B evidence. For nonshockable rhythm and patients, TH possibly leads to better survival and functional neurological status (level C evidence) [24]. Given this evidence, our temperature preference is 32–34 °C for post-cardiac arrest care. We should also note that temperature selection is based on individual patient characteristics, with lower temperature preferred in patients who may be at risk for seizure or brain edema, and higher temperature preferred in patients with risk of bleeding and hypotension. A period of normothermia of 72 h after TH/TTM is recommended. Following TH/TTM, the rewarming process should be gradual at a rate no faster than 0.25 °C/h to achieve normothermia and to avoid complications such as electrolytes abnormality, cerebral edema, and seizures [21]. Another key point in the AAN practice guideline is that prehospital cooling with intravenous chilled saline should not be offered, as it is unlikely to lead in improvement of functional neurological outcome with level A evidence [24].
Shivering is often encountered during TH/TTM, as it is the body’s natural defense mechanism against cold. Shivering can cause disruption of therapy by producing heat, thereby increasing body core temperature during the cooling process [25]. Shivering needs to be controlled. Most hospitals have a protocol for shivering, which include topical warming with warm blanket, and various pharmacological agents depending on the degree of shivering such as magnesium sulfate, buspirone, meperidine, sedatives, and paralytics.
30.4.3 Respiratory Care
It is important to note that with TH/TTM, reported PaCO2 values might be higher than actual patient values. The goal is to maintain normocarbia (end-tidal CO2 30–40 mmHg or Pa CO2 35–45 mmHg) unless there are other patient’s clinical factors that demand permissible hypercapnia (e.g., ARDS) or hypocapnia (e.g., hyperventilation in treating for cerebral edema) [21].
30.4.4 Cerebral Perfusion, Cerebral Edema, and Increased Intracranial Pressure
Hypotension, hypoxia, and hypercapnia may worsen cerebral damage and should be avoided [21]. Although, a MAP >65 mmHg has been suggested post-cardiac arrest, it is likely not sufficient for adequate cerebral perfusion. One study suggested that a MAP of 80–100 mmHg is beneficial, at least for the first 24 h after arrest [26]. In comatose patients with clinical signs of herniation, or cerebral edema on CT scan, ICP monitoring may be helpful to guide therapy in identifying optimum cerebral perfusion [27].
30.4.5 Fever Control
Fever is associated with poor neurological outcome as it may worsen secondary brain damage after cardiac arrest. Antipyretics and surface or invasive cooling measures should be used post-cooling to prevent fever and also used in patients not deemed candidate for TH/TTM [21, 23].
30.4.6 Glucose Control
Glucose management remains controversial in critically ill patients. Hyperglycemia has been linked to poor outcome after ischemic brain injury [28] and tightly controlled glucose leading to frequent hypoglycemia is known to be harmful. The target range for blood glucose is unknown post-cardiac arrest and current guidelines report the uncertainty of the benefit of any specific target range [21].
30.4.7 Seizure Control
Seizures, status epilepticus, and other epileptiform activity are common after cardiac arrest, with some features associated with poor neurological outcomes. Prophylactic antiepileptic drugs are not recommended post-cardiac arrest, but any patient noted to have seizure activity should be treated with standard antiepileptic medication. Any comatose patient after ROSC suspected of having seizure or not regaining consciousness should have an EEG promptly performed, interpreted, and monitored frequently or continuously [21].
These mechanisms of injury, and the physiological factors that can further exacerbate them, offer insights into the management post-cardiac arrest in the ICU. The next question following these management strategies is the prognostication of outcome.
30.5 Neuroprognostication
One of the challenges of neurointensivists is to determine the neurological outcome of post-cardiac arrest patients, which is time dependent, i.e., a patient with a cerebral performance category (CPC) score of 3 at discharge may improve to a better performance category at 3 months [29].
Most prognostications in acute brain injury are a prediction of poor outcome, which includes death, persistent coma, persistent vegetative state, or a state of temporary or permanent independence [30, 31]. It is crucial to give the correct prognosis because too much optimism may lead to survival of neurologically devastated patients, whereas a pessimistic prediction may result in withdrawal of life-sustaining therapy (WLST) and death of a patient who would otherwise have favorable outcome [32, 33]. Several authors have noted that WLST in clinical practice is a self-fulfilling prophecy, because of the bias it introduces in both trials and observational studies, whereby as a consequence of a perceived poor outcome, treatment is limited or withdrawn, leading to realization of the predicted poor outcome [34].
