Updates on Management of Anoxic Brain Injury after Cardiac Arrest
Abstract
Brain injury is a leading cause of mortality and morbidity among cardiac arrest survivors. Management of these patients in the acute phase is challenging as is predicting their outcomes especially with the application of hypothermia. Therapeutic hypothermia has been proven beneficial but target temperature, timing, and duration that produce the best outcome are unclear and the subject of ongoing research. We review the recent advances in therapy and update the guidelines for management of these patients.
Introduction
More than 400,000 people suffer from cardiac arrest each year in the United States.1 Overall survival rates remain dismal: 22% in in-hospital cases and 10% in out-of-hospital cases, respectively.1 A significant cause of mortality is secondary to brain injury, which is a reflection of the brain’s intolerance to ischemia and its complex response to reperfusion.2, 3 Over the past decades, several studies have been performed to better understand the process, prognosis, and management of anoxic brain injury after cardiac arrest in the hopes of improving outcome. Our aim is to update the reader about management of anoxic brain injury in the acute setting in these patients.
Pathophysiology of Neuronal Injury
Cessation of cerebral circulation leads to depletion of oxygen stores within 20 seconds and manifests as loss of consciousness.4 At five minutes, glucose and oxygen are lost causing a disruption in ATP production and subsequent dysfunction of ATP-dependent membrane pumps, which maintain cellular membrane integrity.4, 5 This eventually leads to influx of calcium ions into the ischemic brain cells with the release of glutamate, an excitotoxic neurotransmitter, to NMDA receptors. Elevated intracellular calcium amplifies injury by activating secondary messengers that further promote calcium influx and interfering with the mitochondrial electron transport energy chain leading to formation of oxygen free radicals.5 These free radicals cause further cell damage that can trigger cell death.4–6 Secondary injury, which can occur hours or days after the initial event, results from other causes of oxygen compromise such as hypotension, brain edema and cerebrovascular dysregulation.2
Neurologic Manifestations of Anoxic Brain Injury
The hippocampus, cerebral cortex, thalamus, and cerebellum are regions especially susceptible to hypoxic injury and could cause clinical symptoms such as alteration of awareness, coma, seizures and strokes.2, 5 Long-term outcomes due to damage to thalamo-cortical networks or bilateral cortical regions can range from memory deficits to persistent vegetative state.2, 5 Involvement of the basal ganglia leads to movement disorders such as myoclonus, Parkinsonism, chorea and tics. The brainstem, on the other hand, is more tolerant of global ischemia; hence, the preserved function or early recovery of cranial nerves.2,5 Disruption of the autonomic centers such as the hypothalamus or its connections to cortex, subcortical regions or brainstem can result in paroxysmal autonomic instability with dystonia or PAID syndrome, which is characterized by agitation, diaphoresis, hypertension, tachycardia, tachypnea, and extensor posturing.7
Defining Neurologic Outcome
The Glasgow Outcome Scale (GOS) and the Cerebral Performance Category (CPC) are both five-point scales widely used in assessing neurologic outcome.8 In the GOS, a score of 5 is associated with good cerebral performance with mild or no disability. Lower scores are indicative of more severe disabilities with a score of 1 associated with death or brain death. In CPC, the scoring system is reversed where lower scores are more favorable. Poor outcome (death, coma or vegetative state) is defined as GOS of 1–2 or CPC of 4–5 according to established guidelines.8
Prognostication: Clinical Manifestations, Timing and Tools
Cardiac arrest can cause significant neurological injuries leaving patients cognitively disabled and in a dependent state. (See Figure 1.) Very few patients recover neurologically intact.9 Financial implications for families, insurers and health care systems are substantial.9 The direction of care is dependent on accurate prognostic information, which relies mainly on clinical examination and other ancillary tests. The prognostication algorithm as suggested by American Academy of Neurology can be affected by application of hypothermia in cardiac arrest patients due to delayed clearance of sedatives and paralytics, which can mask their neurological examination.9 Predicting prognoses early in these patients may lead to premature withdrawal of life-sustaining measures in about 20% of patients who would otherwise have good recovery.8 Neurological examination for prognostication purposes should be delayed for 72 hours after normothermia is achieved in the absence of sedation and paralysis. The presence of some prognostic signs such as myoclonic status epilepticus occurring within 24 hours after arrest or absence of all brainstem reflexes may justify earlier evaluation for prognostication than recommended 72 hours. 10
Reproduced from Nolan JP et al. Post-cardiac arrest syndrome: Epidemiology, pathophysiology, treatment, and prognostication. Resuscitation 2008; 79:350—379 with permission from Elsevier.
Brainstem Reflexes and Motor Response
The loss of all brainstem functions, even during the early phase of recovery, is associated with poor outcome.9, 10 The absence of bilateral pupillary light reflex 72 hours after cardiac arrest may be the most accurate predictor of poor outcome.9, 11 Further information can be obtained from checking corneal reflexes, although it may be less reliable if paralytic medication effects persist in the system. 11 The reliability of the presence or absence of other brainstem reflexes such as the vestibulo-ocular, cough and gag reflexes has not been well documented.10
The motor component of the Glasgow Coma Scale (GCS) also provides additional prognostic information. At 48–72 hours after cessation of sedatives and paralytics, a motor response that localizes to pain (GCS motor score ≥5) is suggestive of a favorable outcome while extension to painful stimuli (GCS motor score ≤2) is a poor prognostic sign.9, 11
Myoclonus and Seizures
Post-hypoxic myoclonic seizures are common following cardiac arrest. These are abrupt, irregular, muscle twitches that can be focal or generalized (involving the face, limbs, and trunk).9 These can be sporadic and have little prognostic value.8, 9 Status myoclonus, on the other hand, manifested as repetitive, unrelenting, generalized myoclonic seizures is highly associated with poor outcome even in patients with intact brainstem reflexes and good motor response, especially if it occurs within the first 24 hours of the cardiac event.8–11 The occurrence of early status myoclonus has decreased since the advent of therapeutic hypothermia. This is likely from concurrent use of sedatives and paralytics, which can mislead physicians during prognostication.10, 11
Tonic-clonic and focal epileptic seizures frequently occur during the re-warming process but do not carry the same negative prognostic implication as early myoclonic status.10
Electroencephalography (EEG)
The predictive value of EEG is limited since it can be affected by factors such as timing of recording and interference from sedatives and hypothermia.11 Seizures occur in 10–40% of cardiac arrest patients.8 Electrographic seizures consist of epileptiform discharges occurring repetitively and continuously for at least 10 seconds.11 Prolonged or recurrent electrographic seizures or electrographic status epilepticus (ESE), which can occur with or without clinical seizures is associated with poor outcome especially if it develops from a burst suppression pattern.11 Other patterns associated with poor outcome such as alpha-coma, burst suppression or generalized suppression have been reported.8
Besides seizure detection, EEG can also be used to assess nervous system reactivity. Reactivity is confirmed when there is a reproducible change in the amplitude or frequency of the EEG background after stimulation, such as eye opening, clapping, vocal sounds, touch or pain.8, 11 The absence of reactivity during and after hypothermia is strongly associated with poor outcome in some studies,8,11 but the evidence about EEG reactivity in post-anoxic coma is limited and there need to be further studies before it can be reliably used in clinical settings.8
Somatosensory Evoked Potentials (SSEP)
In contrast to EEG, SSEPs are not affected by hypothermia or sedation. It is a non-invasive test that involves electrical stimulation of a peripheral nerve. For purposes of prognostication, the median nerve is typically tested at the level of the wrist.10 In intact nervous systems, an electrical response (N20 potential) is recorded over the contralateral sensory cerebral cortex approximately 20 milliseconds after median nerve stimulation.8,10 Bilateral absence of N20 SSEP at 24 hours after cardiac arrest is one of the most accurate predictors of poor outcome with specificity nearing 100%.10 The use of SSEP is limited by its low sensitivity, unavailability in smaller institutions, and moderate inter-observer variation.8–10 Despite these limitations, a recent review reveals that among all predictors, absence of N20 SSEP had the most influence on decision-making, including withdrawal of life-sustaining therapies, among physicians and families.11
Bio-Markers
Two of the most extensively studied serum biomarkers of brain injury are neuron-specific enolase (NSE) and S-100B.8 Previous data suggested that a serum NSE level above 33μg/L at 24–72 hours after cardiac arrest is a good indicator of poor outcome.9 However, this cut-off value could not be confirmed in more recent studies, especially in patients treated with hypothermia. Higher NSE cutoffs (>50–80μg/L) are necessary to reliably predict poor outcome.8 S-100B levels are even more variable with cut-off values ranging from 0.2–1.5mg/L.8 The presence of these biomarkers in extra-cerebral sources is a major cause of variability.11 NSE-producing tumors (e.g. small-cell lung carcinoma and neuro-endocrine tumors) and hemolysis cause increased NSE levels.11 In addition, fat and muscle breakdown, which may occur during prolonged chest compressions, can cause erroneously elevated S100-B levels.8–10 Despite its ease of use and independence from sedative drug effects, the wide variability limits its role in decision-making.8,9,11
Neuro-Imaging
During the early phase of cardiac arrest, neuro-imaging results are typically not useful in prognostication as they may be normal or demonstrate subtle abnormalities.12 The purpose of obtaining a head computed tomography (CT) scan immediately after arrest is to exclude any primary brain injury that can result in cardiac arrest or coma. 8,9 The main CT finding of anoxic brain injury is brain edema, which appears as sulcal effacement and decreased grey-white matter differentiation.11,13 However, current data are insufficient to support its use as a prognostication tool in post-cardiac arrest.8
Magnetic resonance imaging (MRI) offers more promise in predicting outcome. Disruption of cell membranes due to oxygen depletion restricts cell membrane water diffusion and results in cytotoxic edema.9,14 This is responsible for the symmetrical grey matter hyper-intensities in diffusion-weighted imaging (DWI), which provides a qualitative measure of brain injury.9,12,14 Different patterns of brain injury relate to clinical outcome. They include isolated cortical injury, isolated deep grey nuclei injury (e.g. caudate nucleus, putamen, and thalamus), and mixed cortical and deep grey nuclei involvement.12,15 The mixed pattern correlates with the most severe form of injury.15
A standardized quantitative assessment of brain injury can be made by measuring the apparent diffusion coefficient (ADC).12 Patients with >10% of brain volume with a decreased ADC value between 650×10−6 mm2/s to 700×10−6 mm2/s had poor outcome (See Figure 2). 16 Most of the decrease in ADC values occur during a certain time period and involve specific brain regions.13, 14 The cortical grey matter was most profoundly affected between three to five days after arrest in patients who had a poor outcome.14 Favorable outcome was noted in those with increased diffusion in the occipital and temporal cortices, hippocampus, putamen, and corona radiata.14
(legend modified from original article)
The color ADC map shows pixels with an ADC value of 300–400 in red, 400–500 in orange, 500–600 in yellow, 600–700 in light green, 700–800 in dark green, 800–1200 in light blue, and 1200–2000 in dark blue.
Panel A: Twenty-two year-old man 64 hours after cardiac arrest who awoke after 3 days and had good neurological recovery at 6 months. Only subtle areas of restricted diffusion are seen in the thalami and cerebellum bilaterally.
Panel B: Thirty-seven year-old woman 52 hours after cardiac arrest who had impaired neurologic outcome. Note the widespread areas of restricted diffusion are seen in the cerebellum, thalami, basal ganglia, sub-cortical white matter, and cortex bilaterally.
Panel C: Sixty-six year-old man 67 hours after cardiac arrest who had poor outcome. Widespread areas of severely restricted diffusion are seen in the cortex, subcortical white matter, internal capsule, thalamus, basal ganglia, brainstem, and cerebellum bilaterally.
Reproduced from Wijman CAC et al. Prognostic Value of Brain Diffusion Weighted Imaging After Cardiac Arrest. Ann Neurol 2009; 65(4): 394–402 with permission from John Wiley and Sons.
More recent advances in brain MRI techniques are being utilized for prognostication among comatose survivors mostly for research purposes. Diffusion Tensor Imaging (DTI) and functional MRI are among emerging areas of investigation. These measure changes in microstructural and functional connectivity, respectively, with more disruptions associated with worse outcomes.17 As we understand these technological advances are understood further, it may be possible to use them to improve prediction of neurologic recovery and outcomes among patients with anoxic brain injury in the future.
Management of Anoxic Brain Injury
Therapeutic hypothermia (TH)
Induced hypothermia as a therapy for acute brain injury was first described by Fay in the 1940s.5 Subsequent use remained limited to small pre-clinical and clinical studies until the findings of Hypothermia After Cardiac Arrest trial conducted in Europe and a concurrent trial from Australia showed that maintenance of temperature between 32°C and 34°C for 12–24 hours increased survival as well as chances of favorable neurologic outcomes in patients with cardiac arrest due to ventricular fibrillation.18, 19
Following these two studies, the International Liaison Committee on Resuscitation (ILCOR) and the American Heart Association (AHA) published an interim scientific statement recommending the use of therapeutic hypothermia in comatose survivors of cardiac arrest.5 Since then, TH has been adopted as a standard of care for post-cardiac arrest cases including cases of pulseless electrical activity. It should be initiated as soon as possible with a target temperature ranging from 32–34°C.2,5 Rapid intravenous infusion of ice-cold 0.9% saline, external cooling blankets or pads with water-filled circulating systems, and intravascular cooling are some of the methods utilized to achieve TH.2,5
There has been much debate regarding the recommended temperature of 32°–34°C since similar results have been observed with milder cooling. A recent multi-center, randomized trial that compared target body temperature controlled at 33°C vs 36°C in patients who were resuscitated after cardiac arrest showed similar mortality rates and neurologic recovery in the two groups.21 The study, however, emphasized the importance of actively controlling the temperature and fever prevention during the first three days of cardiac arrest.
Optimal timing of induction of TH also remains unclear but the current consensus is to start cooling as soon as possible.2, 5 Whether benefits of TH could be gained by applying it in pre-hospital settings was the subject of a recent study in cardiac arrest patients. Results showed that pre-hospital cooling reduced core temperature by hospital arrival and reduced the time to reach a temperature of 34°C but did not improve survival or neurological status in these patients.22
Current guidelines recommend application of TH as part of post-resuscitation care although subtle differences in protocols exist. Target temperature, timing, duration, and method of cooling that leads to the best neurologic outcome are still unknown and the subject of on-going, and possibly future, research.
Fever Management
It has been well documented that hyperthermia exacerbates acute neurological injury and contributes to poor outcomes.23 Each degree over 37°C is correlated with an increased risk for severe disability, coma, or persistent vegetative state.5 Development of fever after cardiac arrest is common. However, fevers are less common in those treated with TH. Fever does not appear to carry the same detrimental consequences in TH-treated patients as compared to those not receiving TH.23 It is possible that the earlier hyperthermia occurs, the greater its consequences.23 Pre-clinical studies reveal that hyperthermia 24 hours after arrest, but not at 48 hours worsens brain injury, which suggests the time-dependent vulnerability of neurons to hyperthermia.23 In both treatment subgroups, antipyretics and surface or invasive cooling measures should be used aggressively to ensure that the body temperature is less than 38°C.5
Seizure Control and Prevention
Prophylactic anti-epileptic drugs (AEDs) are not used commonly in post-cardiac arrest patients. In those who were treated with hypothermia, seizures are commonly observed during the re-warming phase and weaning of sedation.5 An EEG should be performed on any patient who is suspected of having seizures, and those who do not regain consciousness after re-warming to exclude non-convulsive status epilepticus.5, 20 Standard AEDs should also be started but no comparative studies provide guidance as to choice of therapy.10 In cases of status epilepticus, anti-epileptic drugs are combined with sedatives, and the combination suppresses both clinical and electrographic seizures quite well. Data regarding suppression of status epilepticus using conventional AEDs alone are less convincing.10 Myoclonic status, on the other hand, is difficult to treat. Clonazepam is the most effective drug, but sodium valproate, levetiracetam, and propofol may also be effective.2
Cerebral Perfusion
Hypotension after ROSC should be avoided since it worsens cerebral ischemia.5 Cerebral perfusion can be further compromised by dysregulation of cerebral vasculature, which can occur in the acute phase of recovery.5 It may be necessary to maintain the mean arterial blood pressure (MAP) at higher levels to ensure adequate cerebral blood flow.5 A MAP greater than 65mmg Hg is adequate for coronary perfusion but maybe insufficient to provide adequate cerebral blood supply unless therapies are implemented to decrease cerebral metabolic activity, such as sedation and hypothermia.5 MAP values of 80 to 100 mm Hg have been suggested to be beneficial, at least for the first 24 hours after arrest.5
Neuroprotective drugs
Few neuroprotective drugs have been tested in clinical trials. There is inadequate evidence to recommend any pharmacological neuroprotective strategies to reduce brain injury in post-cardiac arrest patients.2
Conclusion
The core of the management in anoxic brain injured patients from cardiac arrest is prompt application of hypothermia in appropriate settings, treatment of seizures, hemodynamic maintenance, and supportive care. The previously established prognostic tools still hold, but TH has influenced the reliability and the timing of clinical evaluation and supportive tests. Prognostication should be done 72 hours after achievement of normothermia, in the absence of sedating or paralyzing drugs, to ensure that the information gathered is accurate in order to avoid premature withdrawal of life-sustaining measures in patients who would otherwise have good outcomes. In addition, certainty of predicting outcomes improves by utilizing a multi-modal approach to prognostication, where information from both clinical examination and supportive tests are taken together to guide decision-making. Lastly, certain areas of management such as target temperature, timing of initiation, and method of cooling require standardization and are currently under investigation.
Biography
Joanna Isabelle Ramiro, MD, is PGY-3, Housestaff Resident and Abhay Kumar, MD, is Assistant Professor; both are in the Department of Neurology and Psychiatry at Saint Louis University School of Medicine.
Contact de.uls@22ramuka
Footnotes
Disclsoure
None reported.
Facebook
Twitter
Google+