2006, Vol.24, Issues 1, Brain Injury and Cardiac Arrest

Neurol Clin 24 (2006) xi–xii

Foreword

Brain Injury and Cardiac Arrest

We use many terms to describe brain injury after cardiac arrest: globalcerebral ischemia (complete or incomplete), global ischemia, hypoxic braininjury, hypoxic-ischemic brain injury, and post–cardiac arrest syndrome, toname a few. Yet however we vary our lexography, the single literal transla-tion is a seriously injured brain without a treatment or good functional out-come. Great progress has been made in the treatment of cardiac arrest,particularly reanimating the heart and circulation; however, little haschanged with respect to ‘‘reanimating’’ or protecting the brain. Three de-cades beyond the discovery of cardiopulmonary resuscitation (CPR), pre-dicting neurologic recovery is based almost entirely on what we havelearned from the pathologic evaluation of brain tissue following cardiac ar-rest/CPR death. In the living patient, the clinician is taught the clinical ex-amination at the bedside to determine coma, prognosis, and brain deathstatusdtasks satisfactory for identifying the poor prognosis patient; butthere has been little effort or success developing predictive assessments to se-lect (or not select) treatment for the questionable-outcome patient, the onewho escapes brain death or emerges from a coma. In the past few years, slowbut thoughtful changes offer new opportunities for neurologists and othersinterested in the emergent care of acute brain injuries.

This issue of Neurologic Clinics offers an excellent window into the chang-ing outlook of brain injury after cardiac arrest. Romer Geocadin has assem-bled a group of physician-scientists who are improving the clinical picture ofbrain function after cardiac arrest. It is clear that, within the first decade ofthe twenty-first century, we will have the experience to assess the comatosecardiac arrest survivor and the know-how to optimize return to brain health.Within this time, we can look forward to guidelines for optimal return ofcirculatory function; brain tissue protection; early prognostication; earlymonitoring of recovery; and a detailed mapping, by degrees of cardiac arrestand extent of brain injury, of the functional and neuropsychologic patientconsequences. The data presented herein offer a compelling argument forthe greater involvement of all physicians to learn cardiac arrest treatmentsand interventions directed to brain recovery within the minutes and firsthours after cardiac arrest. For all the patients, families, and physicians treat-ing cardiac arrest, this is great news.

0733-8619/06/$ - see front matter � 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.ncl.2005.12.003 neurologic.theclinics.com

xii FOREWORD

We thank Dr. Geocadin and his colleagues for opening the door to a bet-ter way of doing things.

Daniel F. Hanley, MDDepartments of Neurology

Neurosurgery and Anesthesiology–Critical Care MedicineJohns Hopkins University School of Medicine

600 North Wolfe Street–Meyer 8-140Baltimore, Maryland 21287, USA

E-mail address: [email protected]

mailto:[email protected]

Neurol Clin 24 (2006) xiii–xvi

Preface

Brain Injury and Cardiac Arrest

Guest Editor

The importance of brain injury after resuscitation from cardiac arresthas been recognized since the days of the pioneering works of Peter Safarin 1958, at the Baltimore City Hospital (currently Johns Hopkins BayviewMedical Center) on airway methods of artificial respiration and Koewen-hoven, Jude, and Knickerbocker at the Johns Hopkins Hospital in the early1960s on the closed cardiac massage with external defibrillation. This gaveway to the development of modern cardiopulmonary resuscitation. In their2000 Guidelines for Cardiopulmonary Resuscitation and Emergency Cardio-vascular Care, the American Heart Association highlighted brain injury byproviding that ‘‘The cerebral cortex, the tissue most susceptible to hypoxia,is irreversibly damaged, resulting in death or severe neurological damage.The need to preserve cerebral viability must be stressed in research endeavorsand in practical interventions.’’

Advances in cardiopulmonary resuscitation and critical care contributedto the increasing success in patient survival; however, a significant numberof survivors live with poor neurologic function. Numerous neuroprotectionclinical trials undertaken in the last three decades failed to show functionalbenefit to survivors until the recent success of therapeutic hypothermia. Thistherapy shows that brain injury can be ameliorated, leading to improvedsurvival and functional outcome in survivors. Although much of the re-search in this field has been undertaken by cardiologists, emergency medi-cine physicians, anesthesiologists, and intensivists, injury to the brainrequires more involvement by neurologists.

Romergryko G. Geocadin, MD



xiv PREFACE

The clinical practice of neurology in this area is also limited, mainly in thecare of post–cardiac arrest complications and prognostication of outcome.After cardiac arrest, the neurologist is typically consulted to evaluate unre-sponsive patients only several days after medical stabilization. The neurol-ogist may be called earlier if the possibility of seizures or stroke isentertained. The key questions are: When should the neurologist get in-volved? What can the neurologist offer other than prognostication? Withthe onset of brain ischemia, complex cascades of events amplify the ische-mic injury during the first few hours. Early involvement of the neurologistduring the very acute and early recovery period after cardiac arrest maygreatly enhance multidisciplinary strategies on brain preservation and func-tional recovery. My experience performing both research and patient care inthis area has been very encouraging. I noted that medical, critical care, andemergency medicine specialists know that the neurologist has much to con-tribute with regard to care for brain injury and research to further improvesurvival and quality of life. Despite the best efforts of specialists, irreversiblebrain injuries still occur in survivors that will lead to long-term morbidities.With the early recognition of the potential long-term neurologic problems,the neurologist can play a significant role in improving the quality of life ofsurvivors.

This issue of Neurologic Clinics revisits brain injury after resuscitationfrom cardiac arrest. The goal of this issue is to provide the neurologist witha comprehensive and multidisciplinary review of the current research andclinical practices related to brain injury and cardiac arrest. It is my hope thatthe recent advances in cardiac arrest resuscitation and acute neurology willtransform the traditional role of the neurologist as a diagnostician to a neu-roclinician who can work directly with other services to provide acute inter-ventions that will ameliorate acute brain injury. A similar change has beensuccessfully undertaken with the acute neurologic interventions in the areasof acute stroke and neurointensive care. It is also my hope that this issue willspark more interest in neurologic research in the area to further facilitate thetransformation.

As a comprehensive and multidisciplinary review, this issue is a collec-tion of articles written by leading experts. The first article by Drs. Bhardwajand Harukuni, renowned researchers in ischemic brain injury, does notlimit itself to the mechanism of brain injury but provides therapeutic con-sideration related to the injury mechanisms. The next article takes thereader to the event immediately around the cardiac arrest. As leaders inthe field of emergency medicine, Drs. Ornato and Peberdy provide neurol-ogists with critical insights into resuscitation and neuroprotection out inthe field and in the emergency department. From the emergency depart-ment, the patient moves to the intensive care area. The article by Dr.Schulman and colleagues on the intensive care of these patients will pro-vide a joint cardiac and neurologic critical care approach to post–cardiac

xvPREFACE

arrest patients. After successful resuscitation, consideration for therapeutichypothermia has to be made for all appropriate patients. Dr. StephenBernard, the lead author of one of the landmark clinical studies on therapeutichypothermia, provides an authoritative review. While therapeutic hypother-mia may benefit some patients, many survivors will still have a poor outcome.The following article by Drs. Popp and Bottiger provides a look into thedevelopment of very promising therapies in the area. They have highlightedin their article the clinical trial in Europe, led by Dr. Bottiger, using throm-bolytics to improve functional recovery after cardiac arrest.

The succeeding articles provide updates on the traditional and novel ap-plications of diagnostic tools that are available to enhance the care of thesepatients. The article by Dr. Koenig and colleagues reviews the role of neuro-electrophysiologic tests in prognostication and provides novel approachesand applications of electrophysiologic testing to enhance not only the prog-nostication but also the early detection of brain injury and monitoring of re-covery. The article by Drs. Geraghty and Torbey reviews the serologicalmarkers of brain injury and provides novel applications of neuroimagingto enhance the prognostication.

As the patient moves out of the intensive care unit and the hospital, long-term neurologic complications become the focus of neurologic care. Move-ment disorder experts Drs. Venkatesan and Frucht address the clinicalproblems and the need for further research in the area of movement disor-ders after cardiac arrest. With the more advanced research on the cognitiveand behavioral dysfunction in patients who have undergone cardiac bypassprocedures than in survivors of cardiac arrest and the numerous similaritiesof the two global ischemic conditions, the article by Dr. Selnes and col-leagues provides critical insight into the cognitive and behavioral dysfunc-tion after a global ischemic injury.

The last two articles deal with special topics in this area. In many diseasestates, children need special consideration. The article by Dr. Hickey, a pedi-atric emergency physician, and Dr. Painter, a pediatric neurologist, tacklesissues of brain injury after cardiac arrest in children. With death as a com-mon outcome even in resuscitated patients, the article by Drs. Manno andWijdicks provides approaches to withdrawal of life-sustaining therapiesand the declaration of brain death in neurologic patients.

Many people contributed to this issue. My gratitude goes to the con-tributors, whose dedication to research and patient care has greatlymoved the field forward. I would like to express my deepest gratitudeto Dan Hanley and Nitish Thakor, who showed me that we can help crit-ically ill patients and their seemingly insurmountable clinical problemsthrough research, and to Peter Safar, whose enthusiasm and brilliance en-couraged me to proceed in this path. I would also like to thank the Neu-rologic Clinics Editors, Don Mumford, Bob Gardler, and their staff fortheir patience and assistance. And most importantly, to my wife Effie

xvi PREFACE

and our children, Ginno and Sofia, whose love and inspiration made allthis possible.

Romergryko G. Geocadin, MDDepartments of Neurology, Neurosurgery and

Anesthesiology-Critical Care MedicineJohns Hopkins University School of Medicine

600 North Wolfe Street, Meyer 8-140Baltimore, MD 21287 USA



Neurol Clin 24 (2006) 1–21

Mechanisms of Brain Injuryafter Global Cerebral Ischemia

Izumi Harukuni, MDa, Anish Bhardwaj, MDb,*aDepartment of Anesthesiology and Critical Care Medicine,

Johns Hopkins University School of Medicine, Division of Cardiac Anesthesiology, Tower 711,

Johns Hopkins Hospital, 600 N. Wolfe Street, Baltimore, MD 21287, USAbDepartments of Neurology and Anesthesiology and Critical Care Medicine,

Johns Hopkins University School of Medicine, Neurosciences Critical Care Division,

Meyer 8-140, Johns Hopkins Hospital, 600 N. Wolfe Street, Baltimore, MD 21287, USA

Global cerebral ischemia occurs commonly in patients who have a varietyof clinical conditions including cardiac arrest (CA), shock, and asphyxia andin patients undergoing complex cardiac surgery [1–4]. In addition to injuryto other organs from systemic hypoperfusion, neurologic sequelae frombrain injury are varied and constitute a spectrum that includes coma, sei-zures, ischemic stroke, delirium, and neurocognitive impairment [5–7].The commonest postulated mechanism for ischemic brain injury after CA(with subsequent resuscitation) is global cerebral ischemia from systemichypoperfusion that can occur with or without pre-existing large-vesselocclusive disease. Embolism that arises from the heart, from aortic arch ar-theromas, or from extracorporeal circulation devices occurs more commonlyin the perioperative period following complex cardiac surgery and less com-monly during resuscitation following CA [7]. Irrespective of the etiology ofcerebral ischemia, cellular and molecular processes trigger a cascade ofevents that culminate in a ‘‘final common pathway,’’ resulting in ischemicneuronal injury. Identification of these injury mediators and pathways ina variety of experimental animal models of global cerebral ischemia hasled to investigation of target-specific cytoprotective strategies that are criti-cal to clinical brain injury outcome. Although the authors have previouslypublished on this subject [8], this article expands on the translational sig-nificance of many of the potential neuroprotective strategies that have

This work is supported in part by the US Public Health Service National Institutes of

Health grant NS046379.

* Corresponding author.

E-mail address: [email protected] (A. Bhardwaj).




2 HARUKUNI & BHARDWAJ

provided important insights into the mechanism or mechanisms of ischemicbrain injury and might be of therapeutic benefit in the future.

Global versus focal cerebral ischemia

Ischemia is defined as diminution of cerebral blood flow (CBF) to a crit-ical threshold that propagates brain damage involving the entire brain ora selective region. Global cerebral ischemia entails diminution in CBFover the entire brain, encountered clinically as sequelae during extracorpo-real circulation following CA from ventricular fibrillation or asystole thatlasts 5 to 10 minutes. Global ischemia from CA results in a predictable pat-tern of histologic injury in which specific neuronal populations are affected(selective ischemic necrosis) [9–11] (discussed later). Although reperfusionrestores CBF, it can lead to secondary brain injury from influx of neutro-phils and to increases in reactive oxygen species (ROS), cerebral edema,and hemorrhage. Elevated levels of ROS may lead to damage of intracellu-lar proteins and DNA by way of oxidation and by activating a number ofpathways that lead to cell death. Unlike global cerebral ischemia, focal ce-rebral ischemia entails reduction in regional CBF in a specific vascular ter-ritory and is usually encountered clinically as an ‘‘ischemic stroke’’ due tothromboembolic or artherothrombotic vaso-occlusive disease. NormalCBF ranges from 50 to 75 mL/100 g of brain tissue per minute and can differbetween the white and gray brain matter. Ischemic depolarization occurswhen CBF decreases to levels of approximately 18 mL/100 g of brain tissueper minute, and neuronal cell death ensues if CBF is less than 10 mL/100 gof brain tissue per minute. In focal ischemia, the ischemic vascular bed com-prises an area with severe CBF reduction that consists of an ‘‘ischemic core’’and a more distal ‘‘ischemic penumbra’’ and includes regions that are mar-ginally perfused and might be served by collateral vascular channels. Histo-pathologic outcome following focal ischemia largely depends on ischemicseverity and duration [8–10]. Increasing durations of depolarizing ischemiaare associated with a spectrum of histopathologic correlatesdfrom revers-ible injury to irreversible cerebral infarction.

Experimental models of global cerebral ischemia

Several animal models have been developed to simulate complete humanglobal cerebral ischemia and have provided histologic evidence of and in-sight into mechanisms of brain injury. Rodent models (gerbil, mouse, andrat) provide the advantages of being inexpensive, of rendering consistent re-producibility of injury because they possess consistent cerebral vasculature,and of being homogeneous among strains, with transgenic counterparts al-lowing targeted mechanistic studies for delineating effects of specific genedeletion on ischemic brain injury. Monitoring of critical physiologic varia-bles (pH, arterial blood pressure, arterial blood gases, core body and cranial

3BRAIN INJURY AFTER GLOBAL CEREBRAL ISCHEMIA

temperature, plasma glucose levels) during the peri-ischemic period, how-ever, is suboptimal. Furthermore, assessment of functional neurologic out-comes with a comprehensive behavioral battery presents a challenge in smallanimal models. Because 50% of Mongolian gerbils possess an incompletecircle of Willis, bilateral carotid artery occlusion results in forebrain ische-mia in these animals. The rat model of global cerebral ischemia includes(1) four-vessel occlusion (4-VO) by electrocoagulation of both vertebral ar-teries, with transient occlusion of both carotid arteries 24 hours later [12]; or(2) ‘‘two-vessel occlusion’’ by transient occlusion of both carotid arteriesand accompanying reduction of arterial blood pressure to a level of 40 to50 mm Hg by phlebotomy [13,14]. Both techniques yield high-grade ische-mia of forebrain structures [15]. Larger animal models (rabbit, canine,feline, swine, nonhuman primates) allow for monitoring of critical physio-logic variables during and following the ischemic insult and allow fora more accurate delineation of neurologic deficits following the insult; how-ever, high cost and significant ethical concerns limit their use. The methodsof brain injury in these larger-animal models include ventricular fibrillation[16], aortic occlusion [17], brachiocephalic or subclavian arterial occlusion incombination with hypotension, elevated intracranial pressure by neck-cuffinsufflation, and intraventricular infusion [18]. In these animal models,specific neuronal populations in the brain, including CA1 pyramidalneurons of the hippocampus, medium-sized neurons of the striatum, andthe Purkinje cells of the cerebellum [19], are susceptible to injury. The sen-sitivity of these neuronal populations is varied and dependent on durationand severity of ischemia, and a typical temporal profile is observed followingCA. Neurons in the CA1 zone are the most sensitive to depolarizing ische-mia (3–5 minutes), whereas the medium-sized neurons of the striatum aremore resistant (15–20 minutes) [20]. Following successful resuscitation andcerebral recirculation, progression of irreversible neuronal injury also differsin these selective neuronal populations (eg, neuronal injury is observedwithin 3 hours of establishing recirculation in medium-sized neurons ofthe striatum but delayed up to 48–72 hours in CA1 hippocampal neurons,a phenomenon referred to as ‘‘delayed neuronal death’’) [20].

Neuronal injury mechanisms: apoptosis versus necrosis

Over the past decade, research has demonstrated that consequences of ce-rebral ischemia result in two temporally distinct phases or processes of neu-ronal cell death, which in turn affect surrounding brain tissue. Each phasehas characteristic defining morphologic and molecular features, and the dis-tinction between the two processes is based on morphologic findings on elec-tron microscopy [8,21]. Apoptosis or programmed cell death, a processassociated with genomic fragmentation, is characterized by cell shrinkage,chromatin aggregation, and preservation of cell membrane integrity and mi-tochondria without inflammation and injury to surrounding tissue [21–25].


Conversely, necrosis, a process that is not ‘‘regulated or programmed,’’ istypically observed as a consequence of severe cerebral ischemia and charac-terized by disruption of cellular homeostasis from energy failure due to se-vere mitochondrial injury, which leads to cellular swelling, membrane lysis,inflammation, vascular damage, and edema formation. Although apoptosisand necrosis characteristically represent distinct modes of cellular injury,a large body of literature suggests that these processes represent a spectrumthat coexists in a spatial distribution within injured tissue (neurons in thecore being necrotics and neurons in the penumbra being apoptotic) and istemporally related to duration and severity of the ischemic insult. The ac-cepted tenet is that the excitotoxic cascade triggered by exposure to excitatoryamino acids (EAAs) plays a prominent role in neuronal necrosis. In vitroexposure to glutamate, however, results in apoptosis in neurons that survivethe early necrotic phase and have partial recovery of mitochondrial function[26]. Therefore, maintenance of mitochondrial function is possibly the crit-ical factor in determining the degree and progression of neuronal injurypropagated by excitotoxicity [25]. A common pathway of stimuli leadingto apoptosis has not yet been identified; however, a mitochondrial-depen-dent intrinsic pathway and a receptor-mediated extrinsic pathway are postu-lated [27,28]. In the intrinsic pathway, cerebral ischemia results in generationof a permeability transition pore in the inner mitochondrial membrane thatleads to disruption of the outer mitochondrial membrane due to release ofseveral proapoptotic factors (caspases, endonucleases, cytochrome c, andother proteases related to interleukin-1b converting enzyme) [21,27,28].These events ultimately lead to DNA fragmentation. The ‘‘mitochondrial-independent’’ or extrinsic pathway of apoptosis involves several death re-ceptor families such as Fas. Bypassing the inhibitory effects of Bcl2-relatedproteins, which control proteolytic systems, has also been postulated to playan important role in apoptotic cell death, whereas other family members,such as Bax, augment apoptosis. Another family of protein-cleaving en-zymes, caspases, is expressed at significant levels during cerebral ischemia,and caspase inhibitors produce resistance to ischemic damage [29].

Excitotoxic brain injury

The concept of excitotoxicity, introduced by Olney in 1969 [30], was basedon a set of observations that included neuronal injurywith local application ofglutamate and other acidic amino acids (aspartate, N-methyl-D-aspartate[NMDA], homocysteine, cysteine). Since this descriptionwas published, otherexcitotoxic mediators have been delineated, including catecholamines (dopa-mine, norepinephrine), nitric oxide (NO), and related species. Glutamate, themost abundant EAA in the brain, serves a variety of important functions(metabolic, neurotrophic, and neurotransmitter) and is compartmentalizedin neurons [19,31]. The healthy adult brain has the ability to clear extracellularglutamate by rapid uptake; however, under conditions in which energy stores


are depleted, such as in hypoxia-ischemia, glutamate efflux into the extracel-lular compartment due to cellular depolarization [32], coupled with its im-paired uptake, results in increases in intracellular Ca2þ. Interference withcysteine uptake (causing depletion of cellular glutathione stores responsiblefor protecting against oxidative stress) results in neuronal injury [30,33]. Ex-citotoxic injury is characterized by its maximal effects on neuronal dendritesand soma, with relative sparing of axons, glia, and ependymal and endothelialcells, possibly due to differences in synaptic input, density, distribution ofmembrane glutamate receptors, and intrinsic defensemechanisms.Glutamateactivates three major families of ionophore-linked receptors (NMDA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid [AMPA], and kainate)and metabotropic receptors that activate second messenger systems [34]. Al-though glutamate release simultaneously stimulates NMDA and AMPA re-ceptors, in vitro studies demonstrate that glutamate toxicity occurs in twodistinct phases: (1) excitotoxicity is rapidly triggered by brief intense stimula-tion of NMDA receptors, which is critically dependent on the presence andinflux of extracellular Ca2þ through the NMDA-gated receptor channel com-plex; and (2) a slowly triggered process by the prolonged stimulation ofAMPA/kainate receptors that have limited Ca2þ channels [33]. Metabotropicglutamate receptors modify excitotoxic injury, rather than directly mediatethe deleterious process.

The cascade of events responsible for glutamate excitotoxicity includesthree distinct processes: (1) induction, whereby extracellular glutamate effluxis transduced by receptors on the neuronal membrane to cause intracellularCa2þ overload, which leads to lethal intracellular derangements; (2) ampli-fication of the derangement, with an increase in intensity and involvement ofother neurons; and (3) expression of cell death triggered by cytotoxic cas-cades [33]. Excess release of Ca2þ and its intracellular influx is thought tobe the primary trigger for a variety of complex, deleterious intracellular pro-cesses that result from activation of catabolic enzymes such as phospholi-pases (which lead to cell membrane breakdown, arachidonic acid, and freeradical formation) and endonucleases (which lead to fragmentation of geno-mic DNA and energy failure due to mitochondrial dysfunction) (Fig. 1).Distinct neuronal populations are selectively vulnerable to excitotoxicinjury, possibly from differences in excitatory synaptic inputs, density ofglutamate receptors, or intrinsic defense mechanisms. Perhaps the most com-pelling evidence for the role of glutamate excitoxicity following focal ische-mia and hypoxic-ischemic brain injury from CA is the neuroprotectionobservedwith antiexcitotoxic strategies includingNMDA-orAMPA-receptorantagonists [35].

Acute efflux of dopamine and norepinephrine into the extracellular spacefollowing cerebral ischemia [36–39] possibly plays a role in the propagation ofbrain injury. Indirect evidence of the importance of their role in cerebral is-chemia stems from amelioration of histopathologic injury following attenua-tion of ischemia-induced surges in extracellular dopamine pharmacologically


with barbiturates, isoflurane, and etomidate [40,41]. Furthermore, severallines of evidence suggest that catecholamine release andmetabolismmight un-derlie the selective vulnerability of striatal neurons to an ischemic insult[42,43]. For example, depletion of catecholamine stores by a-methyl-para-tyrosine exerts a strong protective effect on ischemic damage to nerveterminals [43], and reduction of striatal dopamine content by lesioningthe nigrostriatal tract protects intrinsic striatal neurons from injury follow-ing global cerebral ischemia [44]. Although the precise mechanism of neuro-nal injury by dopamine is unclear, by-products of its metabolism, such ashydrogen peroxide, superoxide ion, and hydrogen radicals, have been impli-cated in this deleterious process [45].

NO, a free radical gas synthesized from the amino acid L-arginine by theenzyme NO synthase (NOS), is produced by a variety of sources (vascularendothelium, neurons, glia, macrophages, white blood cells) [46,47]. NOhas several functions in the brain, including regulation of CBF, neurotrans-mission, and modification of inflammation [46,47]. At least three isoforms ofNOS have been identified: the constitutively expressed neuronal and endo-thelial isoforms (encoded on chromosome 12 and 7) and the inducible iso-form (encoded on chromosome 17) [48,49]. Neuronal NOS–containingneurons are widespread in the brain, including in the cerebral cortex, hippo-campus, and striatum [47]. NO plays a dual role in ischemic neuropathol-

Decreased cerebral perfusionDepletion in energy stores

Cerebral Ischemia

Apoptosis Excitotoxicity Inflammation (Programmed Cell Death)

EAAs (Glutamate, Aspartate),Catecholamine

Activation of Metabotropic receptors

Influx of intracellular Ca2+

Protein phosphorylation ProteolysisNO, oxygen free radical productionAltered gene expression

Neuronal Cell Death

Cytokine (IL-1, TNF- ) productionChemokinesAdhesion molecules (selectins, integrins, ICAM-1)ProteasesBBB breakdown

Activation of ionophore-linked channels(NMDA, AMPA)

Modification

Cell shrinkageChromatin aggregationPreservation of cell membraneintegrity Preservation of mitochondriaLack of inflammation

Fig. 1. Schematic diagram representing events leading to ischemic brain injury.


ogydbeneficial, in that it is a potent vasodilator, and cytotoxic, in that itinhibits important enzyme systems such as complexes I and II of the mito-chondrial transport chain (Fig. 2). Formation of the oxidant perioxynitriteby combination of NO and superoxide anion is considered to be an impor-tant trigger in cytotoxicity. Perioxynitrite generation leads to formation ofother ROS, hydroxyl free radicals, and nitrogen dioxide, resulting in ni-trosylation of tyrosine residues in proteins. One way by which NO isthought to kill neurons is energy-dependent activation of the DNA repairenzyme poly(ADP-ribose) polymerase, leading to consumption of ATP,nicotinamide, and cell death [50].

In cerebral ischemia, constitutive NO activity (endothelial and neuronalisoforms) markedly increases as a consequence of NMDA, AMPA, and me-tabotropic glutamate receptor stimulation [51,52] culminating in a rise in in-tracellular Ca2þ, whereas more sustained levels of NO are expressed by themicroglia and other inflammatory cells (Ca2þ-independent inducible isoform)24 hours following the ischemic event [8,48]. Consequently, administration ofNO donors and specific NOS inhibitors has yielded mixed results in experi-mental models of cerebral ischemia. Such varied results are explained by non-selectivity of NOS inhibitors and timing of such interventions in relation toischemia. For example, infusion of the NO precursor and substrate L-argininein the immediate period following experimental focal ischemia attenuates in-farct volume by accentuating CBF byway of dilation of pial vessels [48,53]. Inline with this evidence, male mutant mice that lack the endothelial isoform ofNOS sustain larger ischemic injury than their wild-type counterparts follow-ing focal ischemia [54]. Infarct volume and functional outcomes are improvedin wild-type mice following selective pharmacologic inhibition of neuronalNOS and in mutant mice that do not express genes for the neuronal or induc-ible isoforms [48,55]. Because vascularNOmight favorably affect outcome andthe neuronal isoform might adversely affect outcome, selective pharmacologic

Nitric Oxide

Beneficial Role (Endothelial) Deleterious Role (Neuronal and Inducible)

Direct Cytotoxic EffectsBinding to iron-sulfur complexesDecreased mitochondrial respiratory enzymesDecreased ribonucleotide reductase

Nitrosylation of thiols and ADP ribosylation

Indirect Cytotoxic EffectsIncreased peroxynitriteIncreased hydroxyl radical

Increased CBF

Decreased NMDA current

Decreased platelet aggregationDecreased platelet adhesion

Decreased aconitase

Decreased PARP, leading to DNA damage

Increased nitrogen dioxide

Fig. 2. Beneficial and deleterious effects of nitric oxide in cerebral ischemia. (Reproduced from

Bhardwaj et al. [8] and adapted from Dalkara and Moskowitz [49], with permission.)


inhibitionof theneuronal and inducible isoformsofNOSmightprovide aviabletherapeutic strategy in cerebral ischemia.

Role of inflammation

Cerebral ischemia leads to inflammatory cell infiltrates from nonspecificimmunologic reaction, migration of peripheral leukocytes into the brain,and activation of microglia [56]. Release of inflammatory cytokines (inter-leukin [IL]-1, tumor necrosis factor a [TNF-a]) by ischemic neurons andglia leads to generation of adhesion molecules (selectins, integrins, intercel-lular adhesion molecule 1) in the cerebral vasculature, which results inbreakdown of the blood-brain barrier (BBB), culminating in edema forma-tion [57,58]. Enhanced secretion of cytokines and proteases such as metallo-proteinases causes further disruption of the extracellular matrix and theBBB. Although IL-1 is detrimental to ischemic brain injury, roles for IL-6,a proinflammatory cytokine, and IL-10, an anti-inflammatory cytokine,are less clear. TNF-a appears to have a dual role in the ischemic brain, inthat it is involved in ischemic tolerance [59] and plays a role in propagatingischemic brain injury [60,61].

Glycemic control

A number of studies in animal models of traumatic brain injury [62], focalcerebral ischemia [63], and global cerebral ischemia [64] demonstrate that gly-cemic control is a critical factor in terms of outcome. Postulated mechanismsinclude accentuation of release of EAAs, attenuation of neuroinhibitory neu-rotransmitters [65], massive deposition of neutrophils [66], and early mito-chondrial damage by way of activation of cytochrome c, caspase-9, andcaspase-3 cleavage [67]. Thesemechanistic studies have led to clinical observa-tions that poor glycemic control accentuates brain injury in ischemic stroke[68] and following cardiac surgery [69]. Glycemic control with insulin treat-ment has been demonstrated to improve neurologic outcome in critically illpatients [70] and in patients who undergo cardiac surgery [69]. Although insu-lin therapy has been shown to ameliorate damage in animal models of globalcerebral ischemia [71], further clinical trials are warranted in the setting ofaggressive glycemic control with insulin therapy in patients following CA.

Role of temperature

Experimental studies using animal models of focal and global cerebral is-chemia have provided evidence for the importance of brain temperature onfunctional and histopathologic outcome [72]. Following cerebral ischemia,intraischemic hyperthermia leads to incomplete normalization of high-energy phosphate metabolites and the conversion of selective neuronal ne-crosis to infarction, increased microvascular injury, and edema, resultingin increased mortality [72]. Spontaneous elevations in body temperature


have been reported following experimental global and focal ischemia andare thought to be a consequence of brain injury [73]. Mild (34�C) tomoderate (30�C) induced systemic hypothermia markedly attenuates ische-mic brain injury following experimental CA [74]. Mechanisms of hypother-mia-induced ischemic brain protection may be multifactorial and includemitigation of pre- and postsynaptic excitotoxic processes (attenuation ofbiosynthesis, release and uptake of EAAs), diminished hydroxyl radical pro-duction, protection of lipoprotein membranes, attenuation of intracellularacidosis, and reduction of oxygen demand by the injured brain [2]. Recentclinical trials have demonstrated improved neurologic outcome and de-creased mortality in patients subjected to mild-to-moderate therapeutichypothermia [1,2]. Future clinical trials should incorporate other pharmaco-logic neuroprotective strategies in combination with hypothermia to furtherenhance outcomes.

Ischemic tolerance and preconditioning

The concept of ischemic tolerance, introduced 2 decades ago and based onobservations in the myocardium [75], was extended to ischemic brain injury,whereby brief ischemic insults protected the brain from subsequent andmore severe ischemia [76]. Further experiments in a variety of animal modelsof focal and global cerebral ischemia confirmed these observations [77,78]. Inaddition to sublethal ischemia, other conditions such as hyperthermia [79], hy-pothermia [80], hypoglycemia [81], and pharmacologic agents (eg, antibiotics,erythropoietin, acetylsalicylic acid, volatile anesthetics) [82–85] have beenshown to induce ischemic tolerance. The early phase of ischemic tolerance(within 30minutes following sublethal insult) is thought to be due to flow-me-tabolism–mediated events, whereas delayed tolerance (O24 hours) involvesnew gene induction and protein synthesis [86,87]. Molecules such as adeno-sine, hypoxia inducible factor 1a, TNF-a, ROS, NO, and other receptor-linked events involving NMDA-receptor activation and downstream effectsof intracellular calcium influx have been implicated in ischemic tolerance. Al-though the precisemechanisms of ischemic tolerance have not been elucidatedcompletely, ischemic preconditioning provides a possible venue and therapeu-tic strategy in ameliorating brain injury in a few, select high-risk patients sus-ceptible to ischemic brain injury.

Experimental pharmacologic neuroprotection and its translational

significance

A comprehensive review of this topic is beyond the scope of this article;however, the interested reader is referred to a recent monograph by Weigland colleagues [88]. Several pharmacologic agents have undergone investiga-tion in animal models of global cerebral ischemia to directly or indirectlydetermine efficacy. Methods used to evaluate postischemic hyperemia or hy-poperfusion include recovery of somatosensory-evoked potentials, recovery


of high-energy phosphates, functional neurologic recovery, and histologicinjury [29,88]. The basis for the study of these agents is a logical extension ofmechanistic studies in a variety of in vitro and in vivo systems of cerebral is-chemia. In addition, results from neuroprotective studies with pharmacologicagents provide important insights into mechanisms because they entail dis-ruption of a specific pathway or receptor blockade in the propagation of ische-mic neuronal injury. Although many of these agents are still underinvestigation, data predominantly from studies using the focal ischemia par-adigm underscore the future use of pharmacologic therapies following globalcerebral ischemia.

N-methyl-D-aspartate–receptor antagonists

Pre- and post-treatment with dextrorphan, an NMDA-receptor antago-nist, improves histologic injury in the hippocampus and the cortex in a4-VO rat model and attenuates the reduction in loss of activity of calcium-dependent protein kinases (protein kinase C and calcium-dependent proteinkinase II) [89]. Although the NMDA-receptor antagonist dizoclipine(MK801) has been shown to provide significant histologic neuroprotectionin animal models of global cerebral ischemia [90,91], its clinical use in ische-mic stroke has been shown to produce significant undesirable side effects(delirium, psychosis, hallucinations, and so forth) [29].

Calcium channel antagonists

Because Ca2þ is the final common pathway in excitotoxic neuronalinjury, nimodipine, a blocker of Ca2þ influx, has been studied in the exper-imental paradigm of global cerebral ischemia. Subcutaneous administra-tion of nimodipine failed to demonstrate any histologic or functionalneurologic improvement in the 4-VO rat model of global cerebral ischemia[92,93]; however, in a rabbit model, intravenous treatment with nimodipinereduced time of EEG recovery and attenuated the decrement in extracellu-lar Ca2þ and disruption of the BBB. In this study, arterial blood pressurewas maintained at 100 mm Hg following the ischemic insult, thereby off-setting the detrimental hypotensive effects of nimodipine. A prospective,randomized, double-blinded trial with nimodipine in patients who hadout-of-hospital ventricular fibrillation failed to demonstrate any improve-ment in 1-year survival rate; however, it demonstrated some benefit in pa-tients who had delayed resuscitation (O10 minutes) [94].

g-Aminobutyric acid and g-aminobutyric acid agonists

The premise of using g-aminobutyric acid (GABA) or its agonists as neu-roprotectants is based on their inhibitory properties by way of opening ofthe Cl� channels [88]. Pretreatment with GABA attenuated histologic injuryand improved neurobehavior in a gerbil model of global cerebral ischemia


[95]. Treatment following the insult failed to demonstrate any improvementin these parameters. Chlormethiazole, a GABA agonist with anticonvulsant,hypnotic, and sedative properties, failed to demonstrate any improvement inhistologic injury or neurobehavior in a rat model of global cerebral ischemia[96]. Furthermore, local infusion of chlormethiazole by way of microdialysisdid not alter ischemia-evoked release of dopamine, serotonin, or their me-tabolites in the ischemic striatum [96]. Intraperitoneal administration ofg-hydroxybutyrate improved histologic injury and neurobehavioral out-comes in a 4-VO rat model of global cerebral ischemia [97]. Tiagabine, a se-lective inhibitor of GABA reuptake, failed to demonstrate any improvementin histologic outcome in the gerbil model when given as a pretreatment [98].

Anticonvulsants

The basis for the use of anticonvulsants in ischemic neuroprotection istheir ability to stabilize neurons by way of hyperpolarization of the mem-brane potential by blocking voltage-gated Naþ channels [88]. Treatmentwith phenytoin attenuates accumulation of Kþ in cerebrospinal fluid in an-imals subjected to circulatory arrest. Some studies of phenytoin treatmenthave demonstrated attenuation of brain edema, increased Naþ/Kþ-ATPaseactivity, decreased intracellular Naþ concentration, and attenuated accumu-lation of lactate and free fatty acids [99]. Lamotrigine attenuates ischemia-induced increases in extracellular glutamate levels and improves histologicoutcomes in the gerbil and rat models of global cerebral ischemia [100].

Magnesium

Magnesium sulfate has multimodal actions. It is an NMDA receptor,a calcium antagonist, and a vasodilator. In a rat model of global cerebralischemia, pre- and post-treatment with bolus intravenous magnesium sulfatefollowed by a continuous intravenous infusion attenuated injury to CA1hippocampal neurons [101]. Other investigators have reported that neuro-protection is demonstrated only when magnesium sulfate is administeredin combination with mild hypothermia [102].

Anesthetic agents

Barbiturates have long been known to exert neuroprotection by coupleddecreases in cerebral metabolic rate of oxygen (CMRO2) with CBF and by in-hibiting agonist-induced cerebral vasoconstrictor responses, protein kinaseC,ischemia-induced increases in free fatty acids, and EAA release [29]. In exper-imental global cerebral ischemia, however, results have been disappointing.Treatment with pentobarbital failed to improve survival in a dog model ofglobal cerebral ischemia [103]. Clinical investigation with barbiturateshas also proved to be disappointing, with lack of therapeutic benefit in thehypoxic-ischemic brain injury from CA and near-drowning [29]. Inhalational


anesthetics (halothane, isoflurane) have not been thoroughly investigated inthe setting of global cerebral ischemia. The premise for their use is their abilityto attenuate cerebral metabolic demand, enhance regional CBF, attenuateischemia-evoked EAA and catecholamine efflux, and attenuate calciuminfluxes [29].

Cyclooxygenase inhibitors

Nimesulide, a cyclooxygenase-2 inhibitor, attenuated injury to the CA1region of the hippocampus in a gerbil model when administered orally orintraperitoneally as a pre- or post-treatment (up to 24 hours) [104]. Furtherexperimental studies in other animal models are needed to confirm thesefindings and bring these agents into the clinical paradigm.

Immunosuppressants

Tacrolimus (FK506) and cyclosporine are immunophilin and calcineurininhibitors that attenuate apoptotic cell death. Chronic administration ofboth agents (for 3 days) before the ischemic insult demonstrated neuropro-tection in the CA1 region at 7 days of reperfusion in a rat model of globalcerebral ischemia [105]. These agents also attenuated calcineurin activity inthe CA1, CA3, and dentate gyrus regions of the hippocampus up to 24hours following the ischemic insult. Pretreatment with cyclosporin andFK506 inhibits dephosphorylation of the proapoptotic protein Bad. The in-ability of cyclosporine to cross an intact BBB is a significant therapeuticconcern. These agents require more rigorous testing with treatment in thepostischemia paradigm across different species and animal models of globalcerebral ischemia [88].

Potential future of neuroprotective agents

Hormonal sex steroids

Evidence is mounting that outcome from cerebral ischemia is quantitative-ly different in adult male and female animals, reflecting patterns of someforms of human cerebrovascular disease [8,106]. Accordingly, it has becomeincreasingly apparent that biologic sex is an important factor in pathophysi-ology and outcome following cerebral ischemia [107]. For example, whenboth sexes are studied, ischemic outcome in transgenic mice can be overtlysex dependent, even when the gene of interest (eg, inducible or neuronalNO) is not linked to sexual development [108–111]. These data suggest thatmolecular mechanisms of cell injury may not be the same or have the sameimpact in the male and the female brain. Furthermore, most experimentalstudies underscore the importance of sex steroids (predominantly estrogen)to outcome from focal ischemia [107,112]; recent studies have reported signif-icant neuroprotection in the global cerebral ischemia paradigm [113].


Opioid-receptor agonists

Experimental research over the last 3 decades has implicated the opioi-dergic receptors in the brain to play an important pathophysiologic rolein cerebral ischemia [114,115]. Opiate receptors in the central nervous sys-tem have been divided into three subtypes: mu (m), kappa (k), and delta(d). Although the m- and d-receptor subtypes have been shown to playa role predominantly in antinociception, several experimental studies havedemonstrated neuroprotective effects of k–opioid-receptor agonists in mod-els of global [116] and focal ischemia [117,118]. Differential time course andalterations in opioid-receptor binding after focal cerebral ischemia in miceindicate that k–opioid-receptor binding sites are preserved much longer(12–48 hours) than other subtypes [119], suggesting a potentially longertherapeutic window with k–opioid-receptor agonists. In vitro studies suggestthat k–opioid-receptor agonists modulate the excitotoxic action of gluta-mate, possibly by the presynaptic inhibition of its release by decreasingthe entry of 45Ca2þ into rat cortical synaptosomes [120,121]. It has beendemonstrated that the selective k–opioid-receptor agonist BRL 52,537attenuates in situ NO production at doses that provide ischemic neuropro-tection [122] that is receptor selective [123] but does not attenuate ischemia-evoked dopamine release in the ischemic striatum [124]. Furthermore, BRL52537 provides significant ischemic neuroprotection when administered forup to 6 hours after the onset of focal cerebral ischemia [125], confirminga long therapeutic opportunity to afford ischemic neuroprotection withthese compounds. It has recently been demonstrated that ischemic neuro-protection with BRL 52537 is sex specific; it confers neuroprotection onlyin male animals [126]. Thus, highly selective k–opioid-receptor agonistsmight hold promise as neuroprotective agents in the future in the global ce-rebral ischemia paradigm.

Sigma-receptor agonists

Over the past 2 decades, nonopioid sigma (s)-receptors have been consid-ered to serve an important physiologic function [127]. A number of ‘‘atypi-cal’’ antipsychotics are potent s-receptor ligandsdatypical, in that theyare effective antipsychotic agents but have low propensity to induce extrapy-ramidal side effects [28]. Naturally occurring s-receptor ligands include pro-gesterone and neuropeptide Y [128]. Purification, molecular cloning, andhigh levels of expression of s1-receptor binding sites in sterol-producingtissues have been demonstrated [129]. Subtyping of the s-receptor into as1-receptor and s2-receptor is based on important differences in regional dis-tribution, enantomeric selectivity, molecular weights, and second messengersemployed in signaling [130,131]. Until recently, specific s1-receptor ligands orantagonists have been lacking. During evaluation as antipsychotic agents,several drugs known to be s-receptor ligands were demonstrated to alterNMDA-receptor function [132]; however, it now appears that the effect


might be directly mediated through activity at the s-receptor. The authorshave demonstrated that the potent s1-receptor ligand 4-phenyl-1-(4-phenyl-butyl) piperidine (PPBP) provides robust ischemic neuroprotection in animalmodels of cerebral is chemia [133–135]. Although the exact in vivo mecha-nisms of neuroprotection bys1-receptor ligands are not completely elucidated,several antiexcitotoxic mechanisms have been postulated, including inhibi-tion of ischemia-induced presynaptic glutamate release, attenuation of post-synaptic glutamate-evoked Ca2þ influx, modulation of neuronal responses toNMDA-receptor stimulation, inhibition of dopamine neurotransmission,and prevention of cortical-spreading depression [133–135]. Recently, it wasdemonstrated that PPBP provides ischemic neuroprotection in vivo by wayof attenuation of ischemia-evoked NO production [134] but does not alter is-chemia-evoked dopamine efflux in the ischemic striatum [135]. Based on thesedata, s1-receptor ligands hold promise for the future as neuroprotectants inthe treatment of cerebral ischemia.

In addition to the above agents, a number of others that have been stud-ied in the experimental paradigm of global cerebral ischemia hold promisefor the future. These agents include anesthetics (etomidate, ketamine, pro-pofol), sodium channel blockers (mexitine, lidocaine), a-receptor agonists(dexmedetomidine), and xanthine oxidase inhibitors (allopurinol). Fora complete review of this subject, the reader is referred to the monographby Weigl and colleagues [88].

Summary

Cerebral ischemia results in a rapid depletion of energy stores that trig-gers a complex cascade of cellular events such as cellular depolarizationand Ca2þ influx, resulting in excitotoxic cell death. The critical determinantof severity of brain injury is the duration and severity of the ischemic insultand early restoration of CBF. Induced therapeutic hypothermia followingCA is the only strategy that has demonstrated improvement in outcomesin prospective, randomized clinical trials. Although pharmacologic neuro-protection has been disappointing thus far in a variety of experimental an-imal models, further research efforts are directed at using some agents thatdemonstrate marginal or moderate efficacy in combination with hypother-mia. Although the signal transduction pathways and intracellular molecularevents during cerebral ischemia and reperfusion are complex, potential ther-apeutic neuroprotective strategies hold promise for the future.

Acknowledgments

The authors thank Tzipora Sofare, MA, for her editorial assistance inpreparing this manuscript.


References

[1] Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-

hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63.

[2] The Hypothermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to

improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56.

[3] Salazar JD, Wityk RJ, Grega MA, et al. Stroke after cardiac surgery: short- and long-term

outcomes. Ann Thorac Surg 2001;72:1195–201.

[4] Nussmeier NA. A review of risk factors for adverse neurologic outcome after cardiac sur-

gery. J Extra Corpor Technol 2002;34:4–10.

[5] Murkin JM. Etiology and incidence of brain dysfunction after cardiac surgery. J Cardio-

thorac Vasc Anesth 1999;13:12–37.

[6] Llinas R, Barbut D, Caplan LR. Neurologic complications of cardiac surgery. Prog Cardi-

ovasc Dis 2000;43:101–12.

[7] Wityk RJ, Glodsborough MA, Hillis A, et al. Diffusion- and perfusion-weighted brain

magnetic resonance imaging in patients with neurologic complications after cardiac sur-

gery. Arch Neurol 2001;58:571–6.

[8] Bhardwaj A, Alkayed NJ, Kirsch JR, et al. Mechanisms of ischemic brain damage. Curr

Cardiol Rep 2003;5:160–7.

[9] Pulsinelli WA. Selective neuronal vulnerability: morphological and molecular characteris-

tics. Prog Brain Res 1995;63:29–37.

[10] Petito CK, Pulsinelli WA. Delayed neuronal recovery and neuronal death in rat hippocam-

pus following severe cerebral ischemia: possible relationship to abnormalities in neuronal

processes. J Cereb Blood Flow Metab 1984;4:194–205.

[11] Kaplan B, Brint S, Tanabe J, et al. Temporal thresholds for neocortical infarction in rats

subjected to irreversible focal cerebral ischemia. Stroke 1991;22:1032–9.

[12] PusinelliWA,Brierley JB, PlumF. Temporal profile of neuronal damage in amodel of tran-

sient forebrain ischemia. Ann Neurol 1982;11:491–8.

[13] Kagstorm E, Smith M-L, Siesjo BK. Local cerebral blood flow in the recovery period

following complete cerebral ischemia in the rat. J Cereb Blood Flow Metab 1983;3:

170–82.

[14] Pulsinelli WA, Duffy DE. Regional energy balance in rat brain after transient forebrain

ischemia. J Neurochem 1983;40:1500–3.

[15] Ginsberg MD, Busto R. Rodent models of cerebral ischemia. Stroke 1989;20:1627–42.

[16] Todd MM, Dunlop BJ, Shapiro HM, et al. Ventricular fibrillation in the cat: a model for

global cerebral ischemia. Stroke 1981;12:808–15.

[17] Koch KA, Jackson DL, Schmiedl M, et al. Total cerebral ischemia: effect of alterations in

arterial PCO2 on cerebral microcirculation. J Cereb Blood Flow Metab 1984;4:343–9.

[18] Nishijima MK, Koehler RC, Hurn PD, et al. Postischemic recovery rate of cerebral ATP,

phosphocreatine, pH, and evoked potentials. Am J Physiol 1989;257:H1860–70.

[19] Fonnum F. Glutamate: a neurotransmitter in mammalian brain. J Neurochem 1984;42:

1–11.

[20] Koehler RC, Eleff SM, TraystmanRJ.Global neuronal ischemia and reperfusion. In: Para-

dis NA, Halperin HR, Nowak RM, editors. Cardiac arrest: the science and practice of

resuscitation medicine. Baltimore (MD): Williams and Wilkins; 1996. p. 113–45.

[21] Leist M, Nicotera P. Cell death: apoptosis versus necrosis. In: Welch KMA, Caplan LR,

Reis DJ, et al, editors. Primer on cerebrovascular disease. Philadelphia: Academic Press;

1997. p. 101–4.

[22] Cohen JJ. Overview: mechanisms of apoptosis. Immunol Today 1993;14:126–30.

[23] Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Science 1995;267:

1456–62.

[24] Dirnagl U, Iadecola C, Moskowitz MA. Pathobiology of ischaemic stroke: an integrated

view. Trends Neurosci 1999;22:391–7.


[25] Vaux L, Haecker G, Strasser A. An evolutionary perspective on apoptosis. Cell 1994;76:

777–9.

[26] Ankarcrona M, Dypbukt JM, Bonfoco E, et al. Necrosis and apoptosis elicited by gluta-

mate in cerebellar granule cells: a temporal sequence depending onmitochondrial function.

Neuron 1995;15:961–73.

[27] Graham SH, Chen J. Programmed cell death in cerebral ischemia. J Cereb Blood Flow

Metab 2001;21:99–109.

[28] Sugawara T, FujimuraM, Noshita N, et al. Neuronal death/survival signaling pathways in

cerebral ischemia. NeuroRx 2004;1:17–25.

[29] RossbergMI, BhardwajA,Hurn PD, et al. Principles of cerebroprotection. In:MurrayMJ,

Coursin DB, Pearl RG, et al, editors. Critical care medicine: perioperative management.

Philadelphia: Lipincott Williams & Williams; 2002. p. 225–35.

[30] Olney JW. Brain lesion, obesity and other disturbances in mice treated with monosodium

glutamate. Science 1969;164:719–21.

[31] Choi DW. Glutamate neurotoxicity and diseases of the nervous system. Neuron 1988;1:

623–34.

[32] BenvenisteH,Drejer J, Schoushoe A, et al. Elevation of extracellular concentrations of glu-

tamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by

intracerebral microdialysis. J Neurochem 1984;43:1369–74.

[33] Choi DW. Excitotoxic cell death. J Neurobiol 1992;23:1261–76.

[34] Sommer B, Seeburg PH. Glutamate receptor channels: novel properties and new clones.

Trends Pharmacol Sci 1992;13:291–6.

[35] Choi DW. Methods of antagonizing glutamate neurotoxicity. Cerebrovasc Brain Metab

Rev 1990;2:105–47.

[36] Bhardwaj A, Brannan T, Martinez-Tica J, et al. Ischemia in the dorsal hippocampus is

associated with acute release of dopamine and norepinephrine. J Neural Transm 1990;80:

195–201.

[37] BhardwajA, BrannanT,Weinberger J. Pentobarbital inhibits extracellular release of dopa-

mine in the ischemic striatum. J Neural Transm 1990;82:111–7.

[38] Slivka A, Brannan TS, Weinberger J, et al. Increase in extracellular dopamine in the strai-

tum during cerebral ischemia: a study utilizing cerebral microdialysis. J Neurochem 1988;

50:1714–8.

[39] Globus MY-T, Busto R, Dietrich WD, et al. Intraischemic extracellular release of dopa-

mine and glutamate is associated with striatal vulnerability to ischemia. Neurosci Lett

1988;91:36–40.

[40] Koorn R, Brannan TS, Martinez-Tica J, et al. Effect of etomidate on in vivo ischemia-

induced dopamine release in the corpus striatumof the rat: a study using cerebralmicrodial-

ysis. Anesth Analg 1994;78:73–9.

[41] Koorn R, Kahn RA, Martinez-Tica J, et al. Effect of isoflurane and halothane on in vivo

ischemia-induced dopamine release in the corpus striatum of the rat: a study utilizing cere-

bral microdialysis. Anesthesiology 1993;79:827–35.

[42] Weinberger J, CohenG,Nieves-Rosa J.Nerve terminal damage in cerebral ischemia: greater

susceptibility of catecholamine nerve terminals relative to serotonin nerve terminals. Stroke

1983;14:986–9.

[43] Weinberger J, Nieves-Rosa J, Cohen G. Nerve terminal damage in cerebral ischemia: pro-

tective effect of alpha-methyl-para-tyrosine. Stroke 1985;16:864–70.

[44] Globus MY-T, Ginsberg MD, Dietrich WD, et al. Subsantia nigra lesion protects against

ischemic damage in the striatum. Neurosci Lett 1987;80:251–6.

[45] KahnRA,Weinberger J, Brannan T, et al. Nitric oxidemodulates dopamine release during

global temporary cerebral ischemia. Anesth Analg 1995;80:1116–21.

[46] Dawson TM, Dawson VL, Snyder SH. A novel neuronal messenger in brain: the free rad-

ical, nitric oxide. Ann Neurol 1992;32:297–311.


[47] Dawson TM, Dawson VL. Nitric oxide: actions and pathological roles. Neuroscientist

1995;1:7–18.

[48] Samdani AF, Dawson TM,Dawson VL. Nitric oxide synthase in models of focal ischemia.

Stroke 1997;28:1283–8.

[49] Dalkara T, Moskowitz MA. The role of nitric oxide in cerebral ischemia. In: Welch KMA,

Caplan LR, Reis DJ, et al, editors. Primer on cerebrovascular disease. Philadephia: Aca-

demic Press; 1997. p. 207–8.

[50] Goto S, Xue R, Sugo N, et al. Poly(ADP-ribose) polymerase impairs early and long-term

experimental stroke recovery. Stroke 2002;33:1101–6.

[51] Bhardwaj A, Northington FJ, Ichord RN, et al. Characterization of ionotropic glutamate

receptor-mediated nitric oxide production in vivo in rats. Stroke 1997;28:850–6.

[52] Bhardwaj A, Northington FJ, Martin LJ, et al. Characterization of metabotropic gluta-

mate receptor-mediated nitric oxide production in vivo. J Cereb Blood Flow Metab

1997;17:153–60.

[53] Dalkara T, Morikawa H, Panahian N, Moskowitz MA. Blood flow-dependent functional

recovery in a rat model of cerebral ischemia. Am J Physiol 1994;267:H678–83.

[54] Huang Z, Huang PL, Ma J, et al. Enlarged infarcts in endothelial nitric oxide synthase

knockout mice are attenuated by nitro-L-arginine. J Cereb Blood Flow Metab 1996;16:

981–7.

[55] HuangZ,Huang PL, PanahianN, et al. Effects of cerebral ischemia inmice deficient in neu-

ronal nitric oxide synthase. Science 1994;265:1883–5.

[56] Zheng Z, Lee JE, Yenari MA. Stroke: molecular mechanisms and potential targets for

treatment. Curr Mol Med 2003;3:361–72.

[57] Han HS, Yenari MA. Cellular targets of brain inflammation in stroke. Curr Opin Investig

Drugs 2003;4:522–9.

[58] Danton GH, Dietrich WD. Inflammatory mechanisms after ischemia and stroke. J Neuro-

pathol Exp Neurol 2003;62:127–36.

[59] Ginis I, Jaiswal R, Klimanis D, et al. TNF-alpha-induced tolerance to ischemic injury in-

volves differential control of NF-kappaB transactivation: The role of NF-kappaB associa-

tion with p300 adaptor. J Cereb Blood Flow Metab 2002;22:142–52.

[60] Gary DS, Bruce-Keller AJ, Kindy MS, et al. Ischemic and excitotoxic brain injury is en-

hanced in mice lacking the p55 tumor necrosis factor receptor. J Cereb Blood FlowMetab

1998;18:1283–7.

[61] Bruce AJ, Boling W, Kindy MS, et al. Altered neuronal and microglial responses to

excitotoxic and ischemic brain injury in mice lacking TNF receptors. Nat Med 1996;

2:788–94.

[62] Kinoshita K, Kraydieh S, Alonso O, et al. Effect of posttraumatic hyperglycemia on con-

tusion volume and neutrophil accumulation after moderate fluid-percussion brain injury in

rats. J Neurotrauma 2002;19:681–92.

[63] ChewW, Kucharczyk J, MoseleyM, et al. Hyperglycemia augments ischemic brain injury:

in vivo MR imaging/spectroscopic study with nicardipine in cats with occluded middle ce-

rebral arteries. AJNR Am J Neuroradiol 1991;12:603–9.

[64] Conroy BP, GrafeMR, Jenkins LW, et al. Histopathologic consequences of hyperglycemic

cerebral ischemia during hypothermic cardiopulmonary bypass in pigs. Ann Thorac Surg

2001;71:1325–34.

[65] Guyot LL, Diaz FG, O’ReganMH, et al. The effect of streptozotocin-induced diabetes on

the release of excitotoxic and other amino acids from the ischemic rat cerebral cortex. Neu-

rosurgery 2001;48:385–90.

[66] Lin B, GinsbergMD, Busto R, et al. Hyperglycemia triggers massive neutrophil deposition

in brain following transient ischemia in rats. Neurosci Lett 2000;278:1–4.

[67] DingC,HeQ, Li PA. activation of cell death pathway after a brief period of global ischemia

in diabetic and non-diabetic animals. Exp Neurol 2004;188:421–9.


[68] Baird TA, Parsons MW, Phanh T, et al. Persistent poststroke hyperglycemia is indepen-

dently associated with infarct expansion and worse clinical outcome. Stroke 2003;34:

2208–14.

[69] Nuttall GA, Abel MD, Mullany CJ, et al. Intraoperative hyperglycemia and perioperative

outcomes in cardiac surgery patients. Mayo Clin Proc 2005;80:862–6.

[70] Van den Berghe G, Schoonheydt K, Becx P, et al. Insulin therapy protects the central and

peripheral nervous system of intensive care patients. Neurology 2005;64(26):1348–53.

[71] Auer RN. Non-pharmacologic (physiologic) neuroprotection in the treatment of brain is-

chemia. Ann N Y Acad Sci 2001;939:271–82.

[72] Busto R, Dietrich WD, Globus MY, et al. Small differences in intraischemic brain temper-

ature critically determine the extent of ischemic neuronal injury. J CerebBloodFlowMetab

1987;7:729–38.

[73] Zhao O, Memezawa H, SmithML, et al. Hyperthermia complicates middle cerebral artery

occlusion induced by an intraluminal filament. Brain Res 1994;649:253–9.

[74] Minamisawa H, Smith ML, Siesjo BK. The effect of mild hyperthermia and hypothermia

on brain damage following 5, 10 and 15 min of forebrain ischemia. Ann Neurol 1990;28:

26–33.

[75] Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell

injury in ischemic myocardium. Circulation 1986;74:1136.

[76] Kitagawa K, Matsumoto M, Tagaya M. ‘‘Ischemic tolerance’’ phenomenon found in the

brain. Brain Res 1990;528:21–4.

[77] Liu Y, Kato H, Nakata N, et al. Protection of rat hippocampus against ischemic neuronal

damage by pretreatment with sublethal ischemia. Brain Res 1992;586:121–4.

[78] Nishi S, TakiW,UemuraY, et al. Ischemic tolerance due to the induction of HSP70 in a rat

ischemic recirculation model. Brain Res 1993;615(2):281–8.

[79] ChoppM, ChenH,HoKL, et al. Transient hyperthermia protects against subsequent fore-

brain ischemic cell damage in the rat. Neurology 1989;39:1396–8.

[80] Nishio S, Yunoki M, Chen ZF, et al. Ischemic tolerance in the rat neocortex following hy-

pothermic preconditioning. J Neurosurg 2000;93:845–51.

[81] Bergstedt K, Hu BR, Wieloch T. Initiation of protein synthesis and heat-shock protein-72

expression in the rat brain following severe insulin-induced hypoglycemia. Acta Neuropa-

thol (Berl) 1993;86:145–53.

[82] Huber R, Kasischke K, Ludolph AC, et al. Increase of cellular hypoxic tolerance by eryth-

romycin and other antibiotics. Neuroreport 1999;10:1543–6.

[83] RiepeMW, Kasischke K, Raupach A. Acetylsalicylic acid increases tolerance against hyp-

oxic and chemical hypoxia. Stroke 1997;28:2006–11.

[84] Dawson TM. Preconditioning-mediated neuroprotection through erythropoietin? Lancet

2002;359:96–7.

[85] Kapinya KJ, Lowl D, Futterer C, et al. Tolerance against ischemic neuronal injury can be

induced by volatile anesthetics and is inducible NO synthase dependent. Stroke 2002;33:

1889–98.

[86] StaglianoNE, Perez-PinzonMA,MoskowitzMA, et al. Focal ischemic preconditioning in-

duces rapid tolerance to middle cerebral artery occlusion in mice. J Cereb Blood Flow

Metab 1999;19:757–61.

[87] Barone FC, White RF, Spera PA, et al. Ischemic preconditioning and brain tolerance: tem-

poral histological and functional outcomes, protein synthesis requirement, and interleukin-1

receptor antagonist and early gene expression. Stroke 1998;29:1937–50.

[88] WeiglM, TenzeG, Steinlechner B, et al. A systematic review of currently available pharma-

cological neuroprotective agents as a sole intervention before anticipated or induced cardi-

ac arrest. Resuscitation 2005;65:21–39.

[89] Arnowski J, Waxham MN, Grotta JC. Neuronal protection and preservation of calcium/

calmodulin-dependent protein kinase II and protein kinase C activity by dextrorphan treat-

ment in global ischemia. J Cereb Blood Flow Metab 1993;13:550–7.


[90] Li MM, Payne RS, Reid KH, et al. Correlates of delayed neuronal damage and neuropro-

tection in a rat model of cardiac-arrest-induced cerebral ischemia. Brain Res 1999;826:

44–52.

[91] Dietrich WD, Lin B, Globus MY, et al. Effect of delayed MK-801 (dizocilpine) treatment

with or without immediate postischemic hypothermia on chronic neuronal survival after

global forebrain ischemia in rats. J Cereb Blood Flow Metab 1995;15:960–8.

[92] Calle PA, Paridaens K, De Ridder LI, et al. Failure of nimodipine to prevent brain damage

in a global brain ischemia model in the rat. Resuscitation 1993;25:59–71.

[93] Lazarewicz JW, Pluta R, PukaM, et al. Diverse mechanisms of neuronal protection by ni-

modipine in experimental rabbit brain ischemia. Stroke 1990;21:108–10.

[94] Roine RO, KasteM, Kinnunen A, et al. Nimodipine after resuscitation from out-of-hospi-

tal ventricular fibrillation. A placebo-controlled, double-blind, randomized trial. JAMA

1990;264:3171–7.

[95] Shuaib A, Murabit MA, Kanthan R, et al. The neuroprotective effects of gamma-vinyl

GABA in transient global ischemia: a morphological study with early and delayed evalua-

tions. Neurosci Lett 1996;204(2):1–4.

[96] Thaminy S, Reymann JM,HeresbachN, et al. Is chlormethiazole neuroprotective in exper-

imental global cerebral ischemia? A microdialysis and behavioral study. Pharmacol Bio-

chem Behav 1997;56:737–45.

[97] Vergoni AV, Ottani A, Botticelli AR, et al. Neuroprotective effect of gamma-hydroxybu-

tyrate in transient global cerebral ischemia in the rat. Eur J Pharmacol 2000;397:75–84.

[98] Iqbal S, Baziany A, Gordon S, et al. Neuroprotective effect of tiagabine in transient fore-

brain global ischemia: an in vivo microdialysis, behavioral, and histological study. Brain

Res 2002;946:162–70.

[99] Artru AA, Michenfelder JD. Anoxic cerebral potassium accumulation reduced by phenyt-

oin: mechanism of cerebral protection? Anesth Analg 1981;60:41–5.

[100] Imaizumi S, Kurosawa K, Kinouchi H, et al. Effect of phenytoin on cortical Na(þ)-K(þ)-

ATPase activity in global ischemic rat brain. J Neurotrauma 1995;12:231–4.

[101] Miles AN, Majda BT, Meloni BP, et al. Postischemic intravenous administration of mag-

nesium sulfate inhibits hippocampal CA1 neuronal death after transient global ischemia in

rats. Neurosurgery 2001;49:1443–50.

[102] Zhu H, Meloni BP, Bojarski C, et al. Post-ischemic modest hypothermia (35 degrees C)

combined with intravenous magnesium is more effective at reducing CA1 neuronal death

than either treatment used alone following global cerebral ischemia in rats. Exp Neurol

2005;193:361–8.

[103] Steen PA, Milde JH, Michenfelder JD. No barbiturate protection in a dog model of com-

plete cerebral ischemia. Ann Neurol 1979;5:343–9.

[104] Candelario-Jalil E, Alvarez D, Gonzalez-Falcon A, et al. Neuroprotective efficacy of nime-

sulide against hippocampal neuronal damage following transient forebrain ischemia. Eur J

Pahrmacol 2002;453:189–95.

[105] Uchino H, Minamikawa-Tachino R, Kristian T, et al. Differential neuroprotection

by cyclosporine A and FK506 following ischemia corresponds with differing abilities to

inhibit calcineurin and the mitochondrial permeability transition. Neurobiol Dis 2002;10:

219–33.

[106] McCullough LD, Hurn PD. Estrogen and ischemic neuroprotection: an integrated view.

Trends Endocrinol Metab 2003;14:228–35.

[107] Hurn PD, Macrae IM. Estrogen as neuroprotectant in stroke. J Cereb Blood Flow Metab

2000;20:631–52.

[108] Loihl AK, Asensio V, Campbell IL, et al. Expression of nitric oxide synthase (NOS)-2 fol-

lowing permanent focal ischemia and the role of nitric oxide in infarct generation in male,

female and NOS-2 gene deficient mice. Brain Res 1999;830:155–64.

[109] SampeiK,Mandir AS, AsanoY, et al. Stroke outcome in doublemutant antioxidant trans-

genics. Stroke 2000;31:2685–91.


[110] Kofler J, Otsuka T, Zhang Z, et al. Differential effect of PARP-2 deletion on brain injury

after focal and global cerebral ischemia. J Cereb Blood Flow Metab 2005 [Epub ahead

of print].

[111] McCullough LD, Zeng Z, Blizzard KK, et al. Ischemic nitric oxide and poly(ADP-ribose)

polymerase-1 in cerebral ischemia: male toxicity, female protection. J Cereb Blood Flow

Metab 2005;25:502–12.

[112] Garcia-Segura LM, Azcoitia I, DonCarlos LL. Neuroprotection by estradiol. Prog Neuro-

biol 2001;63:29–60.

[113] Noppens RR, Kofler J, Hurn PD, et al. Dose-dependent neuroprotection by 17beta-

estradiol after cardiac arrest and cardiopulmonary resuscitation. Crit Care Med 2005;

33:1595–602.

[114] Fried RL, Nowak TS. Opioid peptide levels in gerbil brain after transient ischemia: lasting

depletion of hippocampal dynorphin. Stroke 1987;18:765–70.

[115] Zabramski JM, Spetzler RF, SelmanWR, et al. Nalaxone therapy during focal cerebral is-

chemia evaluation in a primate model. Stroke 1984;15:621–7.

[116] Tang AH. Protection from cerebral ischemia by U-50,488E, a specific kappa opioid anal-

gesic agent. Life Sci 1985;37:1475–82.

[117] Mackay KB, Kusomoto K, Graham DI, et al. Effect of kappa-1 opioid agonist CI-977 on

ischemic brain damage and cerebral blood flow after middle cerebral artery occlusion in the

rat. Brain Res 1993;629:10–8.

[118] Baskin DS, Widmayer MA, Browning JL, et al. Evaluation of delayed treatment of focal

cerebral ischemia with three selective k-opioid agonists in cats. Stroke 1994;25:2047–54.

[119] Boutin H, Dauphin F,MacKenzie ET, et al. Differential time-course decreases in nonselec-

tive mu-, delta-, and kappa-opioid receptors after focal cerebral ischemia in mice. Stroke

1999;30:1271–7.

[120] Bradford HF, Crowder JM,White EJ. Inhibitory actions of opioid compounds on calcium

fluxes and neurotransmitter release frommammalian cerebral cortical slices. Br J Pharmac

1986;88:87–93.

[121] Xiang J-Z, Adamson P, Brammer MJ, et al. The kappa-opiate agonist U50488H decreases

the entry of 45Ca into rat cortical synaptosomes by inhibiting N- but not L-type calcium

channels. Neuropharmacol 1990;29:439–44.

[122] Goyagi T, ToungTJ, Kirsch JR, et al. Neuroprotective kappa-opioid receptor agonist BRL

52537 attenuates ischemia-evoked nitric oxide production in vivo in rats. Stroke 2003;34:

1533–8.

[123] Zhang Z, Chen TY, Kirsch JR, et al. Kappa-opioid receptor selectivity for ischemic neuro-

protection with BRL 52537 in rats. Anesth Analg 2003;97:1776–83.

[124] Chen TY, Goyagi T, Toung TJ, et al. Prolonged opportunity for ischemic neuroprotection

with selective kappa-opioid receptor agonist in rats. Stroke 2004;35:1180–5.

[125] ChenCH, Toung TJ, Hurn PD, et al. Ischemic neuroprotectionwith selective kappa-opioid

receptor agonist is gender specific. Stroke 2005;36:1557–61.

[126] Zeynalov E, Nemoto M, Hurn PD, et al. Neuoprotective effect of selective kappa receptor

agonist is gender specific and linked to reduced neuronal nitric oxide. J Cereb Blood Flow

Metab 2005 July 27 [Epub ahead of print].

[127] Vu TH, Weissmann AD, London ED. Pharmacological characteristics and distribution of

s- and phencyclidine receptors in the animal kingdom. J Neurochem 1990;54:598–604.

[128] Su T-P, London ED, Jaffe JH. Steroid binding at sigma receptors suggests a link between

endocrine, nervous, and immune systems. Science 1988;240:219–21.

[129] Hanner M, Moebius FF, Flandorfer A, et al. Purification, molecular cloning, and expres-

sion of the mammalian sigma1-binding site. Proc Natl Acad Sci U S A 1996;93:8072–7.

[130] Su T-P. Delineating biochemical and functional properties of sigma receptors: emerging

concepts. Crit Rev Neurobiol 1993;7:187–203.

[131] Leitner ML, Hohmann AG, Patrick SL, et al. Regional variation in the ratio of s1 to s2

binding in rat brain. Eur J Pharmacol 1994;259:65–9.


[132] Contreras PC, GrayNM,RaganDM, et al. BMY-14802 protects against ischemia-induced

neuronal damage in the gerbil. Life Sci 1992;51:1145–9.

[133] Harukuni I, Bhardwaj A, Shaivitz AB, et al. Sigma1-receptor ligand 4-phenyl-1-(4-phenyl-

butyl)-piperidine affords neuroprotection from focal ischemia with prolonged reperfusion.

Stroke 2000;31:976–82.

[134] Goyagi T, Goto S, Bhardwaj A, et al. Neuroprotective effect of s1-receptor ligand,

4-phenyl-1-(4-phenylbutyl) piperidine (PPBP) is linked to reduced neuronal nitric oxide

production. Stroke 2001;32:1613–20.

[135] Goyagi T, Bhardwaj A, Koehler RC, et al. Potent sigma 1-receptor ligand 4-phenyl-1-

(4-phenylbutyl) piperidine provides ischemic neuroprotection without altering dopamine

accumulation in vivo in rats. Anesth Analg 2003;96:532–8.

Neurol Clin 24 (2006) 23–39

Prehospital and Emergency DepartmentCare to Preserve Neurologic Function

During and Following CardiopulmonaryResuscitation

Joseph P. Ornato, MD, Mary Ann Peberdy, MD*Department of Emergency Medicine and Internal Medicine,

Virginia Commonwealth University Health System, 1200 East Broad Street,

West Hospital, 10th Floor, Room 1042, Richmond, VA 23298, USA

Approximately 400,000 to 460,000 cardiac arrests occur out of the hospitalin the United States each year [1]. Despite major advances in the EmergencyMedical Service (EMS) system, overall survival from out-of-hospital cardiacarrest (OHCA) remains poor, averaging only 5% to 8% in most communities[2]. Sudden death is the first manifestation of underlying cardiovascular dis-ease in most patients who have OHCA [3]. A ventricular tachyarrhythmia(ventricular tachycardia [VT] or ventricular fibrillation [VF]) has been docu-mented to be the triggering event in up to 80% of cases [4].

The purpose of this article is to describe the prehospital and emergencydepartment (ED) approach to resuscitating patients who have OHCA,with particular emphasis on what can be done during resuscitation to pre-serve neurologic function. The authors review the critical lifesaving interven-tions that can be provided by the continuum of emergency cardiac caredfrom citizen first response to EMS and ED care. The OHCA studies thathave focused on improving neurologic outcome are reviewed, and the ma-jor, new National Institutes of Health (NIH) effort in this area is described.

Layperson first responders

Early defibrillation

The public health and EMS strategy for decreasing the mortality andneurologic sequelae from OHCA has focused on enhancing emergency


E-mail address: [email protected] (M.A. Peberdy).




24 ORNATO & PEBERDY

resuscitation measures because prevention and accurate identification of atrisk individuals is difficult. The traditional approach has been to train thepublic to recognize cardiac arrest, call 9-1-1, and perform cardiopulmonaryresuscitation (CPR) while public-safety personnel (fire or police ‘‘first re-sponders,’’ emergency medical technicians [EMTs], or paramedics) rush tothe scene to provide defibrillation and other advanced cardiac life support(ACLS) treatments. This sequence of events is the cornerstone behind theAmerican Heart Association’s chain of survival conceptual model for im-proving outcome from OHCA (Fig. 1) [5].

The patient’s initial cardiac rhythm is a principal determinant of resusci-tation survival and neurologic outcome. For example, in one series of 352consecutive OHCA patients, 67% of the patients who had VT and 23% ofthose who initially were in VF survived to hospital discharge [6]. None of thepatients who presented with an initial pulseless bradycardia or asystole sur-vived to hospital discharge. One obvious hypothesis is that bradyarrhythmiamay be a marker for a prolonged downtime interval or a more severe under-lying disease process. Because ventricular tachyarrhythmias represent themost common, potentially treatable mechanism of sudden cardiac arrestin adults, the best community resuscitation programs deliver rapid defibril-lation to as many patients as possible because the odds of survival decreaseby 7% to 10% for each minute that a patient remains in VF [5].

An increasingly popular community approach to increase the number ofout-of-hospital VF patients who receive early defibrillation is public-accessdefibrillation, so named because the intent is to have laypersons perform earlydefibrillation with automated external defibrillators (AEDs) while awaitingarrival of EMS personnel. Trained public-safety laypersons (eg, law enforce-ment officers, firefighter first responders, flight attendants, airport employ-ees, casino security officers) can treat cardiac arrest victims safely and

Early

Access

Early

CPR

Early

ACLS

Early

Defibrillation

Fig. 1. The chain of survival. (From Cummins RO, Ornato JP, Thies WH, et al. Improving sur-

vival from sudden cardiac arrest: the ‘‘chain of survival’’ concept. A statement for health

professionals from the Advanced Cardiac Life Support Subcommittee and the Emergency

Cardiac Care Committee, American Heart Association. Circulation 1991;83(5):1832; with

permission).

25PREHOSPITAL AND ED CARE

effectively with AEDs in a wide variety of public settings [7–13]. In addition,a randomized clinical trial of public-access defibrillation showed that thenumber of survivors from OHCA in public places doubled after laypersonswere trained to perform CPR and use AEDs compared with being trained toperform only CPR while awaiting arrival of EMS personnel [14]. The onlyproblem is that approximately 80% to 85% of all cardiac arrests occur inthe home rather than in a public place [15]. The ongoing NIH-sponsoredHome AED Trial is attempting to determine whether the family membersof high-risk survivors of anterior wall myocardial infarction can save morelives when an AED is present in the home.

Early cardiopulmonary resuscitation

It is unfortunate that early defibrillation alone is not the answer. Weis-feldt and Becker [16] formulated a three-phased model of resuscitationfrom cardiac arrest based on the changing physiologic needs of the patient:electrical, hemodynamic, and metabolic. For the first few minutes after theonset of VF (electrical phase), myocardial cells are rich in ATP. Defibrilla-tion may be all that is needed for successful resuscitation. After 3 to 4 mi-nutes, however, depletion of myocardial ATP prevents the heart fromresuming effective contractions following defibrillation. Attempts at defi-brillation during this period usually result in asystole or pulseless electricalactivity (PEA). A brief period of effective CPR before defibrillation duringthis second (hemodynamic) phase can boost myocardial ATP levels, in-creasing the likelihood of restoration of spontaneous circulation (ROSC)following defibrillation. Finally, if spontaneous circulation is not restoreduntil after 8 to 9 minutes, then a cascade of devastating cellular metabolicevents frequently leads to irreversible end organ injury (including anoxicbrain damage). Cellular reperfusion protection strategies are needed inthis third (metabolic) phase.

Increasing evidence suggests that a brief (90 seconds to 3 minutes) pe-riod of CPR before attempted defibrillation may be beneficial in treat-ing VF patients whose cardiac arrest was not witnessed by firefighters orparamedics. Cobb and colleagues [17] observed that routine provisionof approximately 90 seconds of CPR before firefighter AED use was associ-ated with increased survival when fire or EMS response time intervals were4 minutes or longer compared with using a ‘‘shock first’’ strategy. Wik andcoworkers [18] conducted a randomized clinical trial in 200 patients who hadout-of-hospital VF in Oslo, Norway, to determine whether survival wouldbe better with immediate defibrillation (n ¼ 96) or CPR first, with 3 minutesof basic CPR by ambulance personnel before defibrillation (n ¼ 104). Forpatients with fire or EMS response time intervals of longer than 5 minutes,more patients were resuscitated successfully in the CPR-first group (58%versus 38%, P ¼ 0.04). Survival to hospital discharge was also better in theCPR-first group (22% versus 4%, P ¼ 0.006).

26 ORNATO & PEBERDY

Emergency Medical Service system response

EMS systems have three traditional components: emergency medical dis-patch, first response, and EMS ambulance response.

Emergency medical dispatch

Early access to EMS is promoted by a 9-1-1 system currently available tomore than 95% of the United States population. Enhanced 9-1-1 systemsprovide the caller’s location to the dispatcher, which permits rapid dispatchof prehospital personnel to locations even if the caller is not capable of ver-balizing or the dispatcher cannot understand the location of the emergency.

A major challenge is the widespread proliferation and use of cell phones.Traditional cell phone technology does not provide the location of the callerto an enhanced 9-1-1 center. Instead, such calls are usually answered by thestate police, who then attempt to determine the location of the emergencyand forward the call to the appropriate 9-1-1 center. Such additional stepsoften result in substantial delays in the dispatch of emergency units to thescene. Fortunately, new technology exists that allows triangulation of callerlocation from several cell phone tower locations. This technology is beingphased in throughout the country and will soon offer a solution to this vex-ing problem.

In most communities, law enforcement or public-safety officials are re-sponsible for operating 9-1-1 centers because in most locations, 85% of callsare for police assistance, 10% are for EMS, and 5% are for fire-relatedemergencies. Dispatchers who staff 9-1-1 centers typically have only minimalmedical background and training and usually operate by following writtencards and protocols that in many cases are designed and updated locally.High-performance centers employ EMTs or paramedics who are speciallytrained and certified as emergency medical dispatchers. They, too, operateunder standardized written protocols, but such protocols are usually devel-oped and upgraded at the national level. Such centers typically have intensequality-assurance programs to ensure that emergency medical dispatchersfollow protocols and procedures correctly and consistently, which is partic-ularly true for the prearrival instructions that are given to cardiac arrest by-standers to instruct them on how to perform CPR while awaiting arrival ofemergency personnel (‘‘phone CPR’’).

Even though CPR performed by layperson bystanders improves theodds of neurologically intact survival from OHCA, only about 25% of by-standers in most United States cities are willing to perform CPR (a nota-ble exception is Seattle, Washington, where 50% of layperson bystandersperform CPR) [19–23]. Unwillingness of laypersons to perform mouth-to-mouth ventilation on strangers is a major part of the problem [24,25]. Itis fortunate that increasing evidence suggests that chest compressionsare much more important than artificial ventilation during the first 5 minafter onset of OHCA in adults [23,26–29]. Chest compressions alone cause


changes in intrathoracic pressure that ventilate the victim adequately forthe first 5 to 8 minutes after cardiac arrest. Chest compression–only CPR isbeing adopted for prearrival resuscitation instruction in many EMS systems.

Public-safety first responders

To minimize time to treatment, most communities allow volunteer orpaid firefighters and other first-aid providers to function as first responders,providing CPR and, increasingly, early defibrillation using AEDs untilEMTs and paramedics arrive. Ideally, there should be a sufficient numberof trained personnel so that a trained first responder can be at the victim’sside within 5 minutes of the call.

Fire department personnel functioning as first responders are the back-bone of the public-safety primary response for most United States commu-nities. Most of these personnel are trained as first responders or as basicEMTs. They have the capability to provide lifesaving first aid, CPR, andearly defibrillation using AEDs. Some communities train their firefightersto an advanced life support (ALS) level.

Another approach has been to supplement the fire department and EMSresponse with police officers trained and equipped to use AEDs. Such per-sonnel can further enhance survival from OHCA compared with survivalachieved by conventional EMS services. White and colleagues [30] studiedthe outcome of all consecutive adult patients who had nontraumatic cardiacarrest treated in Rochester, Minnesota, from November 1990 through July1995. In that city, a centralized 9-1-1 center simultaneously dispatched po-lice and an ALS ambulance for suspected cardiac arrest cases. Accurate in-tervals were obtained by synchronizing all defibrillator clocks with the 9-1-1dispatch center clock. The personnel who arrived first delivered the initialshock. In patients to whom shocks were delivered initially by police, para-medics provided additional treatment if needed. Main outcome measureswere time elapsed before delivery of the first shock, ROSC, and survivalto discharge home. Of 84 patients, 31 (37%) were shocked initially by police.Thirteen of the 31 demonstrated ROSC, without need for ALS treatment.All 13 survived to discharge. The other 18 patients required ALS; 5 (27.7%)survived. Among the 53 patients first shocked by paramedics, 15 hadROSC; 14 survived. The other 38 needed ALS treatment; 9 survived. Thecall-to-shock time interval for all patients was shorter in the police groupthan in the paramedic group (5.6 versus 6.3 minutes, P ¼ 0.038). For allpatients, the call-to-shock time interval was shorter in those who hadROSC than in those who needed ALS (5.4 versus 6.3 minutes, P ¼ 0.011).Survival to discharge was 49% (41 of 84), with 18 of 31 (58%) in thepolice group and 23 of 53 (43%) in the paramedic group. The call-to-shock time interval was shorter for survivors than nonsurvivors (5.8versus 6.4, P ¼ 0.020). ROSC or discharge survival was not significantlydifferent between police- and paramedic-shocked patients.

28 ORNATO & PEBERDY

The presence of ROSC after an initial shock and the call-to-shock timeintervals were major determinants of survival, whether the first shocks wereadministered by police or by paramedics. When ROSC occurred after thepatient was first shocked, 27 of 28 (96%) patients survived, whereas 14of 56 (25%) patients who needed additional ALS interventions survived(P ¼ 0.001). This study showed that a high discharge-to-home survivalrate could be obtained when early defibrillation was provided by policeor paramedics. It is the rapidity of defibrillation that determines outcome,irrespective of who delivers the shock. When initial defibrillation attemptsresulted in ROSC, the overwhelming majority of patients survived (96%).Even brief (eg, 1 minute) decreases in the call-to-shock time interval in-creased the likelihood of ROSC from shocks only, with a consequent de-crease in the need for further ALS intervention. Similar results werenoted with the use of law enforcement early defibrillation in Miami-DadeCounty, Florida [10].

Ambulance responders

Most cities and larger suburban areas provide EMS ambulance services,with providers from the fire department, a private ambulance company, orvolunteers. The most common deployment pattern is a tiered system inwhich some of the ambulances are staffed and equipped at the basic EMTlevel, which includes first aid and early defibrillation with AEDs, and otherunits (transporting or nontransporting) are staffed by paramedics or otherintermediate-level EMTs who, in addition to basic care, can start intrave-nous (IV) drips, intubate, and administer medications. In some systems,the advanced providers can also perform 12-lead ECGs, provide externalpacing for symptomatic bradycardia, and other advanced techniques.

Some high-performance EMS systems have only ALS-staffed ambulances(all-ALS systems). Advantages of such systems are that they provide a uni-form standard of care and, surprisingly, can actually lower costs by eliminat-ing the need to dispatch two units in response to a call in which it is not initiallyclear to dispatchers whether the patient needs ALS [31]. The potential disad-vantage of suchmodels is that they typically have a large number of paramed-ics, each of whom gets to perform their advanced skills less frequently than thesmaller number of paramedics typically found in tiered systems [32].

Rural areas typically provide primarily basic life support ambulance ser-vices, usually by volunteers supplemented by a relatively small number ofALSunits. In some cases, ALS is provided by paramedics or helicopter personnelwho respond to the scene in addition to a basic life support ambulance.

Pharmacologic interventions during resuscitation

The most critical pharmacologic intervention during resuscitation in-volves the use of vasoconstrictors to maximize coronary and cerebral


perfusion pressure during CPR. Although vasoconstrictors decrease for-ward cardiac output, oxygenated blood is routed preferentially to the vitalorgans. Sodium bicarbonate, once a mainstay of pharmacologic therapyduring resuscitation, is now believed to be of limited value. Antiarrhythmicmedications are used for patients who have recurrent or refractory ventric-ular tachyarrhythmias and fail to respond to defibrillation alone. Atropinecontinues to be used for patients who have bradyasystole.

Vasoconstrictors

Epinephrine, which has a- and b-adrenergic activity, is still consideredthe vasopressor of choice for use during resuscitation. It improves coronaryand cerebral blood flow by increasing peripheral vasoconstriction and pre-venting arterial collapse. By enhancing coronary perfusion pressure, epi-nephrine facilitates the resynthesis of high-energy phosphates withinmyocardial mitochondria and enhances cellular viability and contractileforce. The increased myocardial blood flow, however, is at least partially an-tagonized by the increased myocardial oxygen consumption caused by epi-nephrine’s b-adrenergic actions [33].

The optimal dose of epinephrine to augment aortic diastolic blood pres-sure in humans during closed chest compression has been debated. Despiteevidence in animal models and anecdotal case series suggesting that higherdoses of epinephrine might be needed compared with those that have beenadministered traditionally, recent prospective, randomized clinical trialshave not shown convincing evidence to support the routine use of ‘‘high-dose’’ (eg, O1 mg in adults) epinephrine [34,35]. The American Heart Asso-ciation currently recommendations a 0.5- to 1.0-mg IV epinephrine dose inadults every 3 to 5 minutes during ongoing resuscitation. The use of higherdoses of epinephrine after the initial 1-mg dose during resuscitation is notrecommended or discouraged.

Vasopressin (40 U as a one-time IV dose) is considered an acceptable alter-native to epinephrine as the first vasopressor administered to adults in VF.Randomized clinical trials have shown vasopressin to be as effective as epi-nephrineduring resuscitationof adults [36,37] andhas the advantageof requir-ing IV administration only every 10 to 15 minutes. There is even a suggestionthat the combination of vasopressin and epinephrine administered in alternat-ing doses may be more beneficial than the use of either drug alone [37,38].

Acid-base management during cardiopulmonary resuscitation

The marked fall in cardiac output during closed chest compression criti-cally reduces tissue oxygen delivery. Cells shift to anaerobic metabolism,gradual building up lactic acid as a waste product. During anaerobic metab-olism, the carbon dioxide concentration increases rapidly inside cells.

Central (mixed) venous blood during closed chest compression is acidotic(pH approximately 7.15) and hypercarbic (PvCO

2approximately 74 mm Hg).

30 ORNATO & PEBERDY

With hyperventilation, carbon dioxide is removed as blood flows throughthe lungs. Accordingly, arterial blood is less acidotic. Arterial blood pHduring well-performed closed chest compression is usually normal, slightlyacidotic, or mildly alkalotic. Arterial blood can be slightly alkalotic, whereasthe venous blood is acidotic because pulmonary blood flow is only onefourth to one third of the normal amount during closed chest compression(this phenomenon has been termed the ‘‘venous paradox’’) [39,40]. It is nota paradox, however, but part of normal physiology that occurs when anaer-obic metabolism is required (eg, during strenuous exercise), at which timethe intramyocardial pH is much closer to the venous pH than the arterialpH.

In the past, administration of sodium bicarbonate was recommendedduring closed chest compression because of the belief that bicarbonatewould buffer the hydrogen ions produced during anaerobic metabolism. So-dium bicarbonate, however, contains a high concentration of carbondioxide (260–280 mm Hg). In plasma, the carbon dioxide is released and dif-fuses into cells more rapidly than bicarbonate, causing a paradoxic rise inintracellular PCO2 and a fall in intracellular pH. The increases in intracellularPCO2 in heart muscle cells decrease cardiac contractility, cardiac output, andblood pressure. Paradoxic acidosis of cerebrospinal fluid also can occur fol-lowing the use of sodium bicarbonate [41] and may be responsible for pro-longed confusion following a successful resuscitation as the venous acidosisincreases. Sodium bicarbonate causes other potentially harmful effects, in-cluding hyperosmolality, alkalemia, and sodium overload.

At present, no convincing data indicate that treatment with sodium bicar-bonate is of benefit during closed chest compression, and it does not im-prove survival in experimental animals. Sodium bicarbonate should notbe administered during a routine cardiac arrest because it provides minimal,if any, benefit and adds significant risk. If used at all, sodium bicarbonateshould not be used until proven interventions (such as defibrillation, cardiaccompression, and support of ventilation including intubation) and pharma-cologic therapies (such as epinephrine and antiarrhythmic agents) have beenemployed. If used, the initial dose of sodium bicarbonate is 1 mEq/kg. Nomore than half of the original dose should be administered every 10 minutesthereafter.

Alternate buffer agents have not yet been shown to improve survival dur-ing cardiac resuscitation. In a recent Scandinavian study, 502 adults whohad asystole or VF with failure of the first defibrillation attempt were en-tered into a prospective, randomized, double-blinded, controlled trial com-paring the use of a combination buffer agent (tribonat, containing 250 mLof a sodium bicarbonate-trometamol-phosphate mixture with a bufferingcapacity of 500 mmol/L) with 250 mL of 0.9% saline placebo [42]. Of thepatients receiving buffer, 87 (36%) were admitted to the hospital and 24(10%) were discharged from the hospital alive; of the patients receivingsaline, 92 (36%) were admitted and 35 (14%) were discharged alive (no


significant difference between groups). This landmark clinical trial lends fur-ther evidence to the belief that buffer therapy does not improve patient out-come from routine cardiac arrest.

Management of specific resuscitation scenarios

The American Heart Association’s ACLS algorithms provide a ‘‘frame-work for dealing with the life-threatening cardiopulmonary emergencies ina logical sequence’’ [43,44].

Recurrent or refractory ventricular fibrillation or ventricular tachycardia

Electrical countershock is the treatment of choice for VF or pulseless VT.If three initial countershocks at increasing energies (200, 200–300, and 360 Jfor monophasic defibrillation, using the manufacturer’s recommended en-ergy levels for biphasic defibrillation), intubation, epinephrine, and a fourthcountershock fail to terminate the arrhythmia (refractory VF or VT) or if, asin many cases, the arrhythmia rapidly recurs (recurrent VF or VT), then an-tiarrhythmic drug therapy is recommended.

For refractory VF and pulseless VT, an initial lidocaine dose of 1.5 mg/kgis suggested, although the evidence for its benefit is marginal. After ROSC,lidocaine is usually continued as an IV infusion at a rate of 30 to 50 mg/kg/min (2–4 mg/min). The need for additional bolus doses of lidocaine is usu-ally guided by clinical response or by plasma lidocaine concentrations.

More data are available on the use of IV amiodarone to treat patientswho have recurrent arrhythmias. Amiodarone is at least as effective as bre-tylium in terminating refractory or recurrent, life-threatening ventricular ta-chyarrhythmias but causes fewer side effects. In a randomized, controlledclinical prehospital trial conducted on 504 cardiac arrest patients who hadrecurrent or refractory ventricular tachyarrhythmias, the administration ofa single 300-mg bolus of IV amiodarone at the time of the first IV epineph-rine administration resulted in a 26% greater rate of survival to hospital ad-mission compared with standard ACLS therapy [45]. In another recentrandomized clinical trial, the use of amiodarone led to substantially higherrates of survival to hospital admission in patients who had shock-resistantOHCA than the administration of lidocaine [46]. Thus, IV amiodarone isquickly becoming the antiarrhythmic drug of choice for use in adults whohave recurrent or refractory VF. The principal side effects of IV amiodaroneare hypotension and bradycardia, which usually respond readily to therapy(volume infusion and vasopressors, and atropine or electrical pacing,respectively).

Bradyasystole

Survival is poor regardless of therapy for cardiac arrest patients whopresent with bradyasystole It is always important to exclude disconnection

32 ORNATO & PEBERDY

of a lead or monitor electrode before concluding that a ‘‘flat line’’ is thepatient’s rhythm because some patients who have such a tracing mayhave VF (a rhythm more amenable to treatment) masquerading as asystole.Whenever there is any doubt, the monitor lead should quickly be switchedto another lead to confirm the diagnosis before treatment. When the diag-nosis is still in doubt, the patient should be presumed to have VF and trea-ted accordingly. Other general measures recommended for the treatment ofbradyasystole include support of ventilation, properly performed closedchest compression, and frequent doses of epinephrine to maintain arterialperfusion pressure and coronary and cerebral perfusion. Treatment withatropine sulfate may improve outcome in patients who have bradyasystoliccardiac arrest that is due to excessive vagal stimulation, but atropine is lesseffective when asystole or pulseless idioventricular rhythms are the result ofprolonged ischemia or mechanical injury in the myocardium.

For patients who have bradyasystolic cardiac arrest, a 1-mg dose of IVatropine is administered and repeated every 3 to 5 minutes if asystole per-sists. Three milligrams (0.04 mg/kg) of IV atropine is a fully vagolyticdose in most adults patients. The administration of a total vagolytic doseof atropine should be reserved for patients who have bradyasystolic cardiacarrest. Endotracheal atropine produces a rapid onset of action similar tothat observed with IV atropine. The recommended adult dose of atropinefor endotracheal administration is 1.0 to 2.0 mg diluted in 10 mL of sterilewater or normal saline.

Pacing (transvenous, transthoracic, or transcutaneous) rarely influencessurvival in the unwitnessed cardiac arrest patient who has asystole or brady-cardia and is initially found without a pulse. Pacing, however, is extremelyuseful for bradycardic patients who have a pulse and in selected patients inwhom a pacemaker can be placed immediately after the development of theconduction disturbance. In such cases, a precordial thump can also stimu-late ventricular complexes and a pulse (‘‘fist pacing’’).

Pulseless electrical activity

PEA is present when there is organized electrical activity on the ECG butno effective circulation, as manifest by a lack of a detectable pulse. There aremany underlying potential causes, but the most common denominator mayinvolve myocardial ischemia and dysfunction due to intramyocardial in-creases in carbon dioxide. Prognosis is generally poor unless a discreteand treatable etiology for PEA can be discerned and corrected. Becauseof the poor prognosis when a correctable etiology cannot be defined,efforts should be directed toward detecting causes such as hypovolemia,hypoxemia, acidosis, tension pneumothorax, and pericardial tamponade.Normal saline should rapidly be infused IV if there is a suspicion of hypo-volemia. Suspected pneumothorax or pericardial tamponade should be con-firmed by needle aspiration of the chest or pericardium, respectively (or by


ECG if available). If confirmed, more definitive surgical management (chesttube or thoracotomy) is usually required.

Scrutiny of the neck veins may be helpful in attempting to define an eti-ology for PEA. Most patients who have cardiac arrest have high, right-sidedfilling pressures and distended neck veins. When neck veins are not visible inthis setting and PEA is present, hypovolemia should be suspected. In thetrauma cardiac arrest victim who has obvious or suspected hemorrhage,the presence of prominent neck veins should lead to the suspicion of pericar-dial tamponade or tension pneumothorax.

General measures such as support of ventilation, properly performedclosed chest compression, and frequent doses of epinephrine to maintain ar-terial perfusion pressure and coronary and cerebral perfusion are recom-mended for treatment of PEA. Bradycardia may be treated with atropine.Although catecholamines are frequently administered, there are no datato suggest a specific benefit (other than improvement in coronary and cere-bral blood flow during closed chest compression). Calcium chloride has notbeen shown to affect clinical survival in controlled trials. If penetrating car-diac trauma is present, then open chest massage can be life saving; other-wise, it is rarely of value.

Emergency department care

An increasing number of EDs are staffed with residency-trained, board-certified emergency physicians. These highly trained and skilled profes-sionals must complete 3 or 4 years of training at a major hospital witha busy ED. Although they spend most of their time training in the ED,they rotate through all major services of the hospital, including anesthesia,cardiology, trauma, and critical care. They are highly skilled and certified inbasic cardiac life support, ACLS, pediatric advanced life support, and ALS.

The longstanding practice of transporting cardiac arrest patients to EDswith CPR in progress has given way to authorizing paramedics to pro-nounce patients deceased in the field when standard ACLS interventionsare unsuccessful [47,48]. As a result, the most common cardiac arrest patienttransported to an ED is one who has been successfully resuscitated in thefield. Thus, the principal focus of ED care is on postresuscitation stabiliza-tion and management.

The most promising intervention at present for postresuscitation care in-volves the induction and maintenance of mild (32�C–34�C) hypothermia for12 to 24 hours following successful resuscitation in comatose OHCA pa-tients whose initial rhythm was VF. Two international randomized clinicaltrials have documented improved survival and neurologic outcome usinghypothermia in such patients [49,50]. The International Liaison Committeeon Resuscitation has issued an advisory statement supporting the use of in-duced mild hypothermia in such patients [51].

34 ORNATO & PEBERDY

Research interventions to improve neurologic outcome

Randomized clinical trials on out-of-hospital cardiac arrest

Induced mild hypothermia has thus far been the only intervention testedin randomized clinical trials that has been shown to be of limited success inimproving neurologic outcome from OHCA [49,50]. In the 1980s, research-ers at the University of Pittsburgh conducted a series of randomized clinicaltrials in an attempt to improve neurologic outcome in patients resuscitatedfrom OHCA. These trials, dubbed the Brain Resuscitation Clinical Trials(BRCT), applied promising pharmacologic interventions as soon as possibleafter successful resuscitation [52,53]. Each of these interventions appeared toshow promise in experimental studies but failed to show benefit in humans.

In BRCT-I, 262 comatose survivors of cardiac arrest were randomly as-signed to receive standard brain-oriented intensive care or the same standardtherapy plus a single IV loading dose of thiopental (30 mg/kg of body weight)[54]. The study was designed to have an 80% probability of detecting a 20%reduction in the incidence of permanent postischemic cerebral dysfunction.Baseline characteristics were similar in the two treatment groups. At theend of 1 year of follow-up, there was no statistically significant difference be-tween treatment groups in the proportion of patients who died (77% of thethiopental group versus 80% of the standard-therapy group), survived with‘‘good’’ cerebral recovery (20% of the thiopental group versus 15% of thestandard-therapy group), or survived with permanent severe neurologic dam-age (2% of the thiopental group versus 5% of the standard-therapy group).None of these differences was statistically significant.

The BRCT-I investigators noted that glucocorticoids were commonly ad-ministered to patients who had global brain ischemia, although their efficacyhad not been proved. The database of BRCT-I was used for a retrospectivereview of the effects of glucocorticoid treatment on neurologic outcome af-ter global brain ischemia [55]. The analysis included 262 initially comatosecardiac arrest survivors who made no purposeful response to pain afterROSC. The standard treatment protocol left glucocorticoid therapy to thediscretion of the hospital investigator, which resulted in four patient groupsreceiving no, low, medium, or high doses of glucocorticoids in the first 8hours after arrest. Neurologic outcome was scored using a modification ofthe Glasgow Cerebral Performance Category Scale. None of the steroid reg-imens statistically improved the mean group survival rate or the neurologicrecovery rate compared with the no-steroid regimen.

BRCT-II tested the value of calcium-entry blockade in preserving neuro-logic function in OHCA survivors [56,57]. Five hundred twenty patients whohad cardiac arrest and remained comatose after ROSC were randomly as-signed to receive three doses of lidoflazine (an experimental calcium-entryblocker) or a placebo and were followed for 6 months. Four patients werelost to follow-up. Treated patients received an IV loading dose (1 mg/kg of


body weight) of lidoflazine and two subsequent doses (0.25 mg/kg) 8 and 16hours after resuscitation. The investigators were blinded to treatment assign-ment. There was no statistically significant difference between the lidoflazinegroup (n ¼ 259) and the placebo group (n ¼ 257) in the proportion of pa-tients who (1) died during the 6-month follow-up (82 versus 83%); (2) sur-vived with good cerebral recovery (15 versus 13%); or (3) survived withsevere neurologic deficit (1.2 versus 1.9%). Analysis of the best level of recov-ery achieved at any time during follow-up also did not show a significant dif-ference between the treatment groups: 24% of those administered lidoflazineand 23% of those administered placebo recovered good cerebral function(normal or only moderately disabled cerebral performance) at some time.

Despite these disappointing results, the NIH recently funded a landmark$50 million research project intended to test the most promising resuscita-tion drugs and devices in a prehospital environment. The Resuscitation Out-comes Consortium involves 10 major United States and Canadian EMSsystems that will perform randomized clinical trials research over the next5 years. Together, the EMS systems in these 10 sites manage over 12,000OHCA patients per year, giving the consortium the statistical power to con-duct definitive clinical trials. The first cardiac arrest project, which will testwhether a new inspiratory threshold valve that improves blood flow duringresuscitation can improve survival and neurologic outcome, is expected tobegin enrolling patients in early 2006.

Exception to informed consent

Most clinical trials that involve research on emergency patients who can-not give prospective informed consent (including all resuscitation research)were halted during the 1990s in theUnited States by a regulatorymoratorium.The issues and problems surrounding the ban were of concern to federalregulators and many researchers for several years. In May 1993, the USFood and Drug Administration (FDA) prematurely terminated a clinical re-suscitation trial involving a comparison of the Ambu CardioPump (Ambu,Denmark) and standard CPR. The FDA believed that the device involved‘‘significant risk’’ to the subjects and that there were no means for the sub-jects to give informed consent to participate in the trial [58,59]. The rulesand regulations of the FDA or the Department of Health and Human Ser-vices (DHHS) did not provide a legal means for Institutional Review Boardsto approve emergency medical research in humans when it is impossible toobtain informed consent. On October 25, 1994, the Coalition Conference ofAcute Resuscitation and Critical Care Researchers was held in Washington,DC, to discuss informed consent in emergency research. Representativesfrom more than 20 organizations explored the issues and produced consen-sus recommendations for resolving some of the difficult issues that surroundemergency medical research.

36 ORNATO & PEBERDY

These recommendations and many other ideas were presented and dis-cussed at the FDA/NIH Public Forum on Informed Consent in Clinical Re-search Conducted in Emergency Circumstances, held in Bethesda, Maryland,on January 9 and 10, 1995. Soon thereafter, the FDAandDHHS released newrules that allowed limited resumption of resuscitation research in humans us-ing an ‘‘exception to informed consent’’ process. Although initial experiencewith the rules was problematic, the public-access defibrillation trial demon-strated that this exception to informed consent process could be used success-fully in a wide array of clinical practice settings [14].

Summary

Considerable progress has been made in providing high-quality prehospi-tal and emergency cardiac care forOHCAvictims. The use of early CPR, earlydefibrillation, early ACLS, and state-of-the-art postresuscitation care offersthe best promise for improved community survival and neurologic outcomestatistics in the future. The NIH-sponsored Resuscitation Outcomes Consor-tium represents the largest governmentally sponsored effort of its kind thatthat will test the value of promising pharmacologic and device interventionson improving survival and neurologic outcome in OHCA patients.

References

[1] Centers for Disease Control and Prevention (CDC). State-specific mortality from sudden

cardiac deathdUnited States, 1999. MMWR Morb Mortal Wkly Rep 2002;51(6):123–6.

[2] Nichol G, Laupacis A, Stiell IG, et al. Cost-effectiveness analysis of potential improvements

to emergency medical services for victims of out-of-hospital cardiac arrest. Ann EmergMed

1996;27(6):711–20.

[3] Fornes P, LecomteD,NicolasG. [Sudden coronary death outside of hospital: a comparative

autopsy study of subjects with and without previous cardiovascular diseases]. Arch Mal

Coeur Vaiss 1994;87(3):319–24.

[4] Bayes de LunaA, Coumel P, Leclercq JF. Ambulatory sudden cardiac death: mechanisms of

production of fatal arrhythmia on the basis of data from 157 cases. Am Heart J 1989;117:

151–9.

[5] Cummins RO, Ornato JP, Thies WH, et al. Improving survival from sudden cardiac arrest:

the ‘‘chain of survival’’ concept. A statement for health professionals from the Advanced

Cardiac Life Support Subcommittee and the Emergency Cardiac Care Committee, Ameri-

can Heart Association. Circulation 1991;83(5):1832–47.

[6] Weaver WD, Cobb LA, Hallstrom AP, et al. Factors influencing survival after out-

of-hospital cardiac arrest. J Am Coll Cardiol 1986;7:754.

[7] Mosesso VN Jr, Newman MM, Ornato JP, et al. Law Enforcement Agency Defibrillation

(LEA-D): proceedings of the National Center for Early Defibrillation Police AED Issues

Forum. Prehosp Emerg Care 2002;6(3):273–82.

[8] White RD, Hankins DG, Bugliosi TF. Seven years’ experience with early defibrillation by

police and paramedics in an emergency medical services system. Resuscitation 1998;39(3):

145–51.

[9] Valenzuela TD, RoeDJ, Nichol G, et al. Outcomes of rapid defibrillation by security officers

after cardiac arrest in casinos. N Engl J Med 2000;343(17):1206–9.


[10] Myerburg RJ, Fenster J, Velez M, et al. Impact of community-wide police car deployment

of automated external defibrillators on survival from out-of-hospital cardiac arrest.

Circulation 2002;106(9):1058–64.

[11] O’Rourke MF, Donaldson E, Geddes JS. An airline cardiac arrest program. Circulation

1997;96(9):2849–53.

[12] Page RL, Joglar JA, Kowal RC, et al. Use of automated external defibrillators by a US air-

line. N Engl J Med 2000;343(17):1210–6.

[13] Caffrey SL, Willoughby PJ, Pepe PE, et al. Public use of automated external defibrillators.

N Engl J Med 2002;347(16):1242–7.

[14] Hallstrom AP, Ornato JP, Weisfeldt M, et al. Public-access defibrillation and survival after

out-of-hospital cardiac arrest. N Engl J Med 2004;351(7):637–46.

[15] Becker L, EisenbergM, Fahrenbruch C, et al. Public locations of cardiac arrest. Implications

for public access defibrillation. Circulation 1998;97(21):2106–9.

[16] WeisfeldtML, Becker LB. Resuscitation after cardiac arrest: a 3-phase time-sensitivemodel.

JAMA 2002;288(23):3035–8.

[17] Cobb LA, Fahrenbruch CE, Walsh TR, et al. Influence of cardiopulmonary resuscitation

prior to defibrillation in patients with out-of-hospital ventricular fibrillation. JAMA 1999;

281(13):1182–8.

[18] Wik L, Hansen TB, Fylling F, et al. Delaying defibrillation to give basic cardiopulmonary

resuscitation to patients with out-of-hospital ventricular fibrillation: a randomized trial.

JAMA 2003;289(11):1389–95.

[19] Cummins RO, Eisenberg MS, Hallstrom AP, et al. Survival of out-of-hospital cardiac

arrest with early initiation of cardiopulmonary resuscitation. Am J Emerg Med 1985;3(2):

114–9.

[20] Eisenberg MS, Hallstrom AP, Carter WB, et al. Emergency CPR instruction via telephone.

Am J Public Health 1985;75(1):47–50.

[21] Kellermann AL, Hackman BB, Somes G. Dispatcher-assisted cardiopulmonary resuscita-

tion. Validation of efficacy. Circulation 1989;80(5):1231–9.

[22] Valenzuela TD, Spaite DW, Meislin HW, et al. Emergency vehicle intervals versus collapse

to CPR and collapse to defibrillation intervals: monitoring emergency medical services sys-

tem performance in sudden cardiac arrest. Ann Emerg Med 1993;22(11):1678–83.

[23] VanHoeyweghen RJ, Bossaert LL,Mullie A, et al. Quality and efficiency of bystander CPR.

Belgian Cerebral Resuscitation Study Group. Resuscitation 1993;26(1):47–52.

[24] Ornato JP. Should bystanders perform mouth-to-mouth ventilation during resuscitation?

Chest 1994;106(6):1641–2.

[25] Ornato JP, Hallagan LF, McMahon SB, et al. Attitudes of BCLS instructors about mouth-

to-mouth resuscitation during the AIDS epidemic. Ann Emerg Med 1990;19:151–6.

[26] Wik L, Kramer-Johansen J, Myklebust H, et al. Quality of cardiopulmonary resuscitation

during out-of-hospital cardiac arrest. JAMA 2005;293(3):299–304.

[27] Kern KB, Hilwig RW, Berg RA, et al. Importance of continuous chest compressions during

cardiopulmonary resuscitation: improved outcome during a simulated single lay-rescuer sce-

nario. Circulation 2002;105(5):645–9.

[28] Hallstrom A, Cobb L, Johnson E, et al. Cardiopulmonary resuscitation by chest compres-

sion alone or with mouth-to-mouth ventilation. N Engl J Med 2000;342(21):1546–53.

[29] Berg RA, Kern KB, Sanders AB, et al. Bystander cardiopulmonary resuscitation. Is ventila-

tion necessary? Circulation 1993;88(4 Pt 1):1907–15.

[30] White RD, Asplin BR, Bugliosi TF, et al. High discharge survival rate after out-of-hospital

ventricular fibrillation with rapid defibrillation by police and paramedics. Ann Emerg Med

1996;28(5):480–5.

[31] Wilson BD, Graton MC, Overton J, et al. Unexpected ALS procedures on non-emergency

ambulance calls: the value of a single tier system. Prehosp Disaster Med 1991;6:382.

[32] Stout J, Pepe PE, Mosesso VN Jr. All-advanced life support vs tiered-response ambulance

systems. Prehosp Emerg Care 2000;4(1):1–6.

38 ORNATO & PEBERDY

[33] Ditchey RV, Lindenfeld J. Failure of epinephrine to improve the balance between myocar-

dial oxygen supply and demand during closed-chest resuscitation in dogs. Circulation 1988;

78(2):382–9.

[34] Brown CG, Martin DR, Pepe PE, et al. A comparison of standard-dose and high-dose epi-

nephrine in cardiac arrest outside the hospital. The Multicenter High-Dose Epinephrine

Study Group. N Engl J Med 1992;327(15):1051–5.

[35] Stiell IG, Hebert PC, Weitzman BN, et al. A study of high-dose epinephrine in human CPR.

N Engl J Med 1992;327:1047–50.

[36] Stiell IG, Hebert PC, Wells GA, et al. Vasopressin versus epinephrine for inhospital cardiac

arrest: a randomised controlled trial. Lancet 2001;358(9276):105–9.

[37] Wenzel V, Krismer AC, Arntz HR, et al. A comparison of vasopressin and epinephrine for

out-of-hospital cardiopulmonary resuscitation. N Engl J Med 2004;350(2):105–13.

[38] Lurie KG, Voelckel WG, Iskos DN, et al. Combination drug therapy with vasopressin,

adrenaline (epinephrine) and nitroglycerin improves vital organ blood flow in a porcine

model of ventricular fibrillation. Resuscitation 2002;54(2):187–94.

[39] GrundlerW,WeilMH,RackowEC.Arteriovenous carbon dioxide and pHgradients during

cardiac arrest. Circulation 1986;74(5):1071–4.

[40] Grundler W, Weil MH, Yamaguchi M, et al. The paradox of venous acidosis and arterial

alkalosis during CPR. Chest 1984;86:282.

[41] Berenyi KJ, Wolk M, Killip T. Cerebrospinal fluid acidosis complicating therapy of exper-

imental cardiopulmonary arrest. Circulation 1975;52:319–24.

[42] Dybvik T, Strand T, Steen PA. Buffer therapy during out-of-hospital cardiopulmonary re-

suscitation. Resuscitation 1995;29(2):89–95.

[43] TheAmericanHeart Association in collaboration with the International LiaisonCommittee

on Resuscitation. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Car-

diovascular Care. Part 6: advanced cardiovascular life support: 7C: a guide to the Interna-

tional ACLS algorithms. Circulation 2000;102(8 Suppl):I142–57.

[44] TheAmericanHeart Association in collaboration with the International LiaisonCommittee

on Resuscitation. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Car-

diovascular Care. Part 6: advanced cardiovascular life support: 7B: understanding the algo-

rithm approach to ACLS. Circulation 2000;102(8 Suppl):I140–1.

[45] Kudenchuk P. Amiodarone in out-of-hospital Resuscitation of Refractory Sustained Ven-

tricular Tachyarrhythmias (ARREST) study. American Heart Association Scientific Ses-

sions Late-Breaking Clinical Trials Presentation. 1997.

[46] Dorian P, Cass D, Schwartz B, et al. Amiodarone as compared with lidocaine for shock-

resistant ventricular fibrillation. N Engl J Med 2002;346(12):884–90.

[47] Hick JL, Mahoney BD, Lappe M. Factors influencing hospital transport of patients in con-

tinuing cardiac arrest. Ann Emerg Med 1998;32(1):19–25.

[48] Heckerson EW. Termination of field resuscitation. Emerg Med Serv 1997;26(8):51–7.

[49] Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital

cardiac arrest with induced hypothermia. N Engl J Med 2002;346(8):557–63.

[50] The Hypothermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to

improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346(8):549–56.

[51] Nolan JP,Morley PT, VandenHoek TL, et al. Therapeutic hypothermia after cardiac arrest:

an advisory statement by theAdvancedLife Support TaskForce of the International Liaison

Committee on Resuscitation. Circulation 2003;108(1):118–21.

[52] Edgren E, Hedstrand U, Kelsey S, et al. Assessment of neurological prognosis in comatose

survivors of cardiac arrest. BRCT I Study Group. Lancet 1994;343(8905):1055–9.

[53] Abramson NS, Safar P, Kelsey SF, et al. Clinical trials and cerebral resuscitation research.

Ann Emerg Med 1984;13(9 Pt 2):868–72.

[54] Brain Resuscitation Clinical Trials Group. Randomized clinical study of thiopental loading

in comatose survivors of cardiac arrest. Brain Resuscitation Clinical Trial I Study Group.

N Engl J Med 1986;314(7):397–403.


[55] Jastremski M, Sutton-Tyrrell K, Vaagenes P, et al. Glucocorticoid treatment does not im-

prove neurological recovery following cardiac arrest. Brain Resuscitation Clinical Trial I

Study Group. JAMA 1989;262(24):3427–30.

[56] Brain Resuscitation Clinical Trials Group. A randomized clinical trial of calcium entry

blocker administration to comatose survivors of cardiac arrest.Design,methods, and patient

characteristics. The Brain Resuscitation Clinical Trial II Study Group. Control Clin Trials

1991;12(4):525–45.

[57] Brain Resuscitation Clinical Trials Group. A randomized clinical study of a calcium-entry

blocker (lidoflazine) in the treatment of comatose survivors of cardiac arrest. Brain Resus-

citation Clinical Trial II Study Group. N Engl J Med 1991;324(18):1225–31.

[58] Lurie KG, Shultz JJ, CallahamML, et al. Evaluation of active compression-decompression

CPR in victims of out-of-hospital cardiac arrest. JAMA 1994;271(18):1405–11.

[59] OlsonCM.The letter or the spirit: consent for research in CPR. JAMA1994;271(18):1445–7.

Neurol Clin 24 (2006) 41–59

Intensive Care After Resuscitation fromCardiac Arrest: A Focus on Heart

and Brain Injury

Steven P. Schulman, MDa,*,Tracey K. Hartmann, LPN, CCRPb,

Romergryko G. Geocadin, MDc

aDepartment of Medicine (Cardiology), Johns Hopkins University School of Medicine,

Baltimore, MD, USAbDivision of Brain Injury Outcomes, Johns Hopkins Medical Institutions,

Baltimore, MD, USAcDepartments of Neurology, Neurosurgery and Anesthesiology–Critical Care Medicine,

Johns Hopkins University School of Medicine, Baltimore, MD, USA

Although national statistics are not available, community-wide studiessuggest only a minority of patients have return of spontaneous circulation(ROSC) after an out-of-hospital cardiac arrest. Of the estimated 350,000to 450,000 out-of-hospital cardiac arrests, 100,000 patients have an at-tempted resuscitation. Of these, 40,000 patients have ROSC and are admit-ted to ICU. Half of these patients survive the hospitalization and anotherhalf of this group survive without major neurologic sequelae. Therefore,less than 3% of all patients who have out-of-hospital cardiac arrests haveROSC, survive the hospitalization, and have a reasonable functional recov-ery [1]. The fact that many patients who have ROSC ultimately die or fail tohave favorable neurologic recovery, suggests that processes that occur afterhospitalization, especially in an ICU, have an impact on survival and neu-rologic recovery. This article addresses the acute care, with emphasis onthe cardiac and neurologic aspects, that patients who have postcardiac ar-rest are provided in the cardiac ICU.

Dr. Geocadin is supported in part by National Institutes of Health (NIH) grants R44-

NS-38016 and RO1-HL-EB 71568 and Ms. Hartmann is supported in part by NIH grant

R44-NS-38016.

* Corresponding author. Johns Hopkins Hospital, 600 North Wolfe Street, Carnegie 568,

Baltimore, MD 21287.

E-mail address: [email protected] (S.P. Schulman).




42 SCHULMAN et al

The factors that influence survival in a prehospital setting are wellknown, including whether or not the arrest is witnessed and the rapidityin which resuscitative efforts, including defibrillation, are initiated. (Seethe article by Ornato and Peberdy elsewhere in this issue for discussion ofthese factors.) Much less is known about the factors that influence survivaland neurologic recovery during the first hours and days of ICU evaluationand management of patients who have sudden cardiac death and who haveROSC. In spite of the fact that there has been recent improved short-termsurvival from out-of-hospital cardiac arrest to hospital admission, the hos-pital survival rate with favorable neurologic outcomes has been unchangedduring the past several years [2].

The lack of national or worldwide guidelines results in marked variabilityin themanagement of patients who have sudden cardiac death andROSC andwho are admitted to an ICU. As anticipated, this variability affects outcome.In a study from Sweden [3], a single emergency medical services unit admittedto two hospitals, with similar prehospital care for sudden cardiac death. Inthis observational series, 579 patients were admitted alive after cardiac arrestto one hospital and 459 patients were admitted alive after cardiac arrest toa second neighboring hospital during a concurrent time interval. Survivalwas significantly different between the two hospitals. The hospital with im-proved survival to discharge after ROSC for out-of-hospital cardiac arresthad a more aggressive approach to patient management, including a higherpercentage of patients undergoing coronary angiography, echocardiography,electrophysiologic studies, and stress testing. These data suggest that thecourse of patients who have ROSC after an out-of-hospital cardiac arrestand are admitted to an ICU is affected by the level of care when hospitalized.

Cardiac arrest: etiology and severity

Etiology of cardiac arrest

Autopsy studies show that most sudden death survivors have structuralheart disease, with atherosclerotic coronary artery disease by far the mostcommon underlying substrate, seen in approximately 80% of sudden deathvictims [1,4,5]. In addition to atherosclerotic coronary disease, autopsy evi-dence of plaque rupture in men and plaque erosion in women who have sub-sequent coronary thrombosis is the underlying pathology in the majority ofcases of sudden cardiac death [6]. Given the frequent prevalence of coronarydisease, the majority of patients admitted to the ICU should undergo coro-nary angiography to define the coronary anatomy with possible revasculari-zation at some point during the hospitalization. The American HeartAssociation and American College of Cardiology recommend coronary an-giography in all survivors of sudden cardiac death [7]. The timing of coronaryangiography depends on whether or not there is evidence of acute myocardialinfarction or hemodynamic instability and the overall neurologic prognosis.

43INTENSIVE CARE AFTER RESUSCITATION FROM CARDIAC ARREST

In patients who do not have evidence of acute ST-segment elevation, myocar-dial infarction, or cardiogenic shock, delay until a neurologic prognosis is de-termined seems reasonable.

Other underlying pathologic substrates found in cardiac arrest victims arelisted in Box 1. These include nonischemic dilated cardiomyopathy; infiltra-tive cardiomyopathies; primary electrical abnormalities, such as long QTsyndrome; presence of drugs that prolong the QT interval; electrolyte abnor-malities; and toxins. In addition to cardiac substrates, there are severalother causes of sudden death and ventricular arrhythmias. These includepulmonary causes, such as pulmonary embolism, respiratory arrest followed

Box 1. Underlying substrates for out-of-hospital cardiac arrest

1. Coronary artery diseaseAcute myocardial infarctionChronic ischemic cardiomyopathyCoronary vasospasm or dissectionAnamolous coronary artery

2. Nonischemic heart diseaseDilated cardiomyopathyHypertrophic cardiomyopathyArrythmogenic right ventricular dysplasiaInfiltrative cardiomyopathyMyocarditisValvular heart disease (aortic stenosis)

3. Primary electrical abnormalitiesLong QT syndromeBrugada syndromeWolff-Parkinson-White syndromeIdiopathic ventricular tachycardia

4. Drug or toxin inducedCocaineProarrhythmia from antiarrhythmic medicationsQT interval–prolonging drugs, such as erythromycinantibiotics and psychotropic medications

5. Electrolytic or metabolicThyrotoxicosisPoisoningHypokalema, hypomagnesemia

6. MechanicalPulmonary embolismTension pneumothoraxTrauma

44 SCHULMAN et al

by cardiac arrest from pneumonia, and so forth. In addition to mimickingacute myocardial infarction, primary neurologic events, such as subarach-noid hemorrhage, also can cause neurogenic cardiac injury leading to ven-tricular fibrillation and cardiac arrest (Fig. 1) [8]. Because subsequenttherapy depends to a great degree on the cause of the cardiac arrest, suchas heparin for a pulmonary embolism, direct percutaneous coronary inter-vention for ST-segment elevation myocardial infarction, an evaluation todetermine the underlying substrate of sudden cardiac death is importantin the early management of these patients.

Cardiac arrest and brain injury

The duration of cardiac arrest as the clinical marker of global ischemia iscorrelated highly with brain injury [9–12]. The precise duration of cardiacarrest with the cessation of blood flow to the brain represents the primaryinsult and is one of the most important clinical factors in determining theseverity of the brain injury [10,12]. During the Brain Resuscitation ClinicalTrials (BRCT), a duration of cardiac arrest of 6 minutes or longer and aresuscitation time to achieve ROSC of 28 minutes or longer indicate poorneurologic recovery. Shorter cardiac arrest times and resuscitation times, in-dicating lesser injury, are associated with favorable outcomes. A Europeanstudy reports a similar observation of patients having favorable outcomeswith short cardiac arrest times (4.1 minutes) and unfavorable outcomes

Fig. 1. Electrocardiogram of a 50-year-old woman who presented to the emergency room with

an out-of-hospital cardiac arrest from ventricular fibrillation. She was resuscitated and the ini-

tial postresuscitation electrocardiogram is shown. She went to emergent cardiac catheterization

where an intra-aortic balloon pump was placed for cardiogenic shock. Coronary angiography

demonstrated normal coronary arteries in spite of the dramatic ST-segment elevation at the

time of angiography. Upon return to the coronary care unit, neurologic examination revealed

fixed and dilated pupils. An emergent head CT showed a massive subarachnoid hemorrhage.


with cardiac longer arrest time (8.0 minutes) [11]. In the same study, a resus-citation time leading to achieving ROSC with an average of 17 minutes wasmore likely to result in a favorable outcome than resuscitation time greaterthan 34.5 minutes [11].

Clinical evaluation

Cardiac and systemic evaluation

Postresuscitation cardiac examination should focus on blood pressure,heart rate, and clinical evidence of hypoperfusion, such as cool extremitiesand oliguria. Lung examination for edema also is important. Clinical car-diogenic shock is evident with blood pressure less than 90 mm Hg, pulmo-nary edema on examination, and evidence of hypoperfusiondall in thesetting of acute myocardial infarction. Finally, detection of heart murmurs,such as aortic stenosis, and a comprehensive neurologic evaluation add im-portant information. The ICU team and neurologist often work as a team inthe early assessment of patients.

An electrocardiogram not only shows evidence of ischemia but alsoshould be evaluated for prolongation of the QT interval, delta waves inWolff-Parkinson-White syndrome, right ventricular overload seen with pul-monary embolism, hypertrophy and pseudoinfarction pattern seen in hyper-trophic cardiomyopathy, and low voltage pattern with atrial enlargementseen with infiltrative cardiomyopathy.

Laboratory tests focusing on the cardiac cause, including standard electro-lytes and magnesium, are determined and optimized. A toxicology screenshould be sent on many patients, if the diagnosis is not immediately evident.Cocaine use is associated with cardiac arrest and myocardial infarction, thetreatment of which differs from the treatment of typical coronary artery dis-ease [13]. After resuscitation, most patients have evidence of cardiac enzymeelevation, including creatine kinase and troponin I or T. The sensitivity andspecificity of troponin to diagnose myocardial infarction after successfulROSC in patients who have cardiac arrest are 96%and 80%, respectively [14].

For most patients admitted to an ICU after resuscitation for cardiac ar-rest, an echocardiogram early in the hospital course often gives useful infor-mation, such as overall left ventricular function; a wall motion abnormalityconsistent with myocardial infarction; valvular abnormalities, such as aorticstenosis; and the presence of a pericardial effusion. This information not onlygives the clinician information about the possible cause of the cardiac arrestbut also assists in management, particularly if hypotension is present.

Other potentially treatable conditions associated with cardiac arrestinclude acidosis, toxins, cardiac tamponade, moderate to severe hypo-thermia, hypoxia, poisoning, hyperkalemia, pulmonary embolism, and ten-sion pneumothorax [1]. Routine testing, including examination, electrolytes,blood gas, and chest radiograph, often results in a correct diagnosis.

46 SCHULMAN et al

Another frequent diagnostic test obtained in the first 24 hours, depending onthe clinician’s index of suspicion, is CT imaging of the brain and lungs.

Neurologic evaluation

Many studies describe functional recovery in relation to the neurologicclinical function of patients resuscitated from cardiac arrest [15–20]. Manyof the clinical findings of these studies are incorporated into the bedsidepractice of prognosticating functional outcome in this patient population.The interventions provided to patients in the studies describe the evolutionof neurologic recovery, because the vast majority of these patients were sub-jected only to normothermic conditions. Therefore, the data derived froma normothermic population can be applied properly only to patients whoare normothermic. With the development of effective therapies, such as hypo-thermia (see the article by Bernard elsewhere in this issue) and other poten-tial therapies still being studied (see the article by Popp and Bottigerelsewhere in this issue), caution must be used when applying data from pre-vious studies of dissimilar patient populations and interventions.

When evaluating neurologic injury in patients resuscitated from cardiac ar-rest, a complete bedside neurologic evaluation is essential consisting of eval-uation of mental status, which needs to address patients’ ability to arouse andengage in ameaningful interaction with the examiner. The cranial nerves haveto be assessed appropriately in responsive and in unresponsive patients. Cra-nial nerve function and sensorimotor and other reflexes may provide criticalinsight into the extent of injury in unresponsive patients. Although the auto-nomic system may be affected largely by the cardiac arrest, certain neurolog-ically relevant manifestations of the autonomic system are worth observing,such as the patterns of breathing, temperature, heart rate, and blood pressure.Breathing patternsmay indicate injury to specific areas of injury at the level ofthe brainstem, whereas the occurrence of bradycardia and hypertension maysuggest intracranial pressure (ICP) elevation or Cushing’s reflex.

As clinical indicators of functional outcome, the neurological examina-tion must be taken in the proper context of the overall clinical picture.Parameters that may confound findings on physical examination as predic-tors of poor outcome must be taken into consideration, however, includingmedications, especially sedatives and illicit drugs used before arrest; hypo-tension; focal cerebral ischemia; seizures; electrolyte abnormalities; hepaticor renal failure; and acidosis.

In 1985, a landmark study was undertaken by Levy and colleagues de-scribing the neurologic findings of patients who were comatose after resus-citation after cardiac arrest [16]. A similar study was undertaken in 1994, inthe multicenter BRCT, in relation to the assessment of neurologic prognosisin comatose survivors of cardiac arrest [18]. With more studies on this sub-ject, Zandbergen and colleagues [15] provide a systematic review of the pre-diction of poor outcome in anoxic-ischemic coma.


In patients who are unresponsive after resuscitation from cardiac arrest,several neurologic findings may show some predictive value for functionaloutcomes at various times during the recovery period. Pupillary light reflex,brainstem reflexes, and motor response to pain are the best studied and mosthelpful clinical predictors of outcome [15,16,18]. In 1985, Levy and col-leagues reported a 100% positive predictive value for predicting severe orworse outcomes if the pupillary light reflex was absent on the initial exam-ination after resuscitation [16]. Subsequent studies show that that lack ofpupillary light reflex immediately after resuscitation has a low specificityand not always is indicative of poor outcome [18–20]. As the absence ofthe pupillary light reflex in these patients becomes more persistent, especiallyon or after 3 days, the likelihood of a poor outcome approaches 100% [15].Brainstem dysfunction as manifested by the absence of two or more brain-stem reflexes (pupillary light response, corneal reflex and occulocephalicreflex) for more than 6 hours after arrest also are highly predictive ofpoor outcome [15]. Lack of oculocephalic reflex after 8 hours is highly pre-dictive of poor outcome and its specificity improves at 24 hours [17].

Motor response to painful stimuli consistently is shown to be a reliablecomponent of the physical examination of unresponsive patients [21,22].Edgren and colleagues in 1994 [18] reported that lack of any motor responseto painful stimuli at 3 days after arrest was the best and only independentpredictor of poor outcome that could be identified.

Similar findings were reported [16] in patients who were unresponsive at3 days and had no withdrawal or flexor motor response to pain. Investigatorshave tested the predictive value of the Glasgow Coma Scale (GCS) and findthat a GCS score less than 5 for more than 2 to 3 days and the persistence ofa GCS score greater than 8 for more than 1 week also are predictors of pooroutcome [18,23]. GCS score reaches greatest specificity at 3 days [15].

Although 3 days is the traditional minimum time of observation in relationto the absence of pupillary light reflex andmotor response, a meta-analysis byBooth and colleagues in 2004 reported that the absence of 5 clinical signs (ab-sent corneal reflexes at 24 hours, absent pupillary response at 24 hours, absentwithdrawal response to pain at 24 hours, no motor response at 24 hours, andno motor response at 72 hours) is sufficient to predict death or poor outcomeas early as 24 hours [24]. The investigators also suggest that although usefulsigns occur at 24 hours after cardiac arrest, an earlier prognosis should not bemade by clinical examination alone [24].

The absence of findings (pupillary light reflex, brainstem reflex, and motorresponse) may help determine poor functional outcome at 24 or 72 hours.This information is used routinely to aid in the decision of level and durationof care provided to these patients. The observation period is at least 24 hours;therefore, these parameters may have no relevance to neuroprotective thera-pies that ideally are provided acutely (within 24 hours) in the period of car-diac arrest and resuscitation. Also, no clinical findings are widely validatedthat provide health care providers with information that indicates the

48 SCHULMAN et al

likelihood of a favorable outcome in patients. Therefore, there still is a needto undertake research and identify early (within minutes or few hours of in-jury) neurologic markers of brain injury and recovery. These markers, if iden-tified, may aid in the development of more effective therapies.

Optimizing survival and functional outcome in the ICU

Coronary revascularization and reperfusion in survivorsof out-of-hospital cardiac arrest

Because the pathophysiology of cardiac arrest often involves plaque rup-ture and thrombosis [6], and because many patients, after resuscitation fromcardiac arrest, evolve a myocardial infarction, emergent revascularizationmay benefit certain patients. This may be beneficial particularly in improvingthe left ventricular dysfunction that frequently is present after resuscitation.This concept of emergent coronary revascularization was evaluated in a pro-spective trial where a select group of out-of-hospital sudden death survivorsunderwent emergent angiography and possible percutaneous coronary inter-vention [25]. This study was conducted in Paris, France, where physiciansstaffed the ambulances. Successfully resuscitated patients between 30 and75 years of age were eligible if there was no obvious noncardiac cause andthe patients previously had been well. Of the 1762 cases of suspected out-of-hospital sudden cardiac arrest cases responded to by the ambulance studyteam, only 85 patients were eligible to be transferred for emergent cardiaccatheterization. The majority of patients were excluded because of failureto resuscitate and fatal recurrent cardiac arrest while in transport. Of the84 patients who underwent emergent angiography, 60 patients had significantcoronary disease, with coronary artery occlusion found in 40 patients, andwere treated with coronary angioplasty. Themean ejection fraction was signi-ficantly depressed at 34%. The two independent predictors of coronaryartery occlusion on angiography in this select group of patients who hadout-of-hospital cardiac arrest were ST-segment elevation on the admissionelectrocardiogram and chest pain before the arrest. The presence of one ofthese predictors had positive and negative predictive values for coronary ar-tery occlusion of 0.63 and 0.74, respectively. The presence of chest pain beforethe arrest and ST-segment elevation on the electrocardiogram had a positivepredictive value of 0.87 and negative predictive value of 0.61. Nine of 85 pa-tients who demonstrated an occluded coronary artery on angiography hadneither chest pain preceding the arrest nor ST-segment elevation on the ad-mission electrocardiogram. Predictors of survival to hospital discharge in-cluded absence of inotropic drug support during transport and successfulcoronary angioplasty. A longer time from cardiac arrest to resuscitationwas associated with worse survival. The poor predictive value of clinicaland electrocardiographic data in predicting coronary occlusion on angio-graphy is not surprising, given the frequent lack of history in comatose


patients and the many other causes of cardiac arrest that may be associatedwith significant ST changes [8]. One aspect of this study was that of the largenumber of cardiac arrests screened, only 1 in 20 subjects were taken to cardiaccatheterization. This selection bias plus the lack of any data regardingwhether or not routine percutaneous coronary intervention improves neuro-logic outcomes puts into question whether or not coronary angiography andpercutaneous coronary intervention should be performed on an emergentbasis in all patients surviving out-of-hospital cardiac arrests. As with any in-dication for an invasive procedure, the risks and benefits need to be individ-ualized before proceeding to angiography. If methodologies progress suchthat an early and accurate determination that neurologic recovery in partic-ular sudden death survivors is likely, a more aggressive approach to emergentcoronary angiography likely will evolve.

Patients who have ST-segment elevation on electrocardiogram withhemodynamic compromise, such as cardiogenic shock, should be consideredfor emergent coronary angiography. A randomized trial in patients who hadcardiogenic shock shows that an invasive approach with catheterization andcoronary revascularization improves short- and long-term outcomes [26].Other survivors of cardiac arrest who have ROSC and who should be con-sidered for early angiography with possible coronary revascularization in-clude those who have ST-segment elevation within 12 hours of symptomonset with evidence of or suggestion that neurologic recovery is likely.

Some small series and one underpowered randomized study of 35 pa-tients who had out-of-hospital cardiac arrest evaluate whether or not throm-bolytic therapy at the time of emergency room arrival improves outcomes[27]. The premise for this provocative therapy rests on the role of thrombus(coronary ischemia and pulmonary embolism) in the etiology of out-of-hospital cardiac arrest. Although these small studies suggest a possible ben-efit for some patients, a larger randomized placebo controlled trial shows nobenefit using the fibrinolytic agent, tissue plasminogen activator [28]. In thisstudy, 233 patients who had an out-of-hospital cardiac arrest and demon-strated 1-minute or greater pulseless electrical activity, despite initiation ofstandard cardiopulmonary resuscitation, were randomized to placebo ora 100-mg dose of tissue plasminogen activator. The primary endpoint wassurvival to hospital discharge, which occurred in one patient randomizedto fibrinolytic therapy compared with no patients randomized to placebo.These data show the terrible outcomes in patients who have pulseless elec-trical activity and no benefit for fibrinolytic therapy in this patient group.

Associated risk of brain hemorrhage post systemic thrombolysis

Thrombolysis, either local or systemic, after resuscitation from cardiacarrest is a concern because of the potential increased risk of intracranialbleeding. Several uncontrolled studies find thrombolysis safe for the brain.A retrospective analysis reports on 68 patients who received systemic

50 SCHULMAN et al

thrombolytics after resuscitation from cardiac arrest for presumed acutemyocardial infarction. Cardiac reperfusion was achieved in 71% of the pa-tients treated. Intracranial hemorrhage was reported in only one patient,whereas four others had bleeding outside the central nervous system [29].A study of the use of intravenous thrombolytics as therapy for brain injuryafter cardiac arrest currently is taking place in Europe. (See the article byPopp and Bottiger elsewhere in this issue for discussion of this therapy.)

ICU therapy

After resuscitation from cardiac arrest, admitted patients generally are in-tubated with varying degrees of mechanical ventilatory support and requiregeneral supportive measures in the ICU, such as sedation, deep vein throm-bosis prophylaxis, and stress gastritis prophylaxis. In the absence of ran-domized clinical trials and treatment guidelines for many ICU conditionsin these patients, this article discusses considerations for management basedon existing literature and the authors’ own clinical experience.

Hemodynamic instability

Many patients, after resuscitation from cardiac arrest, have significanthemodynamic lability. More than half of patients who are resuscitated re-quire vasopressor support during the first 72 hours of their hospitalization[30]. Many of the early deaths in the ICU result from circulatory collapse.The cause of the hypotension often is multifactorial, including left ventri-cular dysfunction and inappropriate peripheral vasodilatation. The peri-pheral vasodilatation and left ventricular dysfunction result from theglobal ischemia and reperfusion that occur after cardiac arrest and resusci-tation [31]. During reperfusion from an ischemic episode, as occurs duringcardiac arrest, there is a dramatic release of oxygen-free radicals; activationof cytokines, complement, and neutrophils; and activation of endothelialsurface adhesion molecules. Cytokine levels rise rapidly after resuscitationfrom cardiac arrest. In a series of 61 patients who had out-of-hospital car-diac arrest and ROSC, cytokine levels drawn on admission to the hospitalwere markedly elevated and comparable to the levels seen in patients whohave septic shock [31]. In this study, cytokines, such as interleukin 6(IL-6) and tumor necrosis factor a (TNF-a), correlate with the level of lacticacid on admission. Nonsurvivors and patients requiring vasopressor supporthad significantly higher levels of cytokines, such as IL-6 and TNF-a [31]. Theelevation of cytokines was independent of any evidence of infection. Endo-thelial adhesion molecules for neutrophils also are up regulated rapidly aftertotal body ischemia and reperfusion that occurs with ROSC after cardiacarrest [31]. Therefore, one mechanism of vasodilatory shock after ROSCin many patients who have out-of-hospital cardiac arrest and are admittedto an ICU is a systemic inflammatory response. This response is associated


with marked elevation of cytokines, such as IL-6 and TNF-a, which predictmortality. The correlation between lactic acid and cytokine levels suggeststhat one cause of this systemic inflammatory response is whole-body ische-mia and reperfusion that occurs in cardiac arrest survivors.

Also contributing to circulatory collapse after ROSC after cardiac arrestis inappropriate nitric oxide expression and elaboration [32]. The expressionof inducible nitric oxide synthase (iNOS) that is stimulated by cytokinescontributes to the systemic inflammatory response and inappropriate vaso-dilatation. Observations in patients who have cardiogenic shock complicat-ing an acute myocardial infarction show that systemic vascular resistanceoften is inappropriately low, whereas overall ejection fraction is not im-paired horrendously [33]. After exposure to a variety of cytokines that areproduced after ischemia and reperfusion, iNOS is expressed by many celltypes. This can lead to toxic levels of nitric oxide and its metabolites,such as peroxynitrite. High levels of nitric oxide and nitric oxide metabolitesdirectly inhibit myocardial function, suppress oxidative metabolism, reduceresponsiveness to catecholamines, and induce systemic vasodilatation. Ani-mal models support a role for iNOS and elevated nitric oxide levels contri-buting to the hemodynamic collapse seen in many patients who have ROSCafter cardiac arrest. Inhibition of nitric oxide synthase improves left ventri-cular function in the ischemia and reperfusion mode [34]. Current humantrials with nitric oxide synthase inhibition are ongoing to determine whetheror not cardiogenic shock patients benefit from this therapy.

Hypotension

Transient left ventricular dysfunction, commonly present after ROSCfrom cardiac arrest, also contributes to the hypotension and vasopressorsupport often required in this patient population. This left ventricular dys-function typically lasts 48 to 72 hours, followed by gradual improvement ofthe cardiac output [30]. The dose of epinephrine required for resuscitation isthe primary factor that correlates best with postresuscitation left ventriculartransient dysfunction. These data suggest that high levels of catecholaminescause transient left ventricular dysfunction or stunning and contribute to thetransient shock often present in this patient population. These data are sup-ported by the transient left ventricular dysfunction or stunning recentlydemonstrated in patients who have severe emotional stress [35]. Catechol-amine concentrations in this population also are elevated, suggesting thatendogenous and exogenous catecholamines can cause transient myocardialstunning, with improvement in left ventricular function during the next 48to 72 hours.

In patients who have hypotension, a trial of volume expansion often isbeneficial. Although in certain situations, pulmonary artery catheterizationgives important information, routine management of hypotension witha pulmonary artery catheter generally is not indicated [36,37]. Randomized

52 SCHULMAN et al

trials in postsurgical patients, patients who are postmyocardial infarction,and patients who have critical illness show no benefit and possible worseoutcome in those randomized to pulmonary artery catheterization com-pared with conventional therapy [37]. This adverse outcome may resultfrom misinterpretation of the data from the pulmonary artery catheteriza-tion or overuse of vasopressor medications that can worsen ischemia andleft ventricular function in patients who have depressed cardiac outputs. Se-lective use instead of routine use of a pulmonary artery catheter should beconsidered in patients who have ROSC and cardiogenic shock with hypo-tension, pulmonary edema, and depressed left ventricular function. In pa-tients who do not respond to volume and low-dose vasopressors,placement of a pulmonary artery catheter also should be considered. If vaso-pressors are required for refractory vasodilatory shock, combination ther-apy with modest catecholamine infusion plus vasopressin is preferable tohigh doses of catecholamines. In a study of 46 patients who had vasodila-tory shock requiring norepinephrine, patients were randomized to the addi-tion of vasopressin versus continued escalation of norepinephrine doses. Theformer group had less tachycardia, fewer atrial arrhythmias, improved car-diac output, and better markers of organ perfusion [38].

The impact of hypotension on the progression of neurologic injury mayaffect functional outcome significantly. Persistent systemic hypotensionleading to cerebral hypoperfusion can worsen neurologic outcome andshould be avoided. Due to impairment of cerebral autoregulation in patientsafter cardiac arrest the ideal mean arterial pressure (MAP) range for brainpreservation is not known [39]. Good functional neurologic recovery isassociated positively with higher spontaneous arterial blood pressure duringthe first 2 hours after cardiac arrest [40].

Cardiac arrhythmia

Many patients admitted to an ICU have had an antiarrhythmic agent ini-tiated either in the field or emergency department. A randomized trial showsthat an amiodorone bolus of 300 mg is superior to placebo in shock-resistant ventricular fibrillation in patients suffering out-of-hospital cardiacarrest [41]. An amiodorone bolus also is superior to lidocaine in producingROSC [42] and is the antiarrhythmic agent of choice in shock-resistant ven-tricular fibrillation. In patients who receive amiodorone therapy for out-of-hospital cardiac arrest with ROSC, there is a lack of randomized trial dataconcerning length of therapy with an antiarrhythmic agent. Nevertheless,continuing an intravenous infusion for 24 hours seems reasonable. There-after, the amiodorone generally can be stopped and not resumed unlessrecurrent and sustained ventricular arrhythmias become evident. Mostsudden death survivors have coronary artery disease, with myocardial in-farction present in a large percentage of patients. In addition, an enhancedadrenergic state contributes to ventricular fibrillation. Therefore, beta-


blockade should be initiated on admission to an ICU, if blood pressure andhemodynamics permit [43]. Beta-blockade is contraindicated in patients whohave cardiac arrest and ROSC and who also have hypotension, significantbradyarrhythmias, and severe pulmonary edema.

Sepsis and temperature control

Sepsis occurs frequently in patients who have cardiac arrest. Two com-mon sources of sepsis are aspiration during arrest and abdominal sepsis.The latter arises from the bowel ischemia as a consequence of ischemiaand reperfusion. The majority of patients should have routine cultures per-formed on admission. In patients who are hypotensive after ROSC, physi-cians should consider the initiation of broad-spectrum antibiotics thatcover lung and bowel flora until further culture data are available andhemodynamics improve.

As with sepsis, elevated body temperature occurs often. Increased bodytemperature after cardiac arrest is associated with worse outcome and braindeath [44]. In a study of 40 patients, all 20 patients who had a peak axillarytemperature above 39� within the first 72 hours after cardiac arrest becamebrain dead versus only 3 of 20 patients who had a peak temperature lessthan 39�C [44]. Considerations related to the cause of temperature elevationmay have contributed significantly to the poor outcome, but with recentevidence that hypothermia is beneficial, the prevention of hyperthermiawith routine antipyretics and cooling measures are important clinical inter-ventions. In appropriate situations, therapeutic hypothermia can be insti-tuted during the early period after cardiac arrest [45,46]. (See the articleby Bernard elsewhere in this issue for discussion of this therapy.)

Coagulopathy

Frequently contributing to the postarrest syndrome and also mimickingsepsis is a coagulopathy.Many patients, after ROSC, have marked activationof blood coagulation without adequate activation of endogenous fibrinolysis.This can lead to microvascular thrombosis, resulting in further organ dys-function [47]. The activation of the coagulation system also is related to thelarge increase in cytokines seen in patients who have ROSC after cardiac ar-rest. In a study of 67 patients admitted after out-of-hospital cardiac arrestwith ROSC, measures of cytokines and coagulation were obtained [47]. Pa-tients had increased IL-6, coagulation activity (elevated levels of thrombin-antithrombin complex), reduced anticoagulation (depressed antithrombin,proteinC, and protein S), activated fibrinolysis (elevated plasmin-antiplasmincomplex), and inhibited fibrinolysis (increased plasminogen activatorinhibitor 1 levels). Activation of coagulation and fibrinolysis and reducedanticoagulation at admission were more pronounced in nonsurvivors, in par-ticular patients dying in hospital of refractory shock. These data suggest thatnearly all survivors of out-of-hospital cardiac arrest have a systemic

54 SCHULMAN et al

coagulopathy present on admission to the ICU, such as coagulation activa-tion, diminished anticoagulant factors, or increased fibrinolysis. The coagul-opathy likely originates from cytokine up regulation of tissue factor, knownto be up regulated after cardiopulmonary resuscitation [48]. These coagula-tion abnormalities also may contribute to the high early mortality seen inthose surviving out-of-hospital cardiac arrest. (See the article by Popp andBottiger elsewhere in this issue for discussion of management considerationof this problem.)

Antiplatelet and anticoagulation therapies

Aspirin should be given on admission to the hospital and continued daily,unless contraindicated or clinical data exist that patients do not have ob-structive coronary artery disease [49]. Full-dose anticoagulation with hepa-rin is another consideration for patients after sudden cardiac death withROSC. In general, patients who have possible or definite acute coronarysyndromes from plaque rupture are anticoagulated with heparin for48 hours after admission to the hospital [50]. This therapy reduces recurrentischemic events. Initiation of full-dose anticoagulation depends on theclinical suspicion of an acute plaque rupture event resulting in ischemia, ven-tricular fibrillation, and sudden cardiac death. The bleeding risks of antico-agulation with heparin (chest trauma from cardiopulmonary resuscitation)versus the benefit of full-dose anticoagulation in patients who have acutecoronary syndrome need to be considered carefully by clinicians. Heparinneeds to be held until the coagulopathy, if present, resolves.

Seizures and myoclonus

Seizures and myoclonus are common after cardiac arrest, occurring inapproximately one third of patients [51]. (See the article by Koenig andcolleagues elsewhere in this issue for discussion of these clinical problems.)Clinical or electrographic seizure activity persisting more than 30 to60 minutes usually, but not invariably, is associated with poor outcome[52–55]. Status epilepticus, the persistence of a seizure activity, often is con-sidered a predictor of poor outcome, as is persistent myoclonus. Cautionmust be taken, however, in characterizing epileptic and myoclonic activityproperly after arrest. Postanoxic myoclonus (Lance-Adams syndrome), pre-viously regarded as a predictor of poor outcome, may improve as neurologicstatus improves [56]. (See the article by Venkatesan and Frucht elsewhere inthis issue for description of this clinical condition.) Postanoxic myoclonustends to be more common in patients who have respiratory causes of arrest.A state of status myoclonus, however, defined as more than 30 minutes ofmyoclonic activity associated with burst suppression on electroencephalo-gram (EEG) [53], which occurs commonly after cardiac arrest, is consideredan indicator of extremely poor prognosis, and treatment with antiepilepticstends not to influence short-term or long-term outcomes [53,57]. Status


myoclonus tends to last only a few days. Differentiation between these twostates can be difficult, but it is important to identify the difference. EEGmonitoring and repeated evaluation may be helpful. Myoclonic status epi-lepticus, defined as myoclonic jerks merging with tonic-clonic seizures andlasting more than 30 minutes, is associated with failure to recover conscious-ness [17], although some investigators question this [58].

Seizures can delay the recovery of consciousness after cardiac arrest, andsubclincial status epilepticus can depress the neurologic examination falselyand can be a cause for persisting unresponsiveness. An EEG should be ob-tained if seizures are suspected. Once seizures are found to occur, theyshould be treated aggressively to optimize recovery. Reports on the use ofprophylactic antiepileptics therapy are limited, however, and not well clari-fied in the literature.

Cerebral edema and intracranial pressure elevation

Global cerebral ischemia may lead to brain edema. In one study, up to47% of patients resuscitated from out-of-hospital arrest showed cerebraledema on head CT at day 3 [59]. In another study, more patients (92%)who had cerebral edema on head CT were noted among those who had pri-mary respiratory arrest [59]. Cerebral edema can quantified objectively bythe degree of obliteration gray matter–white matter demarcation by brainCT scan [60]. Torbey and colleagues find that the progressive loss of graymatter–white matter demarcation as a reflection of brain injury is associatedwith poor outcome [60]. In these reports, brain edema is a marker of braininjury and associated with poor neurologic outcomes. (See the article byGeraghty and Torbey elsewhere in this issue for a detailed discussion ofthis approach at prognostication.)

Several small studies have attempted to define the occurrence of intra-cranial pressure elevation after resuscitation from cardiac arrest. In a studyof ICP monitoring starting as early as 3 hours for a period of 2 to 7 days,ICP persistently remained below 20 mm Hg in five of the six patients. ICP el-evation to 57 mm Hg was noted in one patient who had seizure activity. Al-though the study was limited by sample size, the absence of intracranialpathology or seizures made ICP elevation unlikely [61]. Another study showedthat ICP elevation was associated with delayed hyperemia by transcranialDoppler ultrasound after resuscitation from cardiac arrest [62]. Therefore,the use of acute hyperventilation and mannitol therapy may be beneficial atthe time of ICP elevation. These therapies are used successfully in other pathol-ogies, but their use in edema related to global ischemia is not well described.Use of steroids does not provide benefit and can lead to adverse outcomes [63].

Hyperglycemia

Elevated serum glucose is associated with unfavorable outcome afterglobal ischemia from cardiac arrest [63,64]. Serum glucose elevation is

56 SCHULMAN et al

believed to be a marker of the severity of injury. In a series of 145 nondia-betic patients evaluated after witnessed ventricular fibrillation cardiac arrest,a strong association between high median blood glucose levels during24 hours and poor neurologic outcome was found [65].

Elevation of serum glucose after acute neurologic injury may be harmful,therefore treatment of hyperglycemia in patients who have nonlacunarstroke, and global ischemia is advocated [66]. In the absence of controlled hu-man trials showing the benefit of glucose control in patients resuscitated fromcardiac arrest, some insights may be taken from the general critical care liter-ature. Although the precise effect of elevated glucose on neurologic injury ofpatients who have postcardiac arrest remains not well defined, tight glucosecontrol is associated with improved survival and outcome in patients whoare critically ill [67]. Until a dedicated controlled clinical trial in the specificarea is undertaken, glucose monitoring in the ICU is strongly suggested,and providing tight control may provide benefits to patients who have post-cardiac arrest.

Age and systemic complications

Many victims of cardiac arrest are elderly and have serious underlyingcomorbidities [9]. A review of noncardiac complications in cardiac arrestsurvivors notes that the most frequent complications are pneumonia, elec-trolyte abnormalities, and gastrointestinal hemorrhage in approximately45%; followed by seizures and elevated liver enzymes in approximately28%; and septicemia, acute renal failure, and acute respiratory distress syn-drome in approximately 5% to 7% of patients [68]. Although advanced age(older than 65) is a risk factor for decreased overall survival after cardiacarrest, age is not an independent risk factor for poor neurologic outcome[12]; therefore, age consideration should be taken into account as plansfor therapy are undertaken.

Future directions in the ICU

Much of the injury resulting in hemodynamic instability, cardiac dysfunc-tion, and brain injury results from ischemia and reperfusion injury. Animalmodels demonstrate that many therapies given before ischemia and reper-fusion improve outcomes. Naturally, pretreatment is impossible in patientsafter cardiac arrest. Several therapies currently in clinical trial may improvecardiac function in the setting of ischemia and reperfusion [69]. Certainly,therapies already demonstrated to reduce reperfusion injury, such as hypo-thermia, need to be applied more widely. Future technologies that assessneurologic prognosis more readily and accurately will assist in patientselection for early coronary angiography. Continued research in the identi-fication of high-risk patients for sudden cardiac death may result in betterpreventative strategies [70].


Reference

[1] Eisenberg MS, Mengert JT. Cardiac resuscitation. N Engl J Med 2001;344:1304–13.

[2] Peberdy MA, Ornato JP. Post-resuscitation care: is it the missing link in the Chain of

Survival? Resuscitation 2005;64:135–7.

[3] Engdahl J, Abrahamsson P, BangA, et al. Is hospital care of major importance for after out-

of-hospital cardiac arrest? Resuscitation 2000;43:201–11.

[4] Spector PS. Diagnosis and management of sudden cardiac death. Heart 2005;91:408–13.

[5] Pinto DS, Josephson JE. Sudden death. In: Hurst’s the heart, 11th edition. McGraw-Hill

Medical Publishing Divison.

[6] Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to vulnerable patient. Circula-

tion 2003;108:1664–72.

[7] Scanlon JP, Faxon DP, Audet AM, et al. ACC/AHA guidelines for coronary angiography.

J Am Coll Cardiol 1999;33:1756–824.

[8] Hirsch G, Heldman AW, Wittstein IS, et al. ST-segment elevation in an unresponsive

patients. Circulation 2003;208:165.

[9] Grubb N. Managing out-of-hospital cardiac arrest survivors: 1. neurological perspective.

Heart 2001;85:6–8.

[10] BRCT-1.Neurological recovery after cardiac arrest: effect of duration of ischemia. Crit Care

Med 1985;13:930–1.

[11] Berek K, Lechleitner P, Luef G, et al. Early determination of neurologic outcome after pre-

hospital cardiopulmonary resuscitation. Stroke 1995;26:543–9.

[12] Rogove HJ, Safar P, Sutton-Tyrrell K, et al. Old age does not negate good cerebral outcome

after ardiopulmonary resuscitation: analyses from the brain resuscitation clinical trials. The

Brain Resuscitation Clinical Trial I and II Study. Crit Care Med 1995;23:18–25.

[13] Lange RA, Hillis LD. Cardiovascular complications of cocaine use. N Engl JMed 2001;345:

351–8.

[14] MullnerM,HirschlMM,Herkner J, et al. Creatine kinase-MB fraction and cardiac troponin

T to diagnose acute myocardial infarction after cardiopulmonary resuscitation. J Am Coll

Cardiol 1996;28:1220–5.

[15] Zandbergen EGJ, Haan EJ, Hijdra A. Systematic review of prediction of poor outcome in

anoxic-ischemic coma with biochemical markers of brain damage. Intensive Care Med

2001;27:1661–7.

[16] Levy DE, Caronna JJ, Singer BH, et al. Predicting outcome from hypoxic-ischemic coma.

JAMA 1985;10:1420–6.

[17] Jorgensen EO, Holm S. Prediction of neurological outcome after cardiopulmonary resusci-

tation. Resuscitation 1999;41:145–52.

[18] Edgren E, Hedstrand U, Kelsey S, et al. Assessment of neurological prognosis in comatose

survivors of cardiac arrest. BRCT 1 [comments]. Lancet 1994;343:1055–9.

[19] Jorgensen EO. Course of neurological recovery and cerebral prognostic signs during cardio-

pulmonary resuscitation. Resuscitation 1997;35:9–16.

[20] Steen-Hansen JE, Hansen NN, Vaagenes P, et al. Pupil size and light reactivity during

cardiopulmonary resuscitation: a clinical study. Crit Care Med 1988;16:69–70.

[21] Willoughby JO, LeachBR.Relation of neurological findings after cardiac arrest to outcome.

Br Med J 1974;3:437-9.

[22] Bertini G, Margheri M, Giglioli C, et al. Prognostic significance of early clinical manifesta-

tions in postanoxic coma: a retrospective study of 58 patients resuscitated after prehospital

cardiac arrest. Crit Care Med 1989;17:627–33.

[23] Mullie A, Verstringe P, Buylaert W. Predictive value of Glasgow Coma score for awakening

after out-of-hospital cardiac arrest. Lancet 1988;1:137–40.

[24] Booth CM, Boone RH, Tomlinson G, et al. Is this patient dead, vegetative, or severely neu-

rologically impaired? Assessing outcome for comatose survivors of cardiac arrest. JAMA

2004;291:870–9.

58 SCHULMAN et al

[25] Spaulding CM, Joly LM, Rosenberg A, et al. Immediate coronary angiography in survivors

of out-of-hospital cardiac arrest. N Engl J Med 1997;336:1629–33.

[26] Hochman JS, Sleeper L,Webb J, et al. Early revascularization in acutemyocardial infarction

complicated by cardiogenic shock. N Engl J Med 1999;341:625–34.

[27] Fatovich DM, Dobb GJ, Clugston RA. A pilot randomized trial of thrombolysis in cardiac

arrest (the TICA trial). Resuscitation 2004;61:309–13.

[28] Abu-Laban RB, Christenson JM, Innes GD, et al. Tissue plasminogen activator in cardiac

arrest with pulseless electrical activity. N Engl J Med 2002;346:1522–8.

[29] Voipio V, KuismaM, Alaspaa A, et al. Thrombolytic treatment of acute myocardial infarc-

tion after out-of-hospital cardiac arrest. Resuscitation 2001;49:251–8.

[30] Laurent I, Monchi M, Chiche JD, et al. Reversible myocardial dysfunction in survivors of

out-of-hospital cardiac arrest. J Am Coll Cardiol 2002;40:2110–6.

[31] Adrie C, Adib-Conquy M, Laurent I, et al. Successful cardiopulmonary resuscitation after

cardiac arrest as a ‘‘sepsis-like’’ syndrome. Circulation 2002;106:562–8.

[32] Hochman JS. Cardiogenic shock complicating acute myocardial infarction. Circulation

2003;107:2998–3002.

[33] Picard MH, Davidoff R, Sleeper LA, et al. Echocardiographic predictors of survival and

response to early revascularization in cardiogenic shock. Circulation 2003;107:279–84.

[34] Schulz R,Wambolt R. Inhibition of nitric oxide synthesis protects the isolated working rab-

bit heart from ischemia-reperfusion injury. Cardiovasc Res 1995;30:432–9.

[35] Wittstein IS, ThiemannDR, JAC, et-al. Neurohumoral features of myocardial stunning due

to sudden emotional stress. 2003;352:539–48.

[36] Cohen MG, Kelly RV, Kong DF, et al. Pulmonary artery catheterization in acute coro-

nary syndromes: insights from the GUSTO IIb and GUSTO III trials. Am J Med 2005;

118:482–8.

[37] Richard C, Warszawski J, Anguel N, et al. Early use of the pulmonary artery catheter and

outcomes in patients with shock and respiratory distress syndrome. JAMA 2003;290:

2713–20.

[38] Dunser MW, Mayr AJ, Ulmer H, et al. Arginine vasopressin in advanced vasodilatory

shock: a prospective, randomized, controlled study. Circulation 2003;107:2313–9.

[39] Sundgreen C, Larsen F, Herzog T, et al. Autoregulation of cerebral blood flow in patients

resuscitated from cardiac arrest. Stroke 2001;32:128–32.

[40] Mullner M, Sterz F, Binder M, et al. Arterial blood pressure after human cardiac arrest and

neurological recovery. Stroke 1996;27:59–62.

[41] Kudenchuk PJ, Cobb LA, Copass MK, et al. Amiodorone for resuscitation after out-of-

hospital cardiac arrest due to ventricular fibrillation. N Engl J Med 1999;341:871–8.

[42] Dorian P, Cass D, Schwartz B, et al. Amiodorone as compared with lidocaine for shock-

resistant ventricular fibrillation. N Engl J Med 2002;346:884–90.

[43] Hjalmarson A. Effects of beta-blockade on sudden cardiac death during acute myocardial

infarction and the post-infarction period. Am J Cardiol 1997;80:35j–9j.

[44] Takasu A, Saitoh D, Kaneko N, et al. Hyperthermia: is it an ominous sign after cardiac

arrest? Resuscitation 2001;49:273–7.

[45] Group THACA. Mild therapeutic hypothermia to improve the neurologic outcome after

cardiac arrest. N Engl J Med 2002;346:549–56.


cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63.

[47] Adrie C, MonchiM, Laurent I, et al. Coagulopathy after successful cardiopulmonary resus-

citation following cardiac arrest. J Am Coll Cardiol 2005;46:21–8.

[48] Grando S, Nanzaki S, Morimoto Y, et al. Tissue factor and tissue factor pathway in-

hibitor levels during and after cardiopulmonary resuscitation. Thromb Res 1999;96:

107–13.

[49] Baigent C, Collins R, Appleby P, et al. ISIS-2: 10 year survival among patients with sus-

pected acute myocardial infarction in randomised comparison of intravenous streptokinase,


oral aspirin, both, or neither. The ISIS-2 (Second International Study of Infarct Survival)

Collaborative Group. BMJ 1998;316:1337–43.

[50] Blankfield RP. Benefit of heparin plus aspirin vs aspirin alone in unstable angina. JAMA

1996;276:1873–4.

[51] LongstrethWT.Neurologic complications of cardiac arrest. In: AminoffMJ, editor.Neurol-

ogy and general medicine. Philadelphia: Churchill Livingstone; 2001. p. 51–70.

[52] Celesia GG, Grigg MM, Ross E. Generalized status myoclonus in acute anoxic and toxic-

metabolic encephalopathies. Arch Neurol 1988;45:781–4.

[53] Krumholz A, Stern BJ, Weiss HD. Outcome from coma after CPR: relation to seizures and

myoclonus. Neurology 1988;38:401–5.

[54] Arnoldus EPJ, LammersGJ. Postanoxic coma: good recovery despitemyoclonic status. Ann

Neurol 1995;38:697–8.

[55] Wijdicks EFM, Parisi JE, Sharbrough FW. Prognostic value of myoclonus status in coma-

tose survivors of cardiac arrest. Ann Neurol 1994;35:239–43.

[56] Werhahn KJ, Brown P, Thompson PD, et al. The clinical features and prognosis of chronic

posthypoxic myoclonus. Mov Disord 1997;12:216–20.

[57] Krumholz A, Berg A. Further evidence that for status epilepticus ‘‘one size fits all’’ doesn’t

fit. Neurology 2002;58:515–6.

[58] Goh WC, Heath PD, Ellis SJ, et al. Neurologic outcome prediction in a cardiorespiratory

arrest survivor. Br J Anaesth 2002;88:719–22.

[59] Morimoto Y, Kemmotsu O, Kitami K, et al. Acute brain swelling after out-of-hospital

cardiac arrest: pathogenesis and outcome. Crit Care Med 1993;1:104–10.

[60] TorbeyMT, SelimM, Knorr J, et al. Quantitative analysis of the loss of distinction between

gray and white matter in comatose patients after cardiac arrest. Stroke 2000;31:2163–7.

[61] Sakabe T, Tateishi A, Miyauchi Y, et al. Intracranial pressure following cardiopulmonary

resuscitation. Intensive Care Med 1987;13:256–9.

[62] Lida K, Satoh H, Arita K, et al. Delayed hyperemia causing intracranial hypertension after

cardiopulmonary resuscitation. Crit Care Med 1997;25:971–6.

[63] Jastremski M, Sutton-Tyrrel K, Vaagenes P, et al. Glucocorticoid treatment does not im-

prove neurological recovery following cardiac arrest. Brain Resuscitation Clinical Trial I

study group. JAMA 1989;262:3427–30.

[64] Longstreth WT, Inui TS, Cobb LA, et al. Neurologic recovery after out-of-hospital cardiac

arrest. Ann Intern Med 1983;98:121–32.

[65] Mullner M, Sterz F, Binder M, et al. Blood glucose concentration after cardiopulmonary

resuscitation influences functional neurological recovery in human cardiac arrest survivors.

J Cereb Blood Flow Metab 1997;17:430–6.

[66] Kagansky N, Levy S, Knobler H. The role of hyperglycemia in acute stroke. Arch Neurol

2001;58:1209–12.

[67] van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in the critically ill

patients. N Engl J Med 2001;345:1359–67.

[68] BjorkRJ, Snyder BD,CampionBC, et al.Medical complications of cardiopulmonary arrest.

Arch Intern Med 1982;3:500–3.

[69] Bolli R, Becker L, Gross G, et al. Myocardial protection at a crossroads. Circulation 2004;

95:125–34.

[70] Solomon SD, Zelenkofske S, et al. Sudden death in patients with myocardial infarction and

left ventricular dysfunction, heart failure, or both. N Engl J Med 2005;352:2581–8.

Neurol Clin 24 (2006) 61–71

Therapeutic Hypothermia afterCardiac Arrest

Stephen Bernard, MD, FACEM, FJFICMIntensive Care Unit, Dandenong Hospital, David Street, Dandenong, Victoria 3175, Australia

Out-of-hospital sudden cardiac arrest (SCA) is common, occurring in ap-proximately 1 in 2000 adults per year [1]. This number is likely to increasein the future because of an ageing population. Also, the number of patientswho are initially resuscitated from SCA is expected to increase as a result ofearly defibrillation programs [2,3]. At present, 30% to 40% of SCA patientswho have an initial cardiac rhythm of ventricular fibrillation achieve a suc-cessful return of a spontaneous circulation in the field; however, only 1% to10% of patients survive to hospital discharge [4–6]. There are very few sur-vivors (1%–2%) if the initial cardiac rhythm on arrival of paramedics isasystole [7]. On the other hand, those patients who regain consciousnesshave a good long-term prognosis [8].

Previously, the management of SCA patients after hospital arrival waslargely supportive. Recently, randomized, prospective, controlled trials ofinduced hypothermia (IH) to 33�C for 12 to 24 hours has been shown to im-prove outcome significantly in SCA patients when the initial cardiac rhythmis ventricular fibrillation [9,10].

This article reviews the history of the use of IH after SCA, the physiologiceffects of IH, and current cooling techniques. A protocol is described for themanagement of the post-SCA patient that includes the implementation of IH.

The history of hypothermia after cardiac arrest

In 1950, Bigelow introduced hypothermia as ameans of cerebral protectionduring cardiac surgery [11]. Subsequently, cooling to mild hypothermic levelswas used during the 1950s for a range of neurologic indications, includinghead injury, stroke, and anoxic injury [12]. The first reported use of IH forneurologic injury after cardiac arrest was in 1958 in four patients [13].





62 BERNARD

The following year, the same investigators treated 12 patients who remainedcomatose after resuscitation from in-hospital cardiac arrest with IH main-tained for up to 8 days [14]. In this latter study, there were six survivors(50%), compared with one survivor in seven historical controls. For reasonsthat are unclear, there were few subsequent reports of the use of IH aftercardiac arrest until this therapy was studied in laboratory animals in the early1990s.

Interest in the use of IH after cardiac arrest was revived in 1991, at thetime Sterz and colleagues [15] demonstrated in a dog model that IH inducedafter 15 minutes of cardiac arrest was associated with significant improve-ments in neurologic outcome. Similar outcomes in other laboratory studies[16,17] renewed interest in the application of this therapy in humans.

Several centers then started feasibility and safety studies of the use of IHafter resuscitation from SCA. In 1997, the author’s experience in 22 patientspost arrest suggested that IH was well tolerated; and side effects were able tobe managed adequately in a modern intensive care unit [18]. Other prelim-inary clinical studies confirmed these findings [19–21].

Between 1997 and 2000, prospective, randomized, controlled clinicaltrials were conducted in Australia [9] and Europe [10]. In an Australianstudy, 77 patients were enrolled [9]. There was good outcome (defined as dis-charge from hospital either to home or to a rehabilitation facility) in 49% ofthe patients treated with hypothermia (33�C for 12 hours) compared with26% of patients who were maintained at normal temperature. The adjustedodds ratio for good outcome with IH was 5.25 (95% confidence intervals1.47–18.76); and there was no significant increase in the incidence of sideeffects.

In the European study, there were 273 patients enrolled, with 136 patientsundergoing IH (33�C for 24 hours) and 137 patients maintained at normo-thermia [10]. At 6 months, 55% of the IH patients had good outcome, com-pared with 39% of normothermic controls (risk ratio, 1.4; 95% confidenceinterval, 1.08–1.81). The complication rate did not differ between the twogroups.

On the basis of these studies, in 2003 the International Liaison Commit-tee on Resuscitation endorsed the use of hypothermia for patients who haveneurologic injury after ventricular-fibrillation cardiac arrest [22]. Since thattime, clinical studies have largely focused on determining the optimal timingand technique for the induction of hypothermia after cardiac arrest, as wellas on the applicability to nonventricular-fibrillation arrest patients.

The physiologic effects of hypothermia

The use of IH after cardiac arrest requires an understanding of the phys-iologic effects of IH on each organ system. The following provides an over-view that is relevant for the physician during the first 24 hours of patientmanagement after resuscitation from SCA.

63THERAPEUTIC HYPOTHERMIA AFTER CARDIAC ARREST

Neurologic system

Although the precise effects of hypothermia on the injured brain are un-certain, there is considerable laboratory evidence that continuing neurologicinjury occurs in the early post-arrest period (the ‘‘reperfusion injury’’) [23].The injury is caused by several biochemical cascades that are believed to bevery temperature sensitive [24], providing the main scientific rationale fortreatment with hypothermia for hours to days after resuscitation.

Other proposed mechanisms by which hypothermia may be effective areuncertain, but may include improvements in cerebral oxygen delivery. Forexample, cerebral edema may occur in the postarrest period, particularlywhen the cause of the arrest was asphyxia [25]. Hypothermia is known todecrease intracranial pressure [26] and this proposed mechanism may beanother way in which hypothermia proves helpful in such patients.

Respiratory system

Pulmonary complications after SCA resuscitation include aspirationpneumonitis and chest wall injuries from prolonged external cardiac mas-sage. Although nosocomial pneumonia may be seen when IH is used for pe-riods longer than 48 hours [27], there seems to be minimal risk forventilator-acquired pneumonia when IH is used for periods up to 24 hours.

Cardiovascular system

The cause of SCA may not be apparent on the basis of history, physicalexamination, and ECG; and 90% of post-SCA patients do not have ECGcriteria suggesting acute coronary syndrome [9]. On the other hand, earlycoronary angiography in patients who have had SCA may reveal un-expected lesions that are amenable to therapy. In one study, it was foundthat 48% of patients had an acute coronary artery lesion that was consideredresponsible for the ventricular arrhythmia [28]. There is also evidence thataggressive interventional cardiac care after SCA resuscitation is associatedwith improved outcomes [29]. Therefore, in the comatose post-arrest patientwho may or may not have ECG criteria for an acute coronary syndrome,consideration should be given to urgent interventional revascularization.

Myocardial dysfunction is common after resuscitation from SCA [30],and inotropic drug therapy may be required in many patients for the main-tenance of adequate cerebral perfusion pressure. The hemodynamic effectsof IH include a decrease in heart rate and an increase in systemic vascularresistance, with stroke volume and mean arterial blood pressure maintained[9]. There does not seem to be an increase in the number of patients whohave cardiac arrhythmias during IH therapy [9].

There is some evidence that IH may be beneficial to the post-arrest heartand the brain during the reperfusion period. This benefit was demonstratedin studies of IH (using an intravascular cooling device) during interventional

64 BERNARD

cardiology [31]. Also, IH has been used for the treatment of cardiogenicshock that was not responsive to usual therapy [32].

Acid base

There is debate on whether blood gases should be corrected for temper-ature; however, it is only of practical importance when moderate hypother-mia (!32�C) is used, for example, in deep hypothermia for cardiac orneurologic surgery. Hypothermia after SCA does not significantly increasemetabolic acidosis or lactate levels [9].

Renal/electrolytes

An increase in the serum creatinine level is usually seen during the first 24hours after SCA; however, there is rarely a need for renal replacement ther-apy [9]. Of more importance, the potassium level decreases during inductionof hypothermia [9], and therefore its level needs to be monitored every 1 to 2hours during the initial period of care. Magnesium and phosphate levels alsodecrease during IH and may need supplementation [33].

Hematologic

Decreased temperature has a small effect on clotting times and plateletcounts during prolonged (O48 hours) hypothermia [27]. In any case,many patients who are post SCA will receive antiplatelet and/or anticoagu-lation therapy as part of therapy for suspected acute coronary syndrome.The clinical trials of IH after cardiac arrest demonstrated that there wasno added risk for bleeding in hypothermic patients compared with normo-thermic controls [9,10]

Gastrointestinal

Hypothermia increases the blood glucose as a result of decreased insulinrelease from the pancreas [34]. It is also known that hyperglycemia is asso-ciated with worse outcome after (focal) anoxic injury [35]. Because tight con-trol of blood glucose (between 4 and 6mmol/L) has been shown to improveoutcome in critically-ill surgical patients [36], it is reasonable (although ofunproven benefit) to monitor strictly and treat hyperglycemia after anoxicbrain injury using an intravenous (IV) infusion of insulin. However, the ex-act blood glucose target is uncertain.

Cooling techniques

Surface cooling

Laboratory studies suggest that outcomes are further improved when IHis commenced as early as possible after cardiac arrest [15,37]. Therefore,


vigorous patient cooling should be undertaken immediately after hospitalarrival to decrease core temperature toward 33�C. In practice, however,the rapid induction of hypothermia using surface cooling in adults is prob-lematic. In Australian hypothermia trials [9,18], a neuromuscular blockingdrug and extensive surface cooling with icepacks was used; but it providedfor relatively slow core cooling (approximately 0.9�C/h) and was consideredvery inconvenient by attending medical and nursing staff.

The European trial of hypothermia after cardiac arrest used a refrigeratedair blanket; it was also a very slow technique, with a decrease in core tem-perature of only 0.3�C/h [10]. Many of the patients in this study also re-quired icepacks after 4 hours if cold-air cooling had not been effective.

More effective surface cooling seems possible using cooling blankets,which are specially manufactured to provide adhesion to the skin (ArcticSun, Medivance, Inc., Colorado, USA). Preliminary data indicate thatthis device provides accurate temperature control in patients who are febrile[38], and further studies using this technology for IH after cardiac arrest arecurrently being conducted.

Surface cooling using helmet devices does not appear to provide particu-lar significant protection to the brain directly [39] but does decrease coretemperature slowly [40].

Core cooling

A relatively simple, inexpensive technique for the induction of mild hypo-thermia using core cooling is the use of large volume, ice cold intravenousfluid (LVICF). In a pilot study of 22 patients, a large-volume (30 mL/kg)ice-cold (4�C) crystalloid fluid (lactated Ringer’s solution) was rapidlyinfused intravenously, together with a large dose of a long-acting neuro-muscular blocker (vecuronium bromide) to prevent shivering [41]. This ther-apy decreased core temperature by 1.6�C and increased blood pressure,without any patient developing pulmonary edema. The fluid was infused intoa peripheral intravenous cannula using a pressure bag to ensure a highflow rate through a standard IV-giving set.

The use of LVICF has also been studied by Kim and colleagues [42]. Thatstudy determined the effect of infusing 2000 mL of ice-cold saline rapidlyinto a peripheral vein on temperature and hemodynamics in 17 hospitalizedsurvivors of out-of-hospital cardiac arrest. Cardiac function was assessed bytransthoracic echocardiography before and after fluid administration. Infu-sion of cold saline resulted in a mean temperature drop of 1.4�C at 30 min-utes after the initiation of infusion. There was no increase in central venouspressure, pulmonary pressures, or left atrial filling pressures as assessed byechocardiography.

Contraindications to LVICF therapy would include the presence of pul-monary edema or patients who have chronic renal failure (on dialysis) whomay be unable to excrete a large fluid load.

66 BERNARD

Alternative core cooling techniques for the rapid induction of IH are cur-rently under investigation and include intravascular cooling devices [43].These devices consist of a large catheter inserted into the femoral veinand recirculation of cold saline. A feedback circuit with core temperature in-put allows automatic temperature control. These devices provide for veryrapid temperature decreases [31,43]. Other extracorporeal circulation strat-egies have been described [44,45], but these techniques would generally bebeyond the capabilities of most emergency departments. In addition, theyentail expensive machines, require specialized physician training for catheterinsertion, and take considerable time to establish.

During the cooling phase, patients require sedation to prevent shivering.Because all patients will be ventilated mechanically, muscle relaxants mayalso be given as required. In a study of awake patients undergoing IH forstroke, an infusion of magnesium facilitated the cooling procedure [46];however, the relevance of this intervention in the patient who is being ven-tilated is uncertain.

Core temperature measurement

In patients who have neurologic injury for which the use of IH isplanned, it is essential that core temperature be monitored continuouslyand accurately. There is a minimal temperature gradient among brain,esophageal, and bladder temperatures [47]; therefore, it is recommendedthat esophageal or bladder temperature be monitored continuously after ar-rival at hospital. Tympanic temperature monitoring may be inaccuratewhen the head is surrounded by icepacks and cold water has entered theexternal auditory canal.

A protocol for cooling after cardiac arrest

The following protocol has developed from the experience of clinicaltrials [9,18,27,41] and comprises an initial evaluation in the emergencydepartment, immediate interventions focusing on cerebral resuscitation,history taking, secondary survey, and definitive cardiac care. The protocolincludes early, rapid induction of hypothermia to 33�C immediately afterarrival at the hospital.

Initial patient evaluation

Airway/breathingAll patients who have been resuscitated from prolonged SCA will remain

comatose and will require endotracheal intubation for airway protection,oxygenation, and ventilation control. Mechanical ventilation with 100%oxygen at a tidal volume of 10 mL/kg and a rate of 8 to 10 breaths per min-ute are initial ventilator settings that should ensure normocapnea. The


production of carbon dioxide is decreased by 30% when core temperature is33�C, therefore the ventilator rate may need to be decreased to only 6 to 8breaths per minute, guided by end-carbon dioxide readings and arterialblood gas analysis.

To facilitate mechanical ventilation and assist in the rapid induction ofhypothermia, a large dose of a nondepolarising long-acting muscle relaxant(ie, vecuronium bromide) should be administered immediately after initialneurologic assessment. Concurrently, a chest radiograph should be per-formed during initial evaluation to exclude right main bronchus intubationand to diagnose aspiration pneumonitis or pulmonary edema.

CirculationThere is evidence from laboratory studies thatmild hypertension after SCA

improves neurologic outcome [48,49]. One strategy that both induces mildhypothermia and improves blood pressure is the rapid infusion of LVICFas outlined in the discussion above. If hypotension (mean arterial pressure[MAP] !70 mmHg) persists despite this fluid therapy, then an inotropicdrug should be infused. In the author’s studies [9,18,41], epinephrine wasused without apparent adverse cardiac effect. If the patient is initially hyper-tensive, an infusion of propofol G glyceryl trinitrate should be considered.

Initial procedures

After the initial ‘‘ABC’’ resuscitation measures described in the discus-sion above, the following procedures and investigations need to be under-taken. A nasogastric tube should be inserted, because bystander-expiredair ventilation commonly results in air inflation of the stomach. The inser-tion of an arterial line facilitates continuous blood pressure monitoringand the drawing of blood for routine laboratory tests. A 12-lead electrocar-diogram is required to diagnose acute coronary syndromes.

Central venous access may be required for right atrial pressure monitor-ing or inotropic drug infusion. A femoral venous line may be a safer optionthan subclavian or internal jugular puncture, especially if thrombolysis isplanned [50].

Patient history

Information relating to the patient who is in cardiac arrest should besought from paramedics and family members concurrently with the initialresuscitation measures listed in the discussion above. In particular, thetime between collapse and return of spontaneous circulation, initial cardiacrhythm, whether the arrest was witnessed, and whether bystander cardiopul-monary resuscitation was performed are factors that relate to eventual prog-nosis. Other important historical information includes past medical history,current medication, and known allergies.

68 BERNARD

Secondary survey

Following the initial evaluation and treatment, there should be a thor-ough secondary physical examination. In particular, the presence of head in-jury should be excluded, because SCA patients may fall and strike theirheads. In such cases, the neck must be immobilized and computed tomogra-phy of the brain and cervical spine undertaken before consideration ofthrombolytic therapy.

Examination of the chest should focus on the presence of chest trauma asa result of external cardiac massage. This injury may cause bleeding afterthrombolysis; however, even if it is present, the incidence of major hemor-rhage is low [51].

Maintenance of hypothermia

During the induction of IH, the SCA patient will shiver vigorously be-tween 34�C and 35�C, which may be associated with increased oxygen de-mand and possible myocardial ischemia [52]. The maintenance of thistemperature is therefore difficult without pharmacologic paralysis or deepsedation. However, once a core temperature of less than 34�C is reached,adult patients tend to become poikliothermic and not shiver. At tempera-tures about 33�C, the judicious use of sedation rather than pharmacologicparalysis may maintain core temperature. A benzodiazepine, such as mida-zolam, given by continuous infusion is commonly used; but propofol infu-sions should be used very cautiously because of risk for hypotension.

If a cooling device is not available, icepacks will need to be applied to thehead, neck, and torso of the patient if the core temperature starts to increase(O33.5�C). If core temperature decreases (!32.5�C), icepacks should be re-moved and paralyzing drugs withheld while a heated-air blanket is applied.

The optimal duration of IH after SCA resuscitation is uncertain. TheAustralian study used 12 hours [9], while the European study used 24 hours[10]. More recently, an animal study of asphyxial cardiac arrest demon-strated clearly that a 24-hour period was associated with improved neuro-logic recovery compared with 6 hours of IH [53]. Therefore, for mostpatients who have significant neurologic injury, the use of IH for 24 hours,followed by 12 hours of controlled rewarming seems reasonable.

Rewarming from hypothermia

Active rewarming of the hypothermic patient requires the use of a heated-air blanket to increase core temperature. During this time, any shiveringmust be suppressed with sedation. In addition, rewarming may result in pe-ripheral vasodilatation, and (warm) IV fluid therapy may need to be given tomaintain MAP.

There is some evidence that rapid rewarming may be harmful [54]; there-fore, careful monitoring and control of temperature at this time is required.


Summary

The use of IH for 24 hours in patients who remain comatose followingresuscitation from out-of-hospital cardiac arrest improves outcomes. How-ever, the induction of hypothermia has several physiologic effects that needto be considered. A protocol for the rapid induction of hypothermia is de-scribed. At present, the rapid infusion of a large volume (40 mL/kg) of ice-cold crystalloid (ie, lactated Ringer’s solution) would appear to be aninexpensive, safe strategy for the induction of hypothermia after cardiacarrest. Hypothermia (33�C) should be maintained for 24 hours, followedby rewarming over 12 hours. Particular attention must be paid to potassiumand glucose levels during hypothermia.

References

[1] Zheng ZJ, Croft JB, Giles WH, et al. Sudden death in the United States, 1989–1998. Circu-

lation 2001;104:2158–63.

[2] Caffrey SL, Willoughby PJ, Pepe PE, et al. Public use of automated external defibrillators.

N Engl J Med 2002;347:1242–7.

[3] Valenzuela TD, RoeDJ, Nichol G, et al. Outcomes of rapid defibrillation by security officers

after cardiac arrest in casinos. N Engl J Med 2000;343:1206–9.

[4] Bernard SA. Outcome from prehospital cardiac arrest inMelbourne, Australia. EmergMed

1998;10:25–9.

[5] Fredriksson M, Herlitz J, Nichol G. Variation in outcome in studies of out-of-hospital car-

diac arrest: a review of studies conforming to the Utstein guidelines. Am J EmergMed 2003;

21:276–81.

[6] Eckstein M, Stratton SJ, Chan LS. Cardiac arrest resuscitation evaluation in Los Angeles:

CARE-LA. Ann Emerg Med 2005;45:504–9.

[7] Meyers A, Bernard SA, Smith KL, et al. Asystolic cardiac arrest in Melbourne, Australia.

Emerg Med 2001;13:186–9.

[8] Bunch TJ, White RD, Gersh BJ, et al. Long term outcomes of out-of-hospital cardiac arrest

after successful early defibrillation. N Engl J Med 2003;348:2626–33.

[9] Bernard SA, Gray T, Buist MD, et al. Treatment of comatose survivors of out-of-hospital


[10] The Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to

improve the neurological outcome after cardiac arrest. N Engl J Med 2002;346:549–56.

[11] Bigelow WG, Lindsay WK, Greenwood WF. Hypothermia: its possible role in cardiac sur-

gery. Ann Surg 1950;132:849–66.

[12] Vandam DV, Burnap TK. Hypothermia. N Engl J Med 1959;261:546–53.

[13] Benson DW, Williams GR, Spencer FC, et al. The use of hypothermia after cardiac arrest.

Anesth Analg 1958;38:423–8.

[14] WilliamsGR Jr, Spencer FC. Clinical use of hypothermia following cardiac arrest. Ann Surg

1959;148:462–8.

[15] Sterz F, Safar P, Tisherman SA, et al. Mild hypothermic cardiopulmonary resuscitation

improves outcome after cardiac arrest in dogs. Crit Care Med 1991;19:379–89.

[16] GinsbergMD, Sternau LL, GlobusMY. Therapeutic modulation of brain temperature: rel-

evance to ischaemic brain injury. Cerebrovasc Brain Metab Rev 1992;4:189–225.

[17] Weinrauch V, Safar P, Tisherman SA, et al. Cerebral benefit of mild or moderate hypother-

mia and cardiac risk of deep resuscitative hypothermia after cardiac arrest in dogs. Stroke

1992;23:1454–62.

70 BERNARD

[18] Bernard SA, Jones BM, Horne MK. A clinical trial of induced hypothermia in comatose

survivors of prehospital cardiac arrest. Ann Emerg Med 1997;30:146–53.

[19] Yanagawa Y, Ishihara S, Norio H, et al. Preliminary clinical outcome study of mild resus-

citative hypothermia after out-of-hospital cardiac arrest. Resuscitation 1998;39:61–6.

[20] Nagao K, Hayashi N, Kanmatsuse K, et al. Cardiopulmonary cerebral resuscitation using

emergency cardiopulmonary bypass, coronary reperfusion therapy, and mild hypothermia

in patients with cardiac arrest outside the hospital. J Am Coll Cardiol 2000;36:776–83.

[21] Zeiner A, Holzer M, Sterz F, et al. Mild resuscitative hypothermia to improve neurological

outcome after cardiac arrest: a clinical feasibility trial. Stroke 2000;31:86–94.

[22] Nolan JP, Morley PT, Hoek TL, et al. Therapeutic hypothermia after cardiac arrest: an ad-

visory statement by the Advancement Life Support Task Force of the International Liaison

Committee on Resuscitation. Resuscitation 2003;57:231–5.

[23] Safar P, Behringer W, Bottinger BW, et al. Cerebral resuscitation potentials for cardiac

arrest. Crit Care Med 2002;30:S140–4.

[24] Takata K, Takeda Y, Sato T, et al. Effects of hypothermia for a short period on histologic

outcome and glutamate concentration during and after cardiac arrest in rats. Crit CareMed

2005;33:1340–5.

[25] Morimoto Y, Kemmotsu O, Kitami K, et al. Acute brain swelling after out-of-hospital car-

diac arrest: pathogenesis and outcome. Crit Care Med 1993;21:104–10.

[26] CliftonGL,Miller ER, SungCC, et al. Lack of effect of induction of hypothermia after acute

brain injury. N Engl J Med 2001;344:556–63.

[27] Bernard SA, Jones BM, Buist M. Experience with prolonged induced hypothermia in pa-

tients with severe head injury. Crit Care 1999;3:167–72.

[28] Spaulding CM, Joly LM, Rosenberg A, et al. Immediate coronary angiography in survivors

of out-of-hospital cardiac arrest. N Engl J Med 1997;336:1629–33.

[29] Engdahl J, Abrahamsson P, Bang A, et al. Is hospital care of major importance for outcome

after out-of-hospital cardiac arrest? Experience acquired from patients with out-of-hospital

cardiac arrest resuscitated by the same emergencymedical service and admitted to one of two

hospitals over a 16-year period in the municipality of Goteborg. Resuscitation 2000;43:

201–11.

[30] Laurent I, Monchi M, Chiche JD, et al. Reversible myocardial dysfunction in survivors of

out-of-hospital cardiac arrest. J Am Coll Cardiol 2002;40:2110–6.

[31] Dixon SR,Whitbourn RJ, DaeMW, et al. Induction of mild systemic hypothermia with en-

dovascular cooling during primary percutaneous coronary intervention for acute myocar-

dial infarction. J Am Coll Cardiol 2002;40:1928–34.

[32] Dalrymple-HayMJ,DeakinCD,KnightH, et al. Induced hypothermia as salvage treatment

for refractory cardiac failure following paediatric cardiac surgery. Eur J Cardiothorac Surg

1999;15:515–8.

[33] PoldermanKH, Peerdeman SM,Girbes ARJ. Hypophosphatemia and hypomagnesemia in-

duced by cooling of patients with severe head injury. J Neurosurg 2001;94:697–705.

[34] Curry DL, Curry KP. Hypothermia and insulin secretion. Endocrinology 1970;87:750–3.

[35] Baird TA, Parsons MW, Phanh T, et al. Persistent poststroke hyperglycemia is indepen-

dently associated with infarct expansion and worse clinical outcome. Stroke 2003;34:

2208–14.

[36] Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill pa-

tients. N Engl J Med 2001;345:1359–67.

[37] Kuboyama K, Safar P, Radovsky A, et al. Delay in cooling negates the beneficial effect of

mild resuscitative cerebral hypothermia after cardiac arrest in dogs: a prospective, random-

ized study. Crit Care Med 1993;21:1348–58.

[38] Mayer S. Surface cooling for fever control in brain-injured patients. Crit CareMed 2005;33:

1891–5.

[39] Hachimi-Idrissi S, Corne L, Ebinger G, et al. Mild hypothermia induced by a helmet device:

a clinical feasibility study. Resuscitation 2001;51(3):275–81.


[40] Wang H, Olivero W, Lanzino G, et al. Rapid and selective cerebral hypothermia achieved

using a cooling helmet. J Neurosurg 2004;100:272–7.

[41] Bernard SA, BuistMD,MonterioO, et al. Induced hypothermia using large volume, ice-cold

intravenous fluid in comatose survivors of out-of-hospital cardiac arrest: a preliminary re-

port. Resuscitation 2003;56:9–13.

[42] Kim F, Olsufka M, Carlbom D, et al. Pilot study of rapid infusion of 2 L of 4 degrees C

normal saline for induction of mild hypothermia in hospitalized, comatose survivors of

out-of-hospital cardiac arrest. Circulation 2005;112:715–9.

[43] GeorgiadisD, Schwarz S,KollmarR, et al. Endovascular cooling formoderate hypothermia

in patients with acute stroke: first results of a novel approach. Stroke 2001;32:2550–3.

[44] Piepgras A, Roth H, Schurer L, et al. Rapid active internal core cooling for induction of

moderate hypothermia in head injury by use of an extracorporeal heat exchanger. Neurosur-

gery 1998;42:317–8.

[45] HolzerM, BehringerW, JanataA, et al. Extracorporeal venovenous cooling for induction of

mild hypothermia in human-sized swine. Crit Care Med 2005;33:1346–50.

[46] Zweifler RM, Voorhees ME, Mahmood MA, et al. Magnesium sulfate increases the rate of

hypothermia via surface cooling and improves comfort. Stroke 2004;35:2331–4.

[47] Henker R, Brown D, Marion M. Comparison of brain temperature with bladder and rectal

temperatures in adults with severe head injury. Neurosurgery 1998;42:1071–5.

[48] Safar P, Xiao F, Radovsky A, et al. Improved cerebral resuscitation from cardiac arrest in

dogs with mild hypothermia plus blood flow promotion. Stroke 1996;27:105–13.

[49] Safar P, Kochanek P. Cerebral blood flow promotion after prolonged cardiac arrest. Crit

Care Med 2000;28:3104–6.

[50] Desmond J,MegahedM. Is the central venous pressure reading equally reliable if the central

line is inserted via the femoral vein. Emerg Med J 2003;20:467–9.

[51] Scholz KH, Tebbe U, Herrmann C, et al. Frequency of complications of cardiopulmonary

resuscitation after thrombolysis during acute myocardial infarction. Am J Cardiol 1992;69:

724–8.

[52] Frank SM, Fleisher LA, BreslowMJ, et al. Perioperative maintenance of normothermia re-

duces the incidence of morbid cardiac events: a randomised clinical trial. JAMA 1997;277:

1127–34.

[53] Katz LM,YoungA, Frank JE, et al.Neurotensin-induced hypothermia improves neurologic

outcome after hypoxic-ischemia. Crit Care Med 2004;32:806–10.

[54] Ueda Y, Suehiro E, Wei EP, et al. Uncomplicated rapid posthypothermic rewarming alters

cerebrovascular responsiveness. Stroke 2004;35:601–6.

Neurol Clin 24 (2006) 73–87

Cerebral Resuscitation: State of theArt, Experimental Approaches

and Clinical Perspectives

Erik Popp, MD, Bernd W. Bottiger, MD, DEAA*Department of Anesthesiology, University of Heidelberg, Im Neuenheimer Feld 110,

D-69120, Heidelberg, Germany

Every day, up to 1000 persons in the United States and another 1000 inEurope go into cardiocirculatory arrest with subsequent cardiopulmonaryresuscitation (CPR), and only about 2% to12% of them survive. Postarrestbrain damage is a key issue. In the western industrialized nations, CPR isattempted in 40 to 90 of 100,000 inhabitants annually, and restoration ofspontaneous circulation (ROSC) can be achieved in about 25% to 50% ofthese patients. The hospital discharge rate, however, is only 2% to 12%(Fig.1) [1]. Therefore, out of up to 300,000 cardiac arrest victims annuallyin the United States and in Europe, more than 270,000 are not treated suc-cessfully, if one uses complete neurologic restoration as the standard.

The major reason for postarrest in-hospital mortality and morbidity ispersistent brain damage. Brain damage following cardiocirculatory arrestis related to the short period of tolerance to hypoxic stress and specific re-perfusion disorders [2,3]. The individual, social, and economic consequencesof brain damage following cardiac arrest are immense [4–6]. One of the mostimportant issues in cardiac arrest and resuscitation (‘‘whole body ischemiaand reperfusion‘‘) research, therefore, is cerebral resuscitation and the inhi-bition of postarrest cerebral damage [3,7]. Current research focuses on path-ophysiology and problems during reperfusion [7,8].

The mechanisms of brain damage following global cerebral ischemia andcardiac arrest are complex [2,3,7]. Major issues are hypoxia and subsequentnecrosis, reperfusion injury with free radical formation and cellular calciuminflux, release of excitatory amino acids, neuronal apoptosis, and cerebral


E-mail address: [email protected] (B.W. Bottiger).




74 POPP & BOTTIGER

microcirculatory reperfusion disorders [2,3,7,9]. Several clinical trials at-tempted to improve neurologic outcome after cardiac arrest by focusingon brain protection with the therapeutic use of barbiturates and by focusingon reperfusion injury with the use of calcium channel blockers. No positiveeffects on neurological outcome could be established, however [10,11]. Todate, no specific pharmacologic postarrest treatment options are availableto improve neurologic outcome following cardiocirculatory arrest in theclinical setting, with cardiocirculatory arrest being the most relevant clinicalfeature of global cerebral ischemia.

The most important therapeutic options to improve neurologic outcomefollowing cardiac arrest currently under study are focusing on (1) selectiveneuronal vulnerability and delayed neuronal death [8,12,13], (2) on cerebralmicrocirculatory reperfusion [14–17] and (3) on therapeutic hypothermia,the topic with the most clinical evidence [18–20].

From experimental concepts to clinical strategies

Selective neuronal vulnerability and delayed neuronal death

Selective neuronal vulnerability and delayed neuronal death contribute toneuronal damage following global cerebral ischemia resulting from cardiacarrest in the clinical and the experimental settings [8]. Short periods of globalcerebral ischemia resulting from cardiocirculatory arrest induce neuronalcell damage primarily in so-called ‘‘selectively vulnerable’’ brain areas,such as the hippocampal CA1 sector, the thalamic reticular nucleus, anddifferent layers of the neocortex [8,20]. In a recently developed model ofcardiac arrest in rats, strong evidence for neuronal apoptosis in selectivelyvulnerable areas of the brain was found. Using the TUNEL-technique

Fig.1. Survival after out-of-hospital cardiac arrest of cardiac etiology and subsequent cardio-

pulmonary resuscitation (CPR) in the area of Heidelberg, Germany (36-month period). The

great discrepancy between patients who show restoration of spontaneous circulation (ROSC)

and the hospital discharge rate is – in major parts – because of postarrest brain damage [1].

(Adapted from Bottiger BW, Grabner C, Bauer H, et al. Long-term outcome after out-of-hos-

pital cardiac arrest with physician staffed emergency medical services: the Utstein style applied

to a midsized urban/suburban area. Heart 1999;82:674–9; with permission.)

75CEREBRAL RESUSCITATION

(terminal deoxynucleotidyltransferase [TdT]-mediated d-uracil triphos-phate [UTP]-biotin nick end-labeling), a characteristic apoptotic morphol-ogy with DNA-laddering and apoptotic bodies was detected in thehippocampal CA1 sector and the thalamic reticular nucleus [8,9,13]. Posi-tive TUNEL staining is an important indicator of apoptotic degeneration.Quantitative analysis of TUNEL-positive neurons per tissue section re-vealed marked differences with regard to the time course between the hip-pocampal CA1 sector and the thalamic reticular nucleus. TUNEL stainingdemonstrated early onset degeneration in the thalamic reticular nucleus at6 hours; it peaked at 3 days. In contrast, degeneration was delayed in thehippocampal CA1 sector, showing an onset at 3 days and a maximum ofTUNEL-positive cells at 7 days [8,9,13]. These data suggest that apoptosiscontributes to neuronal cell death after cardiac arrest. Moreover, delayedneuronal degeneration reflects a time window in which potential therapeu-tic interventions can be established after cardiac arrest [7].

Cascades of death

The current concepts of apoptosis suggest that there are several steps be-tween the initial ischemic/hypoxic insult and the final DNA fragmentationleading to cell death (Fig. 2) [21–23]. Within this cell death cascadedconsist-ing of various signals, modulatory proteins, and degradation enzymesdseveral molecules and proteins facilitating neuronal survival compete withfactors contributing to cell death. Ultimately, the balance between survivalfactors and death factors determines the fate of the cell. Proteins such as

signals

modulators

proteases

ischemia / hypoxia

survival signals

(e.g. BDNF / IGF)

death signals

(e.g. Fas-FasL)

Survival Death

Bax / Bid / BadBcl-2 / Bcl-XL

caspase inhibitors

(e.g. z-VAD-FMK, p35) Caspase 3, 6, 7

cleavage

of target proteins

„apoptotic bodies“

DNA fragmentationAPOPTOSIS

Fig. 2. A simplified scheme of the pathophysiologic concepts of neuronal apoptosis suggests

that there are several steps between the initial ischemic/hypoxic insult and the final DNA frag-

mentation leading to cell death [9,21,23,24]. Within this cell death cascade, consisting of signals,

modulators, and degradation enzymes, there are several molecules and proteins that facilitate

neuronal survival and compete with factors that contribute to the cell-death cascade. Ultimately,

the balance between survival factors anddeath factors determines the fate of the cell. Proteins such

as Bcl-2 and Bcl-XL promote survival, while Bax, Bid, and Bad promote death. The final step of

this cascade is initiated with the activation of caspase 3.

76 POPP & BOTTIGER

Bcl-2 and Bcl-XL promote survival, whereas Bax and Bad promote death.The final step in this cascade is initiated with the activation of caspase 3. Ac-tivation of caspase 3 leads to a cleavage of poly(ADP-ribose)polymerase(PARP), which is an important DNA-repair enzyme [21,23,24]. Combinedwith the cleavage of other important substrates, activation of caspase 3 isbelieved to be the final trigger of DNA fragmentation and apoptotic celldeath, which can be viewed as a form of endogenous ‘‘cell suicide.’’ This fi-nal caspase activation step can be blocked by synthetic caspase inhibitorsand different viral antiapoptotic proteins, such as the baculovirus proteinp35 and the cowpox virus protein crmA [23]. An important step in confirm-ing that delayed neuronal death after cardiac arrest is based on apoptosiswas the demonstration of upregulated caspase 3-like protease activity in dif-ferent brain regions 24 hours after global cerebral ischemia induced by car-diac arrest [25]. In particular, a significant increase in caspase 3 mRNA andcaspase 3-like proteolytic activity in the hippocampus was observed. Pre-incubation of hippocampal extracts with a specific caspase 3 inhibitorcompletely blocked protease activity. Thus, one of the final steps in anapoptotic cascade, activation of caspase 3, occurs in the brain after cardiacarrest at the transcriptional and the posttranscriptional level [25]. Thisfinding gives further support to the hypothesis that apoptotic degenerationcontributes to neuronal death after cardiocirculatory arrest and suggeststhat antiapoptotic treatment may play an important role in promotingneuronal survival after cardiac arrest in the future.

Inhibiting cerebral apoptosis

In an attempt to inhibit neuronal apoptosis, a transgenic rat line was cre-ated that expressed the antiapoptotic baculovirus protein p35 in neuronspostnatally [26]. Those animals and their nontransgenic littermates under-went a period of 6 minutes of cardiac arrest induced by ventricular fibrilla-tion of the heart. There was a marked difference with regard to the rate ofROSC; and the rate of 7-day survival was significantly higher in p35 trans-genic animals [26]. Therefore, antiapoptotic strategies may be indicated toimprove outcome following global cerebral ischemia and cardiocirculatoryarrest [7,26]. However, additional studies using the intracerebroventricularapplication of artificial caspase inhibitors, such as the specific synthetic cas-pase inhibitor z-DEVD-FMK, did not show a significant positive effect oncaspase-3 activation, neuronal degeneration, and neurologic outcome fol-lowing 6 minutes of cardiocirculatory arrest in rats [27].

Recent data from focal cerebral ischemia demonstrate the possibility ofinhibiting the apoptotic cascade further upstream with the use of neurotro-phins and growth factors. Brain-derived neurotrophic factor (BDNF) is oneof the best characterized neurotrophic factors of the nerve growth factorfamily. BDNF acts on a set of high-affinity receptor kinases (mainly tyro-sine kinase B [TrkB]) to promote survival, differentiation, and neurite


extension in many types of neurons in the mammalian central nervous sys-tem [28–32]. In vivo, BDNF rescues motoneurons and substantia nigradopaminergic cells from traumatic and toxic brain injury [33,34].Intracerebroventricular BDNF administered after focal cerebral ischemiasignificantly reduced infarct volume, primarily in the cortex [35,36]. Anothermechanism of neuroprotection achieved by growth factors after hypoxic/is-chemic events is probably prevention of excitotoxicity [31]. Glutamate-trig-gered excitotoxicity with subsequent Ca2þ overload of cells is thought to beone of the major causes of cellular death after ischemia [37]. BDNF protectsneuronal cells in vitro against glutamate-induced neurotoxicity and the sub-sequently high intracellular calcium levels [31,32]. By inducing an anti-oxidant defense system, BDNF suppresses the glutamate-triggeredaccumulation of peroxide, which contributes to the loss of Ca2þ homeosta-sis [38]. Also, BDNF induces the activation of the IP3 kinase, phospholipaseC, and Ras/MAPKinase (MAPK) pathways via the TrkB receptor, exhibit-ing further neuroprotective cellular stimulation. Those pathways could alsobe activated effectively by the insulin-like growths factor 1 (IGF-1) anderythropoietin (EPO) [39]. Those growth factors/neurotrophins have beenevaluated in models of cardiac arrest with rats. Neither BDNF, nor IGF-1nor EPO, however, could demonstrate beneficial effects on cerebral recoveryand neuronal survival when administered intracerebroventricularily orintraperintoneally after 6 minutes of cardiac arrest [13,40].

Cerebral microcirculatory reperfusion

A major area of experimental resuscitation research has focused on cere-bral microcirculatory reperfusion and associated disorders following cardiacarrest, including endothelial cell swelling, increased leukocyte-endothelialinteractions, and a disseminated intravascular activation of blood coagula-tion [7,14–17].

A most relevant cause of cerebral dysfunction after cardiac arrest is re-flected by the cerebral ‘‘no reflow’’ phenomenon, which describes the regionalmicrocirculatory reperfusion deficits that occur despite adequate systemichemodynamics. Some years ago, Fischer and Hossmann [16] treated catsafter 15 minutes of cardiac arrest and 4 minutes of CPR with hypertonic-hyperoncotic solutions during a 30-minute reperfusion period. Suchsolutions decrease endothelial cell swelling resulting from the high intra-vascular osmotic pressure that is generated by this kind of intervention.In these studies, early cerebral microcirculatory reperfusion disorders (ce-rebral ‘‘no reflow’’) were reduced with the administration of these solu-tions, suggesting that therapeutic interventions that focus on a decreasein endothelial cell swelling have positive effects on cerebral microcirculatoryreperfusion after cardiac arrest [16]. As in several other experimental stud-ies, cerebral perfusion pressure immediately at the start of reperfusion iscorrelated negatively with the extent of cerebral microcirculatory

78 POPP & BOTTIGER

reperfusion disordersdthe higher the early reperfusion pressure, the lowerthe amount of cerebral ‘‘no reflow’’ [14,16,41,42].

Leukocytes in trouble

Another important mechanism of reperfusion injury and reperfusion fail-ure in the microcirculation may be leukocyte adherence and leukocyte stick-ing and, therefore, blocking of microvessels [43,44]. Following 10 minutes ofcardiac arrest and 6 hours of reperfusion, an increase in the number of poly-morphonuclear leukocytes in the brain suggests that leukocytes may playa role in early cerebral microcirculatory reperfusion failure after cardiac ar-rest, as they play a role in reperfusion injury in other organs and differentmodels [8]. This finding has been supported by clinical studies demonstrat-ing that cardiocirculatory arrest and successful CPR are associated witha marked increase in the serum levels of polymorphonuclear neutrophil leu-kocyte (PMN) elastase, complement split products, terminal complementcomplex (sC5b-9), and soluble intercellular adhesion molecules [45–47].

Coagulation without fibrinolysis

The most important pathophysiologic mechanism responsible for cere-bral microcirculatory reperfusion disorders seems to be the activation ofblood coagulation without adequate activation of endogenous fibrinolysis[48–52]. Intravascular fibrin formation and microthromboses are distributedthroughout the entire microcirculation after cardiocirculatory arrest, and in-terventions that focus on hemostasis may be indicated during reperfusion.In the 1950s, Crowell and coworkers [53] demonstrated beneficial effectsof anticoagulatory interventions for the first time in animals. Following10 minutes of cardiac arrest, only a few dogs survived without heparin pre-treatment, while the survival rate was 16% and 67% when doses of heparin,2 mg/kg and 5 mg/kg of body weight respectively, were given before cardiacarrest [53]. In a later study, Crowell demonstrated beneficial effects of pre-treatment with thrombolytic agents before cardiac arrest. In the controlgroup, 14 of 15 animals died after 15 minutes of cardiac arrest [54]. The sur-viving animal suffered from severe neurologic damage. In contrast, only 2 of14 animals died if streptokinase had been administered before cardiac arrest.Almost all neurologic deficits after cardiac arrest in this group disappearedwithin 2 months after stabilization [54]. Lin and coworkers [55] demon-strated that the administration of streptokinase combined with dextranreduces the duration of a flat line electroencephalogram (EEG) and im-proves cerebral blood flow after cardiac arrest in dogs. Safar and coworkers[56] observed an improvement in neurologic outcome in dogs receiving hep-arin, dextran, and hypertensive reperfusion as a combined therapeutic ap-proach following 12 minutes of cardiac arrest.

Based on clinical experience [57,58], the effect of thrombolysis dur-ing CPR on the extent of the cerebral ‘‘no reflow’’ phenomenon was


investigated in cats [16]. Following 15 minutes of cardiac arrest, animalswere allowed to reperfuse spontaneously for 30 minutes. Treated animals re-ceived a bolus injection of recombinant tissue-type plasminogen activator(rt-PA), 1 mg/kg of body weight, combined with heparin, 100 U/kg ofbody weight, during CPR, followed by another 1 mg/kg dose of rt-PA dur-ing reperfusion [16]. The administration of rt-PA and heparin led to a signif-icant reduction (8% versus 29%) in the cerebral ‘‘no reflow’’ phenomenon inthe entire forebrain. This positive effect was particularly relevant in basalganglia and brainstem, and bleeding complications did not occur either[16]. Therefore, there is profound experimental evidence suggesting that he-mostatic disorders may affect overall outcome, and particularly cerebraloutcome, after cardiac arrest.

Clinical investigations

Based on this promising data, a prospective pilot intervention trial in pa-tients undergoing CPR after out-of-hospital cardiac arrest was performed[59]. Overall, 90 patients were included. Heparin and rt-PA were given in40 patients. In the rt-PA group, ROSC was achieved in 68%; and 58% ofpatients were able to be admitted to a cardiac intensive care unit, as com-pared with 44% and 30% of controls. These differences were significant.At 24 hours after cardiac arrest, 35% of patients who were treated withrt-PA (versus 22% of controls) were still alive, and 15% of rt-PA-treated pa-tients (versus 8% of controls) were discharged from the hospital. There wereno bleeding complications related to the CPR procedures. These datawere supported by a retrospective case-control study of 108 patients whowere treated with rt-PA [60]. A randomized and controlled multicenter clin-ical trial on thrombolysis during CPR is currently underway. Overall, morethan 1000 patients will be enrolled in more than 40 centers. The results ofthis large-scale, randomized, controlled clinical trial to improve microcircu-latory reperfusion following cardiac arrest (Thrombolysis in Cardiac Arrest[TROICA] trial) will be available in 2006 [61–63].

Therapeutic mild hypothermia

The use of therapeutic hypothermia following different hypoxic-ische-mic insults has played an important role in various concepts of nonspecificprotection of cells for a long time [64]. Within several cell and animal ex-perimental models, hypothermia was shown to inhibit a wide range of in-tracellular death cascades. A large body of evidence suggests that theprotective effects of cooling are much greater than can be explained bythe reduction in oxygen and glucose metabolism of about 5% to 7%per �C alone [65–72]. Mild therapeutic hypothermia has shown to fosterneuroprotection via inhibition of apoptosis [66], reduction of free radicals[67,68] and excitatory neurotransmitters [69–71], and stabilization of mem-branes [72].

80 POPP & BOTTIGER

Several case reports of accidental hypothermia and good neurologic out-come even after long periods of ischemia [73] led to clinical attempts to usedeep hypothermia (!28�C) in the setting of severe head injury [74] and peri-operatively during cardiac surgery and neurosurgery [75–77]. Although theuse of deep therapeutic hypothermia after cardiac arrest in the last centurydid not lead to an improved outcome [78], recent data show positive effectsof mild therapeutic hypothermia [18,79–81] after cardiac arrest and otherlife-threatening events [82]. The data from the European multicenter trial(‘‘hypothermia after cardiac arrest,’’ HACA trial [19]), as well as thosefrom Australia [18], clearly demonstrate a decrease in mortality (Fig. 3)and a better neurologic outcome for patients who are cooled to 32�C to34�C for 12 or 24 hours. The European trial included 275 patients with wit-nessed cardiac arrest and ventricular fibrillation and who were comatose athospital admission. One hundred thirty-eight of them were cooled to a blad-der temperature of 32�C to 34�C for 24 hours (Fig. 4), whereas 137 remainednormothermic. The hypothermia group showed a good neurologic outcome(able to live independently and work at least part-time; 55% versus 39%)significantly more often with a number needed to treat (NNT) of 6 (relativerisk 1.40; 95% confidence interval 1.08 to 1.81) and higher 6 months survivalrate (59% versus. 49%; relative risk 0.74; 95% confidence interval 0.58-0.95;NNT ¼ 7). Even when the duration of hypothermia is shortened to 12 hours(target temperature 33�C), as done in the Australian study [18], survivalrate with good neurologic outcome is improved (hypothermia 49% versusnormothermia 26%; n ¼ 43 versus 34). In 2003, such findings led to the im-plementation of mild therapeutic hypothermia (32�C–34�C) in the Interna-tional Liaison Committee on Resuscitation (ILCOR) recommendations and

Fig. 3. Kaplan-Meier blot of the cumulative survival in the normothermia and hypothermia

groups of the European trial on hypothermia after cardiac arrest [19]. (From The Hypothermia

After Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic

outcome after cardiac arrest. N Engl J Med 2002;346:549–56; with permission.)


guidelines for the treatment of unconscious patients after prehospital cardi-ac arrest [20]:

On the basis of the published evidence to date, the ILCOR ALS Task Forcehas made the following recommendations:� Unconscious adult patients with spontaneous circulation after out-of-hospital cardiac arrest should be cooled to 32-34.8�C for 12-/24 h

when the initial rhythm was VF.� Such cooling may also be beneficial for other rhythms or in-hospitalcardiac arrest.

Therefore, therapeutic hypothermia should be implemented in clinicalpractice. It is now the first clinically relevant therapeutic approach to im-prove cerebral and overall outcome after cardiac arrest.

Glucose control

Hyperglycemia is a common problem in the postarrest period [83,84]. Ex-perimental and clinical data demonstrate that postischemic blood glucoseconcentration plays an important role in modulating ischemic cerebral in-farction and selective neuronal death [85]. Data from several experimentalanimal studies suggest the use of insulin infusion to overcome those devas-tating effects of high levels of glucose after ischemic brain damage [86,87].Recently a large study of critically ill patients admitted to surgical ICUsdemonstrated a significant mortality benefit when glucose levels were setto a range between 80 and 110 mg/dL using insulin infusion [88]. Up tonow there is no prospective, controlled human trial after cardiac arrest to

Fig. 4. Bladder temperature in the normothermia and hypothermia groups of the European trial

on hypothermia after cardiac arrest [19]. The T bars indicate the 75th percentile in the

normothermia group and the 25th percentile in the hypothermia group. (From The Hypo-

thermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the

neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56; with permission.)

82 POPP & BOTTIGER

support this practice directly. The biologic plausibility of the benefit fromcontrolling hyperglycemia with insulin infusion is strong, however; andtherefore there is a ‘low grade’ recommendation to adjust blood glucose lev-els to between 90 and 145 mg/dL [89].

Future perspectives

More experimental approaches are focusing on the use of a hibernatingstate after cardiac arrest. Various animals use hibernation during times ofextreme environmental conditions like low temperature or food and waterdeprivation. Hibernation in mammals is characterized by a physiologic statewith significant reduction in body-core temperature and metabolic rate [90]resulting from a reduced set-point of thermoregulation. Those physiologicchanges provide strong resistance to cerebral ischemia [91]. Reducing body-core temperature by way of modulation of the set-point, as done withinhibernation, might be a superior concept to forced hypothermia becauseof the avoidance of the homeostatic mechanisms counteracting reductionsin body temperature. Physiologic stress accompanied by a myriad of re-sponses like shivering and increased catecholamine and cortisol levels duringforced hypothermia might decrease cooling speed and efficacy of thehypothermic treatment [92]. Therefore the evaluation of controlled hypo-thermia- or hibernation-inducing drugs in future studies seems mandatory.Several substances have been tested for their efficiency in inducing regulatedhypothermia/hibernation. The hibernation-induction trigger, an 88 kdpeptide found in the serum of hibernating ground squirrels, can increasethe survival time in a multiorgan preparation model with dogs [93]. Withinthe same model, it was shown that a delta opiode receptor agonist likeD-Ala2, D-Leu5-enkephalin (DADLE) extends hypothermic preservationtime of the lung [94]. A modified neurotensin 77 given intravenously isable to induce hibernation for several hours in rats [95] and improves neu-rologic outcome after hypoxic-ischemia [96]. Recently an article in Sciencedemonstrated that H2S induces a suspended–animation-like state in miceby way of inhibiting oxidative phosphorylation [97]. Taken together,there are several at least theoretical possibilities for inducing regulatedhypothermia by way of modulation of the thermoregulatory set-point.

Clinical practice

Neuronal injury following global cerebral ischemia continues to be a cen-tral problem of patients in the postresuscitation phase. Particular attentionmust be paid to measures that serve to preserve neurologic function. Besidesall measures focusing on rapid restoration of spontaneous circulation, suchas the use of defibrillators to treat ventricular fibrillation and the choice ofa suitable vasopressor, several postarrest treatment options have been ex-plored in recent years. Probably the most effective treatment after cardiacarrest, as shown by large randomized trials, is the use of therapeutic mild


hypothermia. Current guidelines of ILCOR recommend the use of therapeu-tic mild hypothermia for all unconscious patients after cardiac arrest.

To date, there is no specific neuroprotective treatment available. How-ever, good practice of critical care, such as the control of blood glucose,blood pressure, and oxygenation, should be provided. Promising animal ex-perimental data concerning the use of thrombolytic agents during cardiopul-monary resuscitation has led to a large European multicenter trial(TROICA trial) that will provide its data in 2006.

References

[1] Bottiger BW, Grabner C, Bauer H, et al. Long-term outcome after out-of-hospital cardiac

arrest with physician staffed emergency medical services: the Utstein style applied to a mid-

sized urban/suburban area. Heart 1999;82:674–9.

[2] Hossmann KA. Ischemia-mediated neuronal injury. Resuscitation 1993;26:225–35.

[3] Safar P. Cerebral resuscitation after cardiac arrest: a review. Circulation 1986;74:Iv138–53.

[4] Earnest MP, Yarnell PR, Merrill SL, et al. Long-term survival and neurologic status after

resuscitation from out-of-hospital cardiac arrest. Neurology 1980;30:1298–302.

[5] Hallstrom A, Boutin P, Cobb L, et al. Socioeconomic status and prediction of ventricular

fibrillation survival. Am J Public Health 1993;83:245–8.

[6] Madl C, GrimmG, Kramer L, et al. Early prediction of individual outcome after cardiopul-

monary resuscitation. Lancet 1993;341:855–8.

[7] Safar P, BehringerW, Bottiger BW, et al. Cerebral resuscitation potentials for cardiac arrest.

Crit Care Med 2002;30:140–4.

[8] Bottiger BW, Schmitz B, Wiessner C, et al. Neuronal stress response and neuronal cell dam-

age after cardiocirculatory arrest in rats. J Cereb Blood Flow Metab 1998;18:1077–87.

[9] Padosch SA, Bottiger BW. Neuronal apoptosis following cerebral ischaemia: pathophysiol-

ogy and possible therapeutic implications. Current OpinionAnesthesiology 2003;16:439–45.

[10] Brain Resuscitation Clinical Trial I Study Group. Randomized clinical study of thiopental

loading in comatose survivors of cardiac arrest. N Engl J Med 1986;314:397–403.

[11] Brain Resuscitation Clinical Trial II StudyGroup. A randomized clinical study of a calcium-

entry blocker (lidoflazine) in the treatment of comatose survivors of cardiac arrest. N Engl J

Med 1991;324:1225–31.

[12] Bottiger BW, Teschendorf P, Krumnikl JJ, et al. Global cerebral ischemia due to cardiocir-

culatory arrest in mice causes neuronal degeneration and early induction of transcription

factor genes in the hippocampus. Brain Res Mol Brain Res 1999;65:135–42.

[13] Popp E, Padosch SA, Vogel P, et al. Effects of intracerebroventricular application of brain-

derived neurotrophic factor on cerebral recovery after cardiac arrest in rats. Crit Care Med

2004;32:S359–65.

[14] Krep H, Brinker G, Pillekamp F, et al. Treatment with an endothelin type A receptor-antag-

onist after cardiac arrest and resuscitation improves cerebral hemodynamic and functional

recovery in rats. Crit Care Med 2000;28:2866–72.

[15] Fischer M, Bottiger BW, Popov Cenic S, et al. Thrombolysis using plasminogen activator

and heparin reduces cerebral no-reflow after resuscitation from cardiac arrest: an experimen-

tal study in the cat. Intensive Care Med 1996;22:1214–23.

[16] Fischer M, Hossmann KA. Volume expansion during cardiopulmonary resuscitation re-

duces cerebral no-reflow. Resuscitation 1996;32:227–40.

[17] Safar P, Stezoski W, Nemoto EM. Amelioration of brain damage after 12 minutes’ cardiac

arrest in dogs. Arch Neurol 1976;33:91–5.

[18] Bernard SA, Gray TW, BuistMD, et al. Treatment of comatose survivors of out-of-hospital


84 POPP & BOTTIGER

[19] The Hypothermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to im-

prove the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56.

[20] Nolan J, Morley P, Vanden Hoek T, et al. Therapeutic hypothermia after cardiac arrest: an

advisory statement by the Advanced Life Support Task Force of the International Liaison

Committee on Resuscitation. Circulation 2003;108:118–21.

[21] Ashkenazi A, Dixit VM. Death receptors: signaling and modulation. Science 1998;281:

1305–8.

[22] Padosch SA, Vogel P, Bottiger BW. Neuronal apoptosis following cerebral ischemia. Basis,

physiopathology and treatment strategies. Anaesthesist 2001;50:905–20.

[23] Eldadah BA, Faden AI. Caspase pathways, neuronal apoptosis, and CNS injury. J Neuro-

trauma 2000;17:811–29.

[24] MacManus JP, LinnikMD.Gene expression induced by cerebral ischemia: an apoptotic per-

spective. J Cereb Blood Flow Metab 1997;17:815–32.

[25] Gillardon F, Bottiger BW, Schmitz B, et al. Activation of CPP-32 protease in hippocampal

neurons following ischemia and epilepsy. Brain Res Mol Brain Res 1997;50:16–22.

[26] Vogel P, Putten H, Popp E, et al. Improved resuscitation after cardiac arrest in rats express-

ing the baculovirus caspase inhibitor protein p35 in central neurons. Anesthesiology 2003;99:

112–21.

[27] Teschendorf P, Popp E, Motsch J, et al. Effects of the inhibition of caspases on neuronal de-

generation and outcome following global cerebral ischemia due to cardiocirculatory arrest in

rats. Anesthesiology 2001;95:A788.

[28] Ebendal T. Function and evolution in the NGF family and its receptors. J Neurosci Res

1992;32:461–70.

[29] Hyman C, HoferM, Barde YA, et al. BDNF is a neurotrophic factor for dopaminergic neu-

rons of the substantia nigra. Nature 1991;350:230–2.

[30] Ip NY, Li Y, Yancopoulos GD, et al. Cultured hippocampal neurons show responses to

BDNF, NT-3, and NT-4, but not NGF. J Neurosci 1993;13:3394–405.

[31] Lindholm D, Dechant G, Heisenberg CP, et al. Brain-derived neurotrophic factor is a sur-

vival factor for cultured rat cerebellar granule neurons and protects them against gluta-

mate-induced neurotoxicity. Eur J Neurosci 1993;5:1455–64.

[32] Shimohama S, Tamura Y, Akaike A, et al. Brain-derived neurotrophic factor pretreatment

exerts a partially protective effect against glutamate-induced neurotoxicity in cultured rat

cortical neurons. Neurosci Lett 1993;164:55–8.

[33] Oppenheim RW, Yin QW, Prevette D, et al. Brain-derived neurotrophic factor rescues de-

veloping avian motoneurons from cell death. Nature 1992;360:755–7.

[34] Sendtner M, Holtmann B, Kolbeck R, et al. Brain-derived neurotrophic factor prevents the

death of motoneurons in newborn rats after nerve section. Nature 1992;360:757–9.

[35] SchabitzWR, Schwab S, SprangerM, et al. Intraventricular brain-derived neurotrophic fac-

tor reduces infarct size after focal cerebral ischemia in rats. J Cereb Blood FlowMetab 1997;

17:500–6.

[36] Yamashita K,Wiessner C, LindholmD, et al. Post-occlusion treatment with BDNF reduces

infarct size in a model of permanent occlusion of the middle cerebral artery in rat. Metab

Brain Dis 1997;12:271–80.

[37] Choi DW. Calcium and excitotoxic neuronal injury. Ann N Y Acad Sci 1994;747:162–71.

[38] Mattson MP, Lovell MA, Furukawa K, et al. Neurotrophic factors attenuate glutamate-

induced accumulation of peroxides, elevation of intracellular Ca2 þ concentration, and

neurotoxicity and increase antioxidant enzyme activities in hippocampal neurons. J Neuro-

chem 1995;65:1740–51.

[39] Schabitz WR, Hoffmann TT, Heiland S, et al. Delayed neuroprotective effect of insulin-like

growth factor-i after experimental transient focal cerebral ischemia monitored with MRI.

Stroke 2001;32:1226–33.

[40] Popp E, Vogel P, Teschendorf P,et al. Effekte der intrazerebroventrikularen Applikation

von ‘‘insulin-like growth factor 1’’ und seines Spaltproduktes GPE auf die neuronale


Degeneration nach Herz-Kreislaufstillstand. Abstractband 52 Deutscher Anasthesiecongress

April 2005, Munich, Germany.

[41] Hossmann KA, Lechtape-Gruter H, Hossmann V. The role of cerebral blood flow for the

recovery of the brain after prolonged ischemia. Z Neurol 1973;204:281–99.

[42] Leonov Y, Sterz F, Safar P, et al. Hypertension with hemodilution prevents multifocal cere-

bral hypoperfusion after cardiac arrest in dogs. Stroke 1992;23:45–53.

[43] Schmid-Schonbein GW. Capillary plugging by granulocytes and the no-reflow phenomenon

in the microcirculation. Fed Proc 1987;46:2397–401.

[44] Schmid-Schonbein GW, Lee J. Leukocytes in capillary flow. Int J Microcirc Clin Exp 1995;

15:255–64.

[45] Bottiger BW, Motsch J, Braun V, et al. Marked activation of complement and leukocytes

and an increase in the concentrations of soluble endothelial adhesion molecules during car-

diopulmonary resuscitation and early reperfusion after cardiac arrest in humans. Crit Care

Med 2002;30:2473–80.

[46] Lechleitner P, Mair J, Genser N, et al. Granulocyte elastase in acute myocardial infarction.

Z Kardiol 1993;82:641–7.

[47] Mussack T, Biberthaler P, Gippner-Steppert C, et al. Early cellular brain damage and sys-

temic inflammatory response after cardiopulmonary resuscitation or isolated severe head

trauma: a comparative pilot study on common pathomechanisms. Resuscitation 2001;49:

193–9.

[48] Bottiger BW, Bohrer H, Boker T, et al. Platelet factor 4 release in patients undergoing car-

diopulmonary resuscitation–can reperfusion be impaired by platelet activation?ActaAnaes-

thesiol Scand 1996;40:631–5.

[49] Bottiger BW, Bauer H, Motsch J, et al. Role of thrombolysis in resuscitation. Lancet 2001;

358:1371–2.

[50] Bottiger BW,Motsch J, Bohrer H, et al. Activation of blood coagulation after cardiac arrest

is not balanced adequately by activation of endogenous fibrinolysis. Circulation 1995;92:

2572–8.

[51] Gando S, Kameue T, Nanzaki S, et al. Massive fibrin formation with consecutive impair-

ment of fibrinolysis in patients with out-of-hospital cardiac arrest. Thromb Haemost

1997;77:278–82.

[52] Bottiger BW, Martin E. Thrombolytic therapy during cardiopulmonary resuscitation

and the role of coagulation activation after cardiac arrest. Curr Opin Crit Care 2001;7:

176–83.

[53] Crowell JW, Sharpe GP, Lambright RL, et al. The mechanism of death after resuscitation

following acute circulatory failure. Surgery 1955;38:696–702.

[54] Crowell JW, Smith EE. Effect of fibrinolytic activation on survival and cerebral damage fol-

lowing periods of circulatory arrest. Am J Physiol 1956;186:283–5.

[55] Lin SR, O’ConnorMJ, Fischer HW, et al. The effect of combined dextran and streptokinase

on cerebral function and blood flow after cardiac arrest: an experimental study on the dog.

Invest Radiol 1978;13:490–8.

[56] Safar P, Stezoski W, Nemoto EM. Amelioration of brain damage after 12 minutes cardiac

arrest in dogs. Arch Neurol 1976;33:91–5.

[57] Bottiger BW. Thrombolysis during cardiopulmonary resuscitation. Fibrinolysis 1997;

11(Suppl2):93–100.

[58] Bottiger BW, Bohrer H, Bach A, et al. Bolus injection of thrombolytic agents during cardio-

pulmonary resuscitation for massive pulmonary embolism. Resuscitation 1994;28:45–54.

[59] Bottiger BW,Bode C,Kern S, et al. Efficacy and safety of thrombolytic therapy after initially

unsuccessful cardiopulmonary resuscitation: a prospective clinical trial. Lancet 2001;357:

1583–5.

[60] Lederer W, Lichtenberger C, Pechlaner C, et al. Recombinant tissue plasminogen activator

during cardiopulmonary resuscitation in 108 patients with out-of-hospital cardiac arrest.

Resuscitation 2001;50:71–6.

86 POPP & BOTTIGER

[61] Bottiger BW, Spohr F. The risk of thrombolysis in association with cardiopulmonary resus-

citation: no reason to withhold this causal and effective therapy. J Intern Med 2003;253:

99–101.

[62] Spohr F, Bottiger BW. Safety of thrombolysis during cardiopulmonary resuscitation. Drug

Saf 2003;26:367–79.

[63] Spohr F, Bottiger BW. Thrombolytic therapy during or after cardiopulmonary resuscita-

tion. Efficacy and safety of a new therapeutic approach.MinervaAnestesiol 2003;69:357–64.

[64] Williams GR Jr, Spencer FC. The clinical use of hypothermia following cardiac arrest. Ann

Surg 1958;148:462–8.

[65] Rosomoff HL, Holaday DA. Cerebral blood flow and cerebral oxygen consumption during

hypothermia. Am J Physiol 1954;179:85–8.

[66] Xu L, Yenari MA, Steinberg GK, et al. Mild hypothermia reduces apoptosis of mouse neu-

rons in vitro early in the cascade. J Cereb Blood Flow Metab 2002;22:21–8.

[67] Hashimoto T, YonetaniM,NakamuraH. Selective brain hypothermia protects against hyp-

oxic-ischemic injury in newborn rats by reducing hydroxyl radical production. Kobe J Med

Sci 2003;49:83–91.

[68] Kil HY, Zhang J, Piantadosi CA. Brain temperature alters hydroxyl radical production dur-

ing cerebral ischemia/reperfusion in rats. J Cereb Blood Flow Metab 1996;16:100–6.

[69] Takata K, Takeda Y, Sato T, et al. Effects of hypothermia for a short period on histologic

outcome and extracellular glutamate concentration during and after cardiac arrest in rats.

Crit Care Med 2005;33:1340–5.

[70] Busto R, Globus M, Dietrich W, et al. Effect of mild hypothermia on ischemia-induced re-

lease of neurotransmitters and free fatty acids in rat brain. Stroke 1989;20:904–10.

[71] Hachimi-Idrissi S, Van Hemelrijck A, Michotte A, et al. Postischemic mild hypothermia re-

duces neurotransmitter release and astroglial cell proliferation during reperfusion after as-

phyxial cardiac arrest in rats. Brain Res 2004;1019:217–25.

[72] Fischer S, Renz D, Wiesnet M, et al. Hypothermia abolishes hypoxia-induced hyperperme-

ability in brain microvessel endothelial cells. Brain Res Mol Brain Res 1999;74:135–44.

[73] Husby P, AndersenKS, Owen-Falkenberg A, et al. Accidental hypothermia with cardiac ar-

rest: complete recovery after prolonged resuscitation and rewarming by extracorporeal cir-

culation. Intensive Care Med 1990;16:69–72.

[74] Fay T. Observation on generalized refrigeration in cases of severe cerebral trauma. Res Publ

Assoc Res Nerv Ment Dis 1945;4:611–9.

[75] BigelowWG, Callaghan JC, Hopps JA. General hypothermia for experimental intracardiac

surgery; the use of electrophrenic respirations, an artificial pacemaker for cardiac standstill,

and radio-frequency rewarming in general hypothermia. Trans Meet Am Surg Assoc Am

Surg Assoc 1950;68:211–9.

[76] Botterell EH, Lougheed WM, Morley TP, et al. Hypothermia in the surgical treatment of

ruptured intracranial aneurysms. J Neurosurg 1958;15:4–18.

[77] Botterell EH, LougheedWM, Scott JW, et al. Hypothermia, and interruption of carotid, or

carotid and vertebral circulation, in the surgical management of intracranial aneurysms.

J Neurosurg 1956;13:1–42.

[78] Weinrauch V, Safar P, Tisherman S, et al. Beneficial effect of mild hypothermia and detri-

mental effect of deep hypothermia after cardiac arrest in dogs. Stroke 1992;23:1454–62.

[79] Bernard SA, Jones BM, Horne MK. Clinical trial of induced hypothermia in comatose sur-

vivors of out-of-hospital cardiac arrest. Ann Emerg Med 1997;30:146–53.

[80] Eisenburger P, Sterz F, Holzer M, et al. Therapeutic hypothermia after cardiac arrest. Curr

Opin Crit Care 2001;7:184–8.

[81] Leonov Y, Sterz F, Safar P, et al. Mild cerebral hypothermia during and after cardiac arrest

improves neurologic outcome in dogs. J Cereb Blood Flow Metab 1990;10:57–70.

[82] Dixon SR,Whitbourn RJ, DaeMW, et al. Induction of mild systemic hypothermia with en-

dovascular cooling during primary percutaneous coronary intervention for acute myocar-

dial infarction. J Am Coll Cardiol 2002;40:1928–34.


[83] Longstreth WT Jr, Inui TS. High blood glucose level on hospital admission and poor neu-

rological recovery after cardiac arrest. Ann Neurol 1984;15:59–63.

[84] Longstreth WT Jr, Diehr P, Cobb LA, et al. Neurologic outcome and blood glucose levels

during out-of-hospital cardiopulmonary resuscitation. Neurology 1986;36:1186–91.

[85] MullnerM, Sterz F, BinderM, et al. Blood glucose concentration after cardiopulmonary re-

suscitation influences functional neurological recovery in human cardiac arrest survivors.

J Cereb Blood Flow Metab 1997;17:430–6.

[86] Voll CL, AuerRN. The effect of postischemic blood glucose levels on ischemic brain damage

in the rat. Ann Neurol 1988;24:638–46.

[87] Voll CL,Whishaw IQ,AuerRN. Postischemic insulin reduces spatial learning deficit follow-

ing transient forebrain ischemia in rats. Stroke 1989;20:646–51.

[88] van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in the critically ill

patient. N Engl J Med 2001;345:1359–67.

[89] Bell DD, Brindley PG, Forrest D, et al. Management following resuscitation from cardiac

arrest: recommendations from the 2003 Rocky Mountain Critical Care Conference. Can J

Anaesth 2005;52:309–22.

[90] Frerichs KU, Dienel GA, Cruz NF, et al. Rates of glucose utilization in brain of active and

hibernating ground squirrels. Am J Physiol 1995;268:R445–53.

[91] Frerichs KU, Hallenbeck JM. Hibernation in ground squirrels induces state- and species-

specific tolerance to hypoxia and aglycemia: an in vitro study in hippocampal slices. J Cereb

Blood Flow Metab 1998;18:168–75.

[92] Gordon CJ. The therapeutic potential of regulated hypothermia. Emerg Med J 2001;18:

81–9.

[93] Chien S, Oeltgen PR, Diana JN, et al. Two-day preservation of major organs with autoper-

fusion multiorgan preparation and hibernation induction trigger. A preliminary report.

J Thorac Cardiovasc Surg 1991;102:224–34.

[94] WuG, Zhang F, Salley RK, et al. delta Opioid extends hypothermic preservation time of the

lung. J Thorac Cardiovasc Surg 1996;111:259–67.

[95] Tyler BM,Douglas CL, Fauq A, et al. In vitro binding and CNS effects of novel neurotensin

agonists that cross the blood-brain barrier. Neuropharmacology 1999;38:1027–34.

[96] Katz LM,YoungA, Frank JE, et al.Neurotensin-induced hypothermia improves neurologic

outcome after hypoxic-ischemia. Crit Care Med 2004;32:806–10.

[97] BlackstoneE,MorrisonM,RothMB.H2S induces a suspended animation-like state inmice.

Science 2005;308:518.

Neurol Clin 24 (2006) 89–106

Clinical Neurophysiologic Monitoringand Brain Injury from Cardiac Arrest

Matthew A. Koenig, MDa,*, Peter W. Kaplan, MBBSb,Nitish V. Thakor, PhDc

aDepartment of Neurology, Johns Hopkins University School of Medicine,

Baltimore, MD, USAbDepartment of Neurology, Johns Hopkins Bayview Medical Center,

Baltimore, MD, USAcDepartment of Biomedical Engineering, Johns Hopkins University School of Medicine,

Baltimore, MD, USA

With the introduction of closed chest compressions and the advent ofmodern cardiopulmonary resuscitation (CPR) techniques in the 1960s,physicians gained a simple means of restoring circulation after cardiac ar-rest. Although basic CPR techniques and postresuscitation intensive caremanagement have improved during the past 4 decades, outcomes after car-diac arrest have remained static. Only approximately 50% of CPR attemptsrestore spontaneous circulation successfully [1]. Of these survivors, 80% re-main unconscious in the immediate postresuscitative period and only 10%to 20% ever have meaningful neurologic recovery [1,2]. Much effort hasbeen devoted to early prognostication while patients remain in coma aftersuccessful CPR. Given the ethical, familial, and economic burden of con-tinuing intensive care management of patients unlikely to achieve meaning-ful neurologic recovery, there exists a profound need for accurate methodsof early injury stratification.

Investigational tools used in early prognostication include cerebrospinalfluid markers of brain injury, electrophysiologic monitoring early and latein the recovery period, functional and anatomic imaging techniques, micro-dialysis of markers implicated in secondary brain injury, and calculation ofbrain oxygen extraction. This review focuses on the use of electrophysiologictools, electroencephalography (EEG) and somatosensory evoked potentials(SSEP), in monitoring brain function after resuscitation from cardiac arrest.

* Corresponding author. Johns Hopkins Hospital, Department of Neurology, 600 North

Wolfe Street, Baltimore, MD 21287.

E-mail address: [email protected] (M.A. Koenig).




90 KOENIG et al

EEG has been used since the 1960s to prognosticate in patients who are co-matose after cardiac arrest. Many injury stratification schemes have been de-vised and several malignant patterns identified, the best validated of whichare reviewed. The limitation of EEG for prognostication is that althoughsome patterns seem specific for poor outcomes, the prognostic accuracy ofEEG is poor unless those patterns are present. SSEP has been studied morerecently and seems to offer higher specificity for determining bad outcomesbut probably is less accurate than EEG in determining which patients are des-tined to have good recovery. EEG and evoked potentials have remained pop-ular for evaluation of patients in coma after cardiac arrest because they arenoninvasive, widely available, inexpensive to perform and interpret, familiarto referring physicians, and validated for this purpose.

The traditional role of neurologists in the care of cardiac arrest survivors isto prognosticate based on a combination of neurologic signs and electrophys-iologic tests that can be performed hours to days after resuscitation. Withsome exceptions, neurologists rarely are involved during the resuscitative ef-fort or immediate postarrest care. From an investigational standpoint, the pe-riod immediately after resuscitation involves dynamic changes in regionalbrain perfusion, secondary neuronal injury, and electrophysiologic recovery.Rather than focusing on the usefulness of electrophysiologic monitoring forprognostication, investigational techniques are explored for using EEG andevoked potentials in real-time monitoring of coma arousal after cardiac ar-rest, early injury stratification, defining a timewindow for potential neuropro-tective strategies, and monitoring response to therapeutic efforts. In addition,this article reviews advances of quantitative EEG and SSEP techniques intracking injury and recovery in several animal models of cardiac arrest.

Electroencephalography

Over the years, many grading scales have been proposed to predict survivaland neurologic outcomes in comatose survivors of cardiac arrest, but themost popular scale was published by Hockaday and colleagues in 1965 [3].This scale divided the EEG into five distinct grades based on the dominantfrequency and presence of specific unfavorable patterns (Table 1). TheEEG grades subsequently have been demonstrated to correlate with progno-sis, especially when unfavorable EEG patternsdincluding electrocerebral si-lence (ECS) and burst suppressiondare present more than 24 hours aftercardiac arrest. Prior [4] reports EEG findings from 96 comatose survivorsof cardiac arrest. All 25 patients who had sustained ECS and 25 of 26 patientswho had burst suppression remained in coma indefinitely or died. Of the45 patients who had benign EEG tracings, most recovered initially to someextent, but 31 subsequently succumbed to underlying systemic diseases.Møller and coworkers [5] report EEG findings from 185 patients who hadcardiac arrest after myocardial infarction. In this population, 70%of patientswho had grade I or II EEG (n¼ 115) survived to discharge with no neurologic

91BRAIN INJURY FROM CARDIAC ARREST

deficit or minor impairment, whereas only 3% (n¼ 70) who had grades III toV EEG survived. In the high-grade survivors, serial EEG testing demon-strates a shift from high-grade to low-grade patterns over subsequent days.These findings emphasize that EEG patterns are dynamic after cardiac arrest,and application of the standard grading scale early after resuscitation seemsto have lower validity. Different iterations of the coma grading scale have beenapplied in many studies of prognosis after cardiac arrest. Scollo-Lavizzariand Bassetti [6] reviewed the literature on prognosis of comatose survivorsof cardiac arrest by EEG recorded at least 6 hours after resuscitation. Review-ing the records of 408 patients, they found meaningful neurologic recoveryin 78% of grade I, 0% of grades IV and V, and intermediate recovery ingrades II and III (Table 2).

The lack of a uniform patient population renders comparison betweenstudies difficult. In particular, the level of arousal at the time of EEG, dura-tion of coma, and interval between resuscitation and EEG recording allchange the prognostic value. Edgren and colleagues [2] report EEG tracingsin 32 patients recorded 24 hours after cardiac arrest. Patients in this serieshad at least 10 minutes of coma after cardiac arrest, but many were notin coma at the time the EEG was recorded. In this population, a good out-come was achieved in 7 of 8 patients who had grade I EEG, 5 of 13 who had

Table 1

Classification of electroencephalogram findings

Grade I Dominant, normal alpha activity

Dominant, normal alpha activity with theta-delta activities

Grade I Dominant theta-delta activity with still detectable normal alpha activities

Grade III Theta-delta activity without alpha activities

Grade IV Delta activity, low voltage, possibly with short isoelectric intervals

Dominant, monomorphic, nonreactive alpha activity (alpha coma)

Periodic generalized phenomena (spikes, sharp waves, slow waves) with very

low-voltage background activity

Grade V Very flat to isoelectric EEG (less than 10–20 mV)

From Scollo-Lavizzari, G, Bassetti C. Prognostic value of EEG in post-anoxic coma after

cardiac arrest. Eur Neurol 1987;26:161–70.

Table 2

Prognosis with the five electroencephalogram grades given (408 cases from the literature)

Electroencephalogram

category Recovery (%)

Survival with permanent

neurologic damage (%) Death (%)

Grade I 79 10 11

Grade II 51 13 36

Grade III 26 7 67

Grade IV 0 2 98

Grade V 0 0 100

From Scollo-Lavizzari, G, Bassetti C. Prognostic value of EEG in post-anoxic coma after


92 KOENIG et al

grade II, and none of 8 who had higher-grade findings at 24 hours. Alterna-tively, Bassetti and coworkers [7] report on 60 patients who were in coma formore than 6 hours after resuscitation from cardiac arrest. In this study, all20 patients who had grade IV or V EEG either died or remained in comaindefinitely and only 4 of 10 patients who had favorable EEG (grade I orII) had a good neurologic recovery. In another series [8] of 64 patients incoma for greater than 24 hours after cardiac arrest, there was no correlationbetween EEG grade and duration of survival. In this population, only 2 of 8patients who had grade I EEG patterns survived, both with mental impair-ment. The discrepancies between these results emphasize the degree to whichthe prognostic grading scale is changed by the population studied. The du-ration of coma is an independent risk factor for poor outcome, so normal ormildly abnormal EEG findings are an unreliable predictor of good outcomein prolonged coma. In addition, high-grade EEG findings have low sensitiv-ity, identifying only half of patients who are comatose who die or remain ina vegetative state.

Binnie and colleagues [9] drew early attention to the concept that regard-less of the dominant EEG frequency, the failure of EEG tracings to react toexternal stimuli portends a poor prognosis for meaningful recovery. Analyz-ing the EEG characteristics beyond dominant frequency may improve thediagnostic accuracy of low-grade EEG patterns. Synek [10,11] emphasizesthe importance of reactivity and amplitude of the EEG pattern and the pres-ence or absence of epileptiform activity beyond the dominant frequency. Hereports that the diagnostic accuracy of the EEG grading scale improves to98.4% when the scale is altered to include reactivity, epileptiform activity,and amplitude at each of the first three grades [10]. Although numberswere small in each group, survival occurred in 2 of 2 patients who hadreactive grade III EEG, 3 of 4 patients who had unreactive grade III, 2of 4 patients who had grade III EEG and epileptiform activity, and noneof 3 patients who had low-amplitude grade III EEG [11].

Generalized electrical suppression

Some patterns of EEG activity seem to portend an extremely poor prog-nosis in comatose survivors of cardiac arrest, especially when reported morethan 24 hours after resuscitation. Several reports emphasize the prognosticgravity of ECS, burst suppression, and generalized alpha or theta activityunreactive to external stimulation (alpha and theta coma) [12]. There areno generally accepted criteria for ECS, although some investigators attemptto provide voltage criteria. Young [13] reports survival in 1 patient who hadincomplete EEG suppressiondvoltage between 10 and 20 mVdafter24 hours but none who had voltage lower than 10 mV. There are no reportsin the adult or pediatric literature of survival with true ECS that was recor-ded more than 24 hours after resuscitation. In normothermic patients whoare free of toxic ingestion or pharmacologic sedation, the presence of ECS


more than 24 hours after cardiac arrest is the only undisputed indication ofno chance of meaningful neurologic recovery.

Generalized burst suppression

Generalized burst suppression after cardiac arrest usually, but not invari-ably, is associated with no prognosis for meaningful neurologic recovery[13]. The term, burst suppression, indicates bursts of electrical activity ofvariable duration separated by periods of generalized suppression lastingat least 1 second [14]. The bursts may range from high-amplitude delta acti-vity or polyspike and slow wave complexes with epileptiform characteristics[14]. The physiology of burst suppression remains to be elucidated fully, butthe generally accepted theory relates burst suppression to functional disso-ciation of the cortex from the intrinsic pacemaker firing of neurons in thereticular thalamus [15,16]. As such, persistent burst suppression seems tobe a marker for severe injury to the thalamus, cerebral cortex, and intercon-necting relay circuits. Burst suppression may occur alone or in combinationwith underlying low-voltage unreactive alpha or theta activity [17,18]. Burstsuppression is associated with a universally fatal outcome or persistent veg-etative state (PVS) in multiple series [19,20]. There are scattered reports ofsurvival and good neurologic recovery in patients who are comatose withburst suppression after cardiac arrest [21,22], especially when the EEG isperformed early after resuscitation and epileptiform activity is absent. Spon-taneous movement is seen variably in association with bursts and can rangefrom simple mouth movements to violent, generalized myoclonic jerks of theextremities. The presence or absence of movement during bursts does notseem to alter the prognostic implications of this EEG pattern, however[23]. Like ECS, it is important to ensure that patients have not been exposedto anesthetic agents or toxic medication overdose before placing undueprognostic importance on burst suppression.

Alpha and theta coma

Reactivity of the EEG carries greater prognostic importance than thedominant frequency alone. Patients who have the normal posteriorly dom-inant pattern of alpha activity that is suppressed by eye opening and en-hanced by eye closure or noxious stimulation have a good prognosis formeaningful recovery in the appropriate clinical context. This pattern is instark contrast to comatose cardiac arrest survivors who have generalizedor frontally dominant alpha patterns with constant, unwavering frequencyand amplitude that is unresponsive to stimulation. Also known as alphacomador theta coma when extending into the fast theta banddthis EEGpattern may be seen in 10% to 15% of cases of postanoxic coma [5,24].The physiology of alpha coma and its distinction from physiologic alpha re-main to be elucidated. It probably is regarded best, however, as a transitionalpattern that evolves into a more prognostically definitive pattern in 90% of

94 KOENIG et al

patients within 6 days [25]. On serial EEG, those patients who die withoutregaining consciousness evolve into burst suppression and survivors regainelectrographic reactivity to external stimuli [12,26]. Chokroverty [27] reportssurvival of 2 of 12 patients who had alpha coma patterns after cardiac ar-rest. In both cases, the EEG was recorded early after resuscitation and serialtesting revealed evolution to more benign patterns. In a large series of pa-tients who underwent continuous EEG recording immediately after CPR,however, alpha coma patterns were detected early after resuscitation in31 patients, 25 of whom eventually regained consciousness and 9 of whomprogressed to full recovery [28]. Overall, however, alpha and theta comaare associated with an 88% chance of death or PVS according to a meta-analysis of the world literature [12].

Electroencephalogram patterns evolve after resuscitation

The proximity of the first EEG recording to the time of resuscitation aftercardiac arrest has a major impact on the distribution and prognostic impor-tance of electrical patterns (Figs. 1 and 2). Several investigators report EEGfindings in patients during cardiac arrest and resuscitation. Losasso andcoworkers [29] report a case of asystole lasting 2 minutes in a patient under-going continuous EEG monitoring during carotid endarterectomy. Within10 seconds of asystole, there was generalized suppression on the EEG,

Fig. 1. Burst-suppression EEG pattern after cardiac arrest demonstrating periodic generalized

bursts of sharply contoured electrical activity separated by periods of EEG suppression.


leading immediately to ECS. Low-voltage, high-frequency activity began toreturn within 15 to 20 seconds of external chest compressions with gradualre-emergence of baseline EEG rhythm. Another patient had EEG monitor-ing during a 27-second episode of asystole [30]. In this case, the EEG re-turned immediately to baseline after return of spontaneous circulation(ROSC). Case reports using bispectral index [31] and compressed spectralarray monitoring of EEG [32] during intraoperative cardiac arrest yield sim-ilar findings. Clute and Levy [33], in an unusual study design, monitoredEEG changes in humans undergoing automated internal cardioverting defi-brillator (AICD) placement. During AICD testing, patients routinely under-went brief induction of ventricular fibrillation resulting in a controlledperiod of cardiac arrest. In this study, the mean time of onset of EEGchange from baseline was 10.2 seconds (range 3.3–21.1 seconds) after induc-tion of ventricular fibrillation. The majority of patients demonstrated slow-ing and attenuation or sudden loss of activity above the delta range. EEGactivity returned to baseline immediately after cardioversion.

Several series are published of serial EEG recordings or continuous mon-itoring of patients within hours of resuscitation from cardiac arrest. Pampi-glione [34] reports the frequent occurrence of burst suppression within hoursof cardiac arrest in a series of children. A study of EEG recorded within3 hours of birth asphyxia in full term infants, however, reports survival

Fig. 2. Alpha coma EEG pattern after cardiac arrest demonstrating unwavering generalized

alpha frequency electrical activity unresponsive to stimulation and lacking the normal posterior

predominance.

96 KOENIG et al

with good neurologic outcomes in 3 infants whose initial EEG had burstsuppression [35]. In these cases, the EEG evolved from burst suppressionto continuous activity within 3 hours. All of these infants, however, hadbeen treated with phenobarbital before the EEG recording. In a series of42 cases of cardiac arrest survivors who underwent EEG within the first12 hours after resuscitation, Cloche and coworkers [22] report the EEG pat-tern changed within the first 48 hours, including five cases in which ECSevolved into a reactive pattern after several hours. Regardless of changesin the EEG pattern, however, they found that patients who had an initialEEG demonstrating ECS or burst suppression had a uniformly fatal out-come. Similar results were reported in a series of 371 patients studiedsoon after resuscitation [36], although exact times were not provided.

A more definitive description of the evolution of human EEG in thehours after resuscitation comes from an important series of publicationsfrom Jørgensen and colleagues. A neurologist was present during the resus-citative effort and monitored EEG and neurologic examination findingscontinuously during and after ROSC. In a series of 37 patients who regainedconsciousness after resuscitation from cardiac arrest, all patients displayedECS immediately upon re-establishment of circulation [37]. During thisperiod, decerebrate posturing and return of cranial nerve reflexes oftenoccurred. The first electrical activity resumed 10 minutes to 8 hours there-after. In most cases, the EEG initially showed burst suppression beforefusion to a continuous rhythm. Decorticate posturing and myoclonus some-times occurred during this phase of recovery. Burst suppression lasted aslong as 16 hours before fusion in 1 patient. In the minority of cases, contin-uous electrical activity occurred without a preceding period of burst sup-pression. In all cases, the frequency and amplitude of continuous activityincreased with time. Re-emergence of continuous activity was correlatedpoorly with arousal, however, and predominant alpha activity often was re-ported before awakening [37]. In a second series of 88 patients who did notregain consciousness after CPR, 71 patients regained electrocortical activityat some point after resuscitation [38]. Of these patients, only 12 regainedcontinuous electrical activity, whereas the remainder continued in burst sup-pression indefinitely. The time of initial electrical activity was significantlylater in these patients, ranging from 15 minutes to 124 hours after resus-citation. In the small number of patients who regained continuous EEG acti-vity, the interim to EEG fusion also was longer (range 4 to 127 hours)compared with patients who emerged ultimately from coma. In a third seriesof 231 patients who underwent EEG monitoring immediately after CPR,those who had immediate return of EEG activity after resuscitation madea full recovery more often than those who had initial ECS [28]. The durationof burst suppression before return of continuous EEG activity also cor-related with prognosis for coma emergence. In these series, the longest inter-val after resuscitation for a patient who subsequently regained consciousnessto remain with ECS was 3.3 hours and 10.5 hours for burst suppression [36].


Real time measures of brain electrophysiology during cardiac arrest andthe immediate period after resuscitation, however, are studied best withanimal models. Models of cardiac arrest have been developed and validatedusing rats, monkeys, dogs, and cats during the past few decades. Improvingtechnology in signal processing and quantitative EEG during the past fewyears has led to new tools to track recovery after cardiac arrest and to definethe window of time for possible neuroprotective intervention. EEG record-ings during experimental cardiac arrest closely mirror the findings reportedby Jørgensen and colleagues in humans immediately after resuscitation. Indogs, ECS occurs within 10 to 30 seconds of cardiac arrest [39]. Severalhours of burst-suppression EEG also are reported in dogs after resuscitation[39–42]. The burst-suppression pattern evolved over 2 to 3 hours, with short-ening of the duration of suppression and increasing frequency, duration,and complexity of burst periods until activity fused to a continuous pattern[41]. In the dog model, the rate of EEG recovery correlated with degree andrate of recovery, as determined by pathologic evidence of ischemic neuronsand neurologic examination deficit scales [42]. Prognosis for neurologic re-covery was correlated directly with the latency of initial EEG activity, con-tinuous activity, and the duration of burst suppression [42–44]. Similar EEGrecovery patterns also are reported in monkey [45] and neonatal piglet mod-els of cardiac arrest [46].

A rat model of electrophysiologic monitoring after cardiac arrest hasbeen developed more recently. This model is validated by demonstrationof histologic evidence of ischemic cell death in the expected brain structuresafter cardiac arrest [47,48]. In this model, ECS is achieved within seconds ofcardiac arrest, followed by continued silence for a variable period afterROSC [49]. As in other models, a period of burst suppression precedes re-turn of continuous EEG activity [50,51]. As in the dog model, the durationof burst-suppression activity before return of continuous EEG is correlatedwith outcomes, using pathologic evidence and neurologic deficit scales toquantify injury [51]. Those rats destined for poor outcomes at 48 hoursdemonstrated longer intervals of suppression and less frequent bursts ateach time point during the recovery period, compared with animals withgood outcomes [51,52]. The interval to return of continuous EEG activityalso correlated with clinical signs of arousal, whereby rats who achievedEEG fusion earlier also demonstrated earlier purposeful hind limb move-ments [52]. The interval to return of continuous EEG also can be used totrack secondary injury. Those rats subjected to hyperglycemia after cardiacarrest, which is known to exacerbate ischemic injury, remained in burst sup-pression longer than normoglycemic controls [50].

Quantitative electroencphalogram

To increase the objectivity and sensitivity of EEG analysis for markers ofrecovery, several quantitative EEG techniques have been applied to the rat

98 KOENIG et al

model of cardiac arrest. Using cepstral distance, a frequency-based measureof difference between baseline EEG and various time points during reco-very, graded injury can be quantified accurately after cardiac arrest [52,53].Rats subjected to graded injury by 1-, 3-, 5-, and 7-minute cardiac arrest du-ration can be distinguished using cepstral distance measures to quantify thetime to EEG recovery. These quantitative EEG measures also correlatewith clinical and pathology extent of injury [53,54]. Power spectral analysisof EEG also has been undertaken in dogs [55]. In this model, dogs treatedwith neuroprotective free radical scavengers had earlier restitution of EEGpower in the high frequency range compared with untreated controls after re-suscitation from cardiac arrest [55]. More recently, quantitative EEG mea-surement of entropy shows promise in prognosticating degree of injury andmonitoring therapeutic responses to neuroprotective therapies [56,57]. Entropy,a mathematic measure of randomness or predictability of EEG patterns,may be an ideal technique for monitoring injury because all of the pathologicpatterns after cardiac arrestdseizure activity, burst suppression, alpha coma,and ECSdhave low entropy compared with continuous EEG patterns seen inanimals with good potential for recovery [57–59]. Recent data from the au-thors’ work show that treatment with mild hypothermiadwhich has demon-strated neuroprotective effects in rats after ischemic brain injurydcausesa more rapid increase in EEG entropy compared with that in controls [60].The potential to use quantitative EEG measures to quantify injury, track re-covery, test neuroprotective strategies, provide early prognostic information,and define the therapeutic time window currently is being explored in humansand animals.

Somatosensory evoked potentials

Short-latency median nerve SSEP studies also can provide a prognosis incomatose survivors of cardiac arrest. This electrophysiologic test examinesfor continuity of the sensory pathways triggered by an electrical stimulusapplied to the median nerve sensory distribution. Response to the sensorystimulus then is monitored through the peripheral nerve, nerve root, spinalcord, subcortical brain structures, and primary sensory cortex. The negativeslope of the scalp-evoked potential (P15-N20) is generated by the thalamo-cortical radiation, whereas the following positive wave (N20-P25) reflectscortex [61]. Damage to subcortical structures, therefore, is reflected in slow-ing of central conduction with reduction of the N20 amplitude, whereas ab-sence or distortion of the N20 peak represents cortical injury [61]. Bilateralloss of cortical N20 peaks, in the absence of known pre-existing cortical in-jury, is interpreted as widespread cortical injury after cardiac arrest. Severalseries demonstrate death or PVS in 100% of comatose cardiac arrest survi-vors who had bilaterally absent N20 peaks hours to days after resuscitation[7,61–64]. A recent meta-analysis [65] of reliable indicators of prognosisafter cardiac arrest finds that only absent cortical SSEP responses, pupillary


light reflexes on day 3, and motor response to pain on day 3 had 100% spec-ificity for poor outcomes. The prognostic importance of reduced N20 ampli-tudes and prolonged central conduction times are less clear. In one series, 8of 9 patients who had 50% to 75% decrement in N20 amplitude remained inPVS at 8 weeks [61]. In a different series, delayed or low amplitude N20 po-tentials were associated with poor outcome in 11 of 12 patients [7]. In ano-ther series, however, N20 amplitude and central conduction times weresimilar in patients who had good outcomes compared with those whodied or remained comatose [66].

SSEP has excellent specificity for poor outcome after cardiac arrest whenN20 peaks are absent, but the presence of normal N20 peaks has poor ac-curacy for predicting good outcomes. In an early series of patients whowere in coma after cardiac arrest, all 6 patients who had normal N20 poten-tials ultimately aroused from coma and had good neurologic recovery [61].In subsequent studies, however, normal SSEP responses are an unreliablepredictor of good outcome [7,67]. Only 50% of patients who had normalN20 potentials ultimately regained consciousness in one series [7] and25% in another [68]. To improve the ability of SSEP to predict more accu-rately good outcomes in comatose cardiac arrest survivors, Madl and co-workers evaluated long-latency N70 peaks in 66 patients [69]. In this series,N70 latency was less than 118 milliseconds in all patients who had good re-covery and was absent or greater than 118 milliseconds in patients who hadpoor outcome [69]. In a larger series of 305 comatose cardiac arrest survi-vors, N70 peaks were present in 98% of those who ultimately madea good neurologic recovery [70]. Long latency responses are believed to re-flect complex corticocortical pathways, important for intellectual functions.

Somatosensory evoked potentials patterns after resuscitation

Like EEG, recovery of SSEP also follows a predictable time course(Fig. 3). For this reason, assessment of SSEP early after resuscitation maylead to a falsely pessimistic interpretation. In a study of serial SSEP testingof comatose cardiac arrest survivors beginning 4 hours after resuscitation,there was significant improvement in SSEP responses in 76% of patientsduring a 24-hour period [71]. In particular, 2 patients who initially had ab-sent N20 responses regained them within 12 hours after ROSC. The speci-ficity for poor outcome with absent N20 responses increased from 0.43 at4 hours to 1.0 at 24 hours [71]. The N20 peak latency declined significantlyduring the first 24 hours in the majority of patients. The detection of N70peaks also increased from 7 of 25 patients at 4 hours to 14 of 25 patientsat 24 hours, although the N70 latency tended to improve during the sametime frame [71]. Only 3 of 7 patients who had eventual good outcomeshad an N70 latency less than the 130 milliseconds at 4 hours, whereas all7 were below this threshold at 24 hours [71]. At the same time, 3 patientswho ultimately had poor outcomes also regained an N70 latency less than

100 KOENIG et al

130 milliseconds after 24 hours, suggesting the specificity of this finding forgood outcomes decreases as the sensitivity rises. These findings contradictthe results of a similar study in 30 patients who underwent SSEP 30 to150 minutes after resuscitation from cardiac arrest [72]. In this series, nopatients who had absent N20 responses on the first examination regainedthose responses with subsequent SSEP testing [72].

More recent studies have tried to further differentiate surviving patientswho will be slightly or moderately disabled from those who would be

Fig. 3. Median nerve SSEP after cardiac arrest demonstrating absence of the cortical N20 peak

(arrow).


severely disabled or in PVS. Guerit and colleagues studied event-related po-tentials produced by a passive auditory oddball paradigm in 103 patients asa function of GCS [73]. In a subset of 83 patients examined within 4 days, abiphasic negative-positive evoked potential (EP) complex was identified tohave high specificity but low sensitivity for bad outcomes. This complex hasa high sensitivity but low specificity for badoutcomes. The cortex is particularlysensitive to anoxia, whereas the brainstem is relatively resistant, explainingwhy the negative predictive value of late cortical EP likely is better than thatof early subcortical EP. These assumptions are confirmed with studies usinglong- and short-latency auditory EP [74–76]. Because of these limitations,Guerit and colleagues adopted a three-modality approach dependent on twoindices: the index of global cortical functioning, graded from I to IV basedon visual evoked potentials and SSEP components after N20; and the in-dex of brainstem conduction, calculated from brainstem auditory evokedpotentials (BAEP) and SSEP components before and including N20 [76].These investigators demonstrate that only the index of global cortical func-tioning was helpful prognostically in anoxic coma. Grades I, II, III, and IVwere observed between the first and the third day of anoxic coma in 60%,40%, 15%, and 0% of patients, respectively, who had good outcomes [76].

More detailed examination of the loss of evoked responses during cardiacarrest and the sequence of restitution of potentials after ROSC can begleaned from animal models. In a cat model of SSEP during cardiac arrest,cortical responses were lost within 12 to 15 seconds of ischemia [77]. Theshorter latency potentials then were lost sequentially during the ensuing 2to 6 minutes. After resuscitation, shorter latency potentials recoveredmore rapidly and cortical evoked potentials regained 75% of their baselineamplitude within 6 hours [77]. In a dog model, all central components of theSSEP waveform were lost shortly after cardiac arrest [55]. The short latencycomponents of the SSEP waveform, reflecting activation of the dorsal col-umns, thalamocortical relay structures, and cortical N20 potentials, recov-ered within 1 hour after resuscitation. The long latency N70 potentialsrecovered only in animals treated with neuroprotective medications butnot in control animals after cardiac arrest.

In simultaneously recorded BAEP in dogs, the amplitude began to de-crease within 5 to 6 minutes after cardiac arrest [55]. Waves II through V,representing projections from the auditory nucleus through the inferior col-liculus, flattened at 7 to 8 minutes. Wave I, representing potentials from theauditory nerve, flattened at 9 to 10 minutes after cardiac arrest. AfterROSC, the BAEP gradually normalized in a caudorostral progression dur-ing the first hour, after which all latencies were normal [55]. This BAEP datais similar to a case report in a human subject resuscitated from cardiac arrestwhile undergoing BAEP testing [78]. This patient lost all BAEP responseswithin minutes of cardiac arrest. Waves I and II returned within 5 minutesof ROSC. Waves III through V also returned after 5 minutes but continuedto have prolonged latencies out to 50 minutes from ROSC [78]. Taken

102 KOENIG et al

together with the time course of EEG restitution, these data reflect the pro-gressive loss of electrical activity in a rostrocaudal pattern during cardiac ar-rest and the caudorostral restitution of electrical activity after resuscitation.Those regions most resistant to electrical dysfunction after cardiac arrest arethe first to return after resuscitation.

With microelectrode placement into the ventroposterolateral (VPL) nucle-us of the thalamus simultaneous to SSEP examination of the primary sensorycortex, the restitution of structures integral to generation of SSEP waveformscan be studied directly. In a rat model of graded injury from cardiac arrest,animals were subjected to 3, 5, and 7 minutes of arrest [54]. At the onset ofasphyxia, thalamic firing was characterized by an increasingly asynchronouspattern followed by loss of firing during a 15-second span [79]. The rhythmicspindle oscillations of the VPL neurons were replaced by increased tonic firingduring the first 45 seconds of asphyxia, followed by VPL silence within 1 min-ute [79]. In all three arrest times, restitution of the N20 potentials recovered to60% of baseline at 2 hours after resuscitation. The VPL response occurredsimultaneous with the recovery of N20 potentials in the short arrest timesbut was delayed relative to cortical responses in the 7-minute model [54,79].These data demonstrate thalamocortical dissociation after cardiac arrest,during the period in which the EEG demonstrates burst suppression [54,79].

Summary

Electrophysiologic testing continues to play an important role in injurystratification and prognostication in patients who are comatose after cardiacarrest. As discussed previously, however, the adage about treating whole pa-tients, not just the numbers, is relevant in this situation. EEG and SSEP canoffer high specificity for discerning poor prognosis as long as they are ap-plied to appropriate patient populations. As discussed previously, EEGand SSEP patterns change during the first hours to days after cardiac arrestand negative prognostic information should not be based solely on studiesperformed during the first 24 hours. Both electrophysiologic techniquesalso are susceptible to artifacts that may worsen the electrical patterns arti-ficially and suggest a falsely poor prognosis. EEG is suppressed by anes-thetic agents and hypothermia, both of which may produce ECS andburst suppression. Patients who experience respiratory arrest from a toxicingestion of narcotics or barbiturates, in particular, may present withhigh-grade EEG patterns initially. Many patients also receive anestheticmedications at the time of tracheal intubation, which may linger beyondtheir normal half-life in patients who have hepatic or renal insufficiencyor concurrent use of interacting medications. SSEP is much less susceptibleto sedative anesthetic agents, but hypothermia is demonstrated to prolongevoked potential latencies [60]. As therapeutic hypothermia becomes morecommon after cardiac arrest, the effect of temperature on electrophysiologictesting needs to be taken into account. The publications discussed


previously also emphasize the need to adjust the prognostic value of electro-physiologic tests to the pretest probability of meaningful neurologic recov-ery in individual patients. Clearly, grade I EEG patterns and normal N20potentials indicate a much better prognosis in patients who have a short du-ration of cardiac arrest, short duration of coma after resuscitation, andwhen the studies are performed within the first few days. In patients who re-main in coma days after resuscitation and lack appropriate brainstemreflexes, however, even the most normal appearing electrophysiologicpatterns do little to change the overall prognosis.

Aside from prognostication, electrophysiologic testing holds great prom-ise in defining the basic anatomy and physiology of coma emergence aftercardiac arrest. In addition, quantitative EEG and automated evoked poten-tials have the potential to render these tools less subjective and arcane andmore applicable for monitoring patients in the period during and immedi-ately after resuscitation. Quantitative EEG also has great potential asa tool to define the time window for neuroprotective intervention and themeans to track the response to such therapies in real time.

References

[1] Safar P. Cerebral resuscitation after cardiac arrest: a review. Circulation 1986;74:138–53.

[2] Edgren E, Hedstrand U, Nordin M, et al. Prediction of outcome after cardiac arrest. Crit

Care Med 1987;15:820–5.

[3] Hockaday JM, Potts F, Bonazzi A, et al. Electroencephalographic changes in acute cerebral

anoxia from cardiac or respiratory arrest. Electroencephalogr Clin Neurophysiol 1965;18:

575–86.

[4] Prior PF. EEG findings in dying and resuscitated adult patients. Electroencephalogr Clin

Neurophysiol 1969;27:333.

[5] Møller M, Holm B, Sindrup E, et al. Electroencephalographic prediction of anoxic brain

damage after resuscitation from cardiac arrest in patients with acute myocardial infarction.

Acta Med Scand 1978;203:31–7.

[6] Scollo-Lavizzari G, Bassetti C. Prognostic value of EEG in post-anoxic coma after cardiac

arrest. Eur Neurol 1987;26:161–70.

[7] Bassetti C, Bomio F,Mathis J, et al. Early prognosis in coma after cardiac arrest: a prospec-

tive clinical, electrophysiological, and biochemical study of 60 patients. J Neurol Neurosurg

Psychiatry 1996;61:610–5.

[8] Thomassen A, Sørensen K, Wernberg M. The prognostic value of EEG in coma survivors

after cardiac arrest. Acta Anaesthesiol Scand 1978;22:483–90.

[9] Binnie CD, Prior PF, Lloyd DSL, et al. Electroencephalographic prediction of fatal anoxic

brain damage after resuscitation from cardiac arrest. BMJ 1970;4:265–8.

[10] Synek VM. Validity of a revised EEG coma scale for predicting survival in anoxic encepha-

lopathy. Clin Exp Neurol 1989;20:119–27.

[11] SynekVM.Value of a revisedEEGcoma scale for prognosis after cerebral anoxia and diffuse

head injury. Clin Electroencephalogr 1990;21:25–30.

[12] Kaplan PW, Genoud D, Ho TW, et al. Etiology, neurologic correkations, and prognosis in

alpha coma. Clin Neurophysiol 1999;110:205–13.

[13] Young GB. The EEG in coma. J Clin Neurophysiol 2000;17:473–85.

[14] Beydoun A, Yen CE, Drury I. Variance of interburst intervals in burst suppression. Electro-

encephalogr Clin Neurophysiol 1991;79:435–9.

104 KOENIG et al

[15] SchaulN. The fundamental neuralmechanisms of electroencephalography. Electroencepha-

logr Clin Neurophysiol 1998;106:101–7.

[16] Niedermeyer E, ShermanDL, Geocadin RG, et al. The burst-suppression electroencephalo-

gram. Clin Electroencephalogr 1999;30:99–105.

[17] Ganji SS, HenryR, Furlow J. Diffuse theta activity and spindle-like bursts during coma after

cardiac arrest. Clin Electroencephalogr 1996;27:89–94.

[18] Zaret BS. Prognostic and neurophysiological implications of concurrent burst suppression

and alpha patterns in the EEG of post-anoxic coma. Electroencephalogr Clin Neurophysiol

1985;61:199–209.

[19] Pampiglione G, Harden A. Resuscitation after cardiac arrest. Lancet 1968;1:1261–4.

[20] BrennerRP, SchwartzmanRJ, Richey ET. Prognostic significance of episodic low amplitude

or relatively isoelectric EEG patterns. Dis Ner Syst 1975;36:582–7.

[21] Chen R, Bolton CF, Young GB. Prediction of outcome in patients with anoxic coma: a clin-

ical and electrophysiological study. Crit Care Med 1996;24:672–8.

[22] Cloche R, Desmonts JM,Hennetier G, et al. Morphology and evolution of the EEG in acute

cerebral anoxia. Electroencephalogr Clin Neurophysiol 1968;5:89.

[23] Reeves AL, Westmoreland BF, Klass DW. Clinical accompaniments of the burst-

suppression EEG pattern. J Clin Neurophysiol 1997;14:150–3.

[24] Sørensen K, Tomassen A, Wernberg M. Prognostic significance of alpha frequency EEG

rhythm in coma after cardiac arrest. J Neurol Neurosurg Psychiatry 1978;41:840–2.

[25] Young GB, McLachlan RS, Kreeft JH, et al. An electroencephalographic classification

system for coma. Can J Neurol Sci 1997;24:320–5.

[26] Young GB, Jordan KG, Doig GS. An assessment of nonconvulsive seizures in the intensive

care unit using continuous EEG monitoring: an investigation into the variables associated

with mortality. Neurology 1996;47:83–9.

[27] Chokroverty S. Alpha-like rhythms in electroencephalograms in coma after cardiac arrest.

Neurology 1975;25:655–63.

[28] Jørgensen EO,Holm S. The natural course of neurological recovery following cardiopulmo-

nary resuscitation. Resuscitation 1998;36:111–22.

[29] Losasso TJ, Muzzi DA, Meyer FB, et al. Electroencephalographic monitoring of cerebral

function during asystole and successful cardiopulmonary resuscitation. Anesth Analg

1992;75:1021–4.

[30] Moss J, Rockoff M. EEG monitoring during cardiac arrest and resuscitation. JAMA 1980;

244:2750–1.

[31] AzimN,Wang CY. The use of bispectral index during a cardiopulmonary arrest: a potential

predictor of cerebral perfusion. Anaesthesia 2004;59:610–2.

[32] Young WL, Ornstein E. Compressed spectral array EEG monitoring during cardiac arrest

and resuscitation. Anesthesiology 1985;62:535–8.

[33] CluteHL, LevyWJ. Electroencephalographic changes during brief cardiac arrest in humans.

Anesthesiology 1990;73:821–5.

[34] Pampiglione G. Electroencephalographic studies after cardiopulmonary resuscitation. Proc

R Soc Med 1962;55:653–7.

[35] Hellstrom-Westas L, Rosen I, Svinningsen NW. Predictive value of early continuous ampli-

tude integrated EEG recordings on outcome after severe birth asphyxia in full term infants.

Arch Dis Child 1995;72:F34–8.

[36] Prior PF. EEG monitoring and evoked potentials in brain ischaemia. Br J Anesth 1985;57:

63–81.

[37] Jørgensen EO, Malchow-Møller A. Natural history of global and critical brain ischaemia.

Part I: EEG and neurological signs during the first year after cardiopulmonary resuscitation

in patients subsequently regaining consciousness. Resuscitation 1981;9:133–53.

[38] Jørgensen EO, Malchow-Møller A. Natural history of global and critical brain ischaemia.

Part II: EEG and neurological in patients remaining unconscious after cardiopulmonary

resuscitation. Resuscitation 1981;9:155–74.


[39] YashonD,Wagner FC,Massopust LC, et al. Electrocorticographic limits of cerebral viabil-

ity during cardiac arrest and resuscitation. Am J Surg 1971;121:728–31.

[40] Lind B, Snyder J, Safar P. Total brain ischaemia in dogs: cerebral physiological and meta-

bolic changes after 15 minutes of circulatory arrest. Resuscitation 1975;4:97–113.

[41] Lin SR, Morris TW, Violante MR. Cerebral water content, blood flow, and EEG changes

after cardiac arrest in the dog. Invest Radiol 1977;12:325–32.

[42] Gurvitch AM,Mutuskina EA, Novoderzhkina IS. Quantitative evaluation of brain damage

in dogs resulting from circulatory arrest to the central nervous system or the whole animal:

2. Electroencephalographic evaluation during early recovery of the gravity and reversibility

of post-ischaemic cerebral damage. Resuscitation 1972;1:219–28.

[43] Brechner VL, Bethune RWM, Kavan EM, et al. The electroencephalographic effect of ar-

rested circulation in the normothermic human and dog. Anesth Analg Curr Res 1961;40:

1–14.

[44] YashonD,White RJ, TaslitzN, et al. Experimental cerebral circulatory arrest: effect on elec-

trocerebral potentials. J Neurosurg 1970;32:74–82.

[45] Myers RE, Yamaguchi SI. Nervous system effects of cardiac arrest inmonkeys. ArchNeurol

1977;34:65–74.

[46] Sherman DL, Brambrink AM, Ichord RN, et al. Quantitative EEG during early recovery

from hypoxic-ischemic injury in immature piglets: burst occurrence and duration. Clin Elec-

troencephalogr 1999;30:175–83.

[47] Kawai K, Nitecka L, Ruetzler CA, et al. Global cerebral ischemia associated with cardiac

arrest in the rat: I. Dynamics of early neuronal changes. J Cereb Blood Flow Metab 1992;

12:238–49.

[48] Radovsky A, Katz L, Ebmeyer U, et al. Ischemic neurons in rat brains after 6, 8, or

10 minutes of transient hypoxic ischemia. Toxicol Pathol 1997;25:500–5.

[49] Rosen I, SmithML,Rehncrona S. Quantitative EEGand evoked potentials after experimen-

tal brain ischemia in the rat; correlationwith cerebralmetabolism andblood flow. ProgBrain

Res 1994;62:175–82.

[50] Siemkowicz E, Hansen AJ. Brain extracellular ion composition and EEG activity following

10 minutes ischemia in normo- and hyperglycemic rats. Stroke 1981;12:236–40.

[51] GeocadinRG, ShermanDL,HansenHC, et al.Neurological recovery byEEGbursting after

resuscitation from cardiac arrest in rats. Resuscitation 2002;55:193–200.

[52] Luft AR, Buitrago MM, Paul JS, et al. Early restitution of electrocorticogram predicts sub-

sequent behavioral recovery from cardiac arrest. J Clin Neurophysiol 2002;19:540–6.

[53] Geocadin RG, Ghodadra R, Kimura T, et al. A novel quantitative EEG injury measure of

global cerebral ischemia. Clin Neurophysiol 2000;111:1779–87.

[54] GeocadinRG,Muthuswamy J, ShermanDL, et al. Early electrophysiological and histologic

changes after global cerebral ischemia in rats. Movement Dis 2000;15(Suppl 1):14–21.

[55] Cerchiari EL, Sclabassi RJ, Safar P, et al. Effects of combined superoxide dismutase and de-

feroxamine on recovery of brainstem auditory evoked potentials and EEG after asphyxial

cardiac arrest in dogs. Resuscitation 1990;19:25–40.

[56] Sherman DL, Atit MK, Geocadin RG, et al. Diagnostic instrumentation for neurological

injury. IEEE Instrum Meas Mag 2002;5:28–35.

[57] Thakor NV, Tong S. Advances in quantitative electroencephalogram analysis methods.

Annu Rev Biomed Eng 2004;6:453–95.

[58] Bezerianos A, Tong S, Thakor NV. Time-dependent entropy estimation of EEG rhythm

changes following brain ischemia. Ann Biomed Eng 2003;31:221–31.

[59] Al-NashashH, Paul J, ZiaiW, et al.Wavelet entropy for subband segmentation of EEGdur-

ing injury and recovery. Ann Biomed Eng 2003;31:653–8.

[60] Shin HC, Tong S, Yamashita S, et al., Quantitative EEG and effect of hypothermia on brain

recovery after cardiac arrest. IEEE Trans Biomed Eng, in press.

[61] Ahmed I. Use of somatosensory evoked responses in the prediction of outcome from coma.

Clin Electroencephalogr 1988;19:78–86.

106 KOENIG et al

[62] Kano T, Shimoda O, Morioka T, et al. Evaluation of the central nervous function in resus-

citated comatose patients by multilevel evoked potentials. Resuscitation 1992;23:235–48.

[63] BerekK, Lechleitner P, LuefG, et al. Early determination of neurological outcome after pre-

hospital cardiopulmonart resuscitation. Stroke 1995;26:543–9.

[64] Rothstein TL, Thomas EM, Sumi SM. Predicting outcome in hypoxic-ischemic coma. A

prospective clinical and electrophysiologic study. Electroencephalogr Clin Neurophysiol

1991;79:101–7.

[65] Zandbergen EGJ, de Haan RJ, Stoutenbeek CP, et al. Systematic review of early prediction

of poor outcome in anoxic-ischaemic coma. Lancet 1998;352:1808–12.

[66] Zingler VC, Krumm B, Bertsch T, et al. Early prediction of neurological outcome after

cardiopulmonary resuscitation: a multimodal approach combining neurobiochemical and

electrophysiological investigations may provide high prognostic certainty in patients after


[67] Madl C, Holzer M. Brain function after resuscitation from cardiac arrest. Curr Opin Crit

Care 2004;10:213–7.

[68] BrunkoE, Zegers de Beyl D. Prognostic value of early cortical somatosensory evoked poten-

tials after resuscitation from cardiac arrest. Electroencephalogr Clin Neurophysiol 1987;66:

15–24.

[69] Madl C, Kramer L, Domanovits H, et al. Improved outcome prediction of unconscious

cardiac arrest survivors with sensory evoked potentials compared with clinical assessment.

Crit Care Med 2000;28:721–6.

[70] Bauer E, Funk GC, Gendo A, et al. Electrophysiological assessment of the afferent sensory

pathway in cardiac arrest survivors. Eur J Clin Invest 2003;33:283–7.

[71] Gendo A, Kramer L, Hafner M, et al. Time-dependency of sensory evoked potentials in

comatose cardiac arrest survivors. Intensive Care Med 2001;27:1305–11.

[72] Nakabayashi M, Kurokawa A, Yamamoto Y. Immediate prediction of recovery of con-

sciousness after cardiac arrest. Intensive Care Med 2001;27:1210–4.

[73] Guerit J-M.Medical technology assessment. EEGand evoked potentials in the intensive care

unit. Neurophysiol Clin 1999;29:301–17.

[74] Rosenberg C, Wogensen K, Starr A. Auditory brain-stem and middle-and long-latency

evoked potentials in coma. Arch Neurol 1984;41:835–8.

[75] Fischer C, Bognar L, Turjman F, et al. Auditory early-andmiddle-latency evoked potentials

in patients with quadrigeminal plate tumors. Neurosurgery 1994;35:45–51.

[76] Guerit J-M, deTourtchaninoff M, Soveges L, et al. The prognostic value of three modality

evoked potentials in severe head injury. Electroencephalogr Clin Neurophysiol 1993;23:

209–26.

[77] Hossmann V, Hossmann KA. Return of neuronal functions after prolonged cardiac arrest.

Brain Res 1973;60:423–38.

[78] Rossini PM, Kila RW, HouseWJ, et al. Alteration of brain stem auditory evoked responses

following cardio-respiratory arrest and resuscitation. ElectroencephalogrClinNeurophysiol

1982;54:232–4.

[79] Muthaswamy J, Kimura T, Ding MC, et al. Vulnerability of the thalamic somatosensory

pathway after prolonged global hypoxic-ischemic injury. Neuroscience 2002;115:917–29.

Neurol Clin 24 (2006) 107–121

Neuroimaging and Serologic Markers ofNeurologic Injury after Cardiac Arrest

Madeleine C. Geraghty, MDa,Michel T. Torbey, MD, MPH, FAHAb,*

aDepartment of Neurology, University of Wisconsin Hospital and Clinics,

600 Highland Avenue, Madison, WI 53792, USAbDepartments of Neurology and Neurosurgery, Medical College of Wisconsin,

9200 West Wisconsin Avenue, Milwaukee, WI 53226, USA

For as long as patients have been comatose after cardiac arrest (CA),physicians have been searching for a reliable way to predict their outcomes.Although nearly half of all CA patients are resuscitated successfully, only10% to 20% of these patients make a good functional recovery [1–3].Most patients either die or survive with severe disability, often after a pro-longed stay in the ICU. The lengthy stays in ICUs have a significant as-sociated cost burden. A 1991 study determined that a hospital stay fora comatose survivor of CA cost up to $95,000 [4]. Multiple scales basedon physical and neurologic examinations were developed to assist in identi-fying patients who have no hope of recovery; but in the era of intensive caremanagement, these scales are often compromised by