Neuroprognostication is a multidisciplinary approach using a combination of clinical examination, neuroimaging, neurophysiological modalities, and biomarkers data. It is important to point out that most existing studies have limited reliability because of the lack of well designed, randomized, and blinded studies. In prognostication studies, the prevailing parameter is the false positive rate (FPR), which indirectly indicates the extent with which one can make mistake if using the particular parameter, i.e., the lower the FPR, the better the tool. Many studies provide FPR that is too high (FPR > 10); furthermore, the 95% confidence intervals of these studies are also very wide.
30.5.1 Clinical Examination
Clinical features used in prognostication post-cardiac arrest include: eyes findings, best motor response, and status myoclonus. After ROSC, the timing of prognostication is crucial and is influenced by the presence or absence of therapeutic hypothermia (TH). In the absence of TH, at least 72 h post-arrest is needed for a reliable exam [29, 35]. In patients treated with TH, the timing of an accurate neuroprognostication is much later because hypothermia reduces the clearance of sedatives and neuromuscular blockade drugs used during the cooling process [36–38]. Most clinicians will wait at least 72 h or longer after the patient has been rewarmed to prognosticate but some important considerations need to be made. In the TTM trial, the prognostication of resuscitated comatose patients was determined at a median of 118 h after cardiac arrest [22].
Ocular brainstem reflexes are important in prognostication. The absence of pupillary light reflex is the most accurate clinical predictor of poor outcome at 72 h or longer from ROSC [39]. Corneal reflex is less specific. Its absence is a poor prognostic sign with high FPR up to 5% [40, 41], because of the effect of muscle relaxants. In the 2015 AHA guidelines, absent pupillary light reflex post-cardiac arrest is a strong predictor of poor outcome with FPR of 0% in patients treated with or without TTM at 72 h, whereas absent corneal reflex at 24 h and 48 h post-arrest predicted a poor outcome, with FPR of 17% and 7%, respectively [21].
Regarding the motor response, an absence or extension response at 72 h or more after ROSC is a sensitive predictor of poor outcome, with low specificity and high FPR (10–40%) because of the effect of neuromuscular blockade and sedative [29, 41, 42]. It is a class III harm based on AHA 2015 guidelines and should not be used alone to predict poor outcome [21].
Evidence of early post-cardiac arrest status myoclonus, defined as continuous repetitive myoclonic jerks >30 min, is a sign of poor outcome with a FPR of 0% at 24 h post-arrest [21, 43, 44], whereas the presence of any simple myoclonus within 72 h after cardiac arrest has a 5% FPR, making it a non-reliable predictor of poor outcome [21].
It must be emphasized that the uncertainty in prognostication of poor outcome in comatose survivors becomes less over time [42]. While 72 h after rewarming seems to be favored, it should also be noted delayed recovery after 72 h of coma with favorable outcomes is being reported [5]. It’s worthwhile highlighting that in the recent TTM study, prognostication of outcome was discouraged before 108 h [22].
30.5.2 Neuroimaging
After ROSC, a computer tomography (CT) or MRI of the brain is obtained to assess any evidence of brain injury and to help prognosticate. Abnormal CT findings include intracranial hemorrhage, or evidence of early post-anoxic injury as diffuse brain edema, which can be quantified as the ratio between the densities of the gray matter and the white matter (GWR) at the level of the basal ganglia, within 1 h after ROSC [45]. An average GWR below 1.14 Hounsfield or a GWR below 1.22 Hounsfield predicted a poor neurological outcome with 100% specificity [45, 46].
On MRI studies, diffusion-weighed imaging (DWI) can detect neuronal cytotoxic edema. The presence of bilateral hippocampal hyperintensities on DWI and fluid-attenuated inversion recovery (FLAIR) sequences is a marker of poor prognostic sign according to one study [47]. Apparent diffusion coefficient (ADC) can quantify the severity of brain injury. Decrease ADC values correspond to area of diffusion restriction likely because of infarction. Normal values range between 700 and 800 × 10−6 mm2 s−1 [5], whereas ADC values <665 × 10−6 mm2 s−1 is a significant predictor for unfavorable neurological outcome [48, 49].
Although these neuroimaging techniques are promising, they are unfortunately not yet well validated and not practical, especially in the early acute phase when patients are critical.
30.5.3 Electrophysiological Studies
EEG patterns associated with poor neurological outcomes, with high sensitivity and specificity, include low voltage output pattern, burst suppression, alpha/theta coma, focal or generalized seizures, generalized periodic epileptiform discharges, status epilepticus, and background unreactivity [50]. A nonreactive EEG background at least 72 h post-cardiac arrest is incompatible with good neurologic recovery, with 100% specificity [50–53].
SSEPs are strong predictors of poor outcome in post-anoxic coma as determined by the bilateral absence of short latency (N20) SSEP response [54]. In the absence of therapeutic hypothermia, a bilateral absence of cortical N20 response is a reliable predictor of poor outcome as early as 24 h from ROSC with a false positive rate (FPR) of 0.7% [29, 31]. In patients treated with therapeutic hypothermia, a composite result of recent studies showed that bilateral absence of N20 SSEP response after rewarming is a reliable predictor of poor outcome with a FPR of 0.9% [41, 55, 56]. With comparison to EEG and clinical examination, SSEPs are better predictors of poor outcome post-cardiac arrest as they are less affected by sedation and temperature [21]. It is critical to note that though SSEPs are superior and promising, the above studies are still limited given the high rates of WLST and self-fulfilling prophecy; therefore, it is still better to wait at least 72 h to prognosticate.
30.5.4 Biomarkers
Two most common biomarkers used in post-cardiac arrest are neuron specific enolase (NSE), derived from neuron, and the S100B protein, derived from astrocyte and Schwann cells [39]. Higher levels of these biomarkers correlate with a higher extent of brain injury and therefore poor outcome [39], whereas patients with good outcomes have lower S100B levels and lower NSE levels on day 1 and 3 post-arrest (p < 0.0083) [57]. However, there are several limitations of these biomarkers, including extra central nervous system sources (from hemolysis, neuroendocrine tumors, muscles, and adipose tissue breakdown), lack of laboratory standards between centers, and false positive results from hemolysis [21], which make them not a good prognosticator tool, but rather a confirmatory tool.
In summary, a combination of absent pupillary light reflex at 72 h after rewarming, absent bilateral SSEPs response, malignant EEG patterns, and abnormal neuroimaging are prognostic factors of poor outcome.
30.6 Future Directions
Neurological outcome after cerebral resuscitation post-cardiac arrest is highly dependent on the prompt administration of high quality CPR leading to ROSC and the management post-ROSC, with strong evidence for mild therapeutic hypothermia (32–34 °C) for 24 h. Clinicians must use a multidisciplinary approach by combining experience with available clinical, neuroimaging, neurophysiological, and biomarkers data to counsel family regarding the prognosis of their loved ones. Timing is key in prognostication as there are many confounding factors in early prognosis, such as evolving disease, pharmacological agents, and reversibility of injury [58], so when in doubt, one should wait. There is a need for better studies that are blinded, with proper sample calculation, without bias, and a better understanding of patients’ outcome, not only functional status, but also cognition and integration back into society.
Key Points
At the epicenter of the high mortality and morbidity from cardiac arrest, is the post-cardiac arrest syndrome, which includes anoxic brain injury, myocardial dysfunction, and a systemic ischemia and reperfusion syndrome.
The brain is prone to injury due to lack of significant intrinsic energy and nutrient stores, therefore highly dependent on a constant supply of oxygen and nutrients.
Multiple factors play a role in the extent and pattern of brain injury post-cardiac arrest including the initial ischemic cascade, the reperfusion injury after return of spontaneous circulation (ROSC), the delay ischemia due to the no reflow phenomenon, and post-resuscitation variable such as pyrexia and hypoglycemia.
Clinicians must use a multidisciplinary approach by combining experience with available clinical, neuroimaging, neurophysiological, and biomarkers data to counsel family regarding the prognosis of their loved ones.
Timing is key in prognostication as there are many confounding factors in early prognosis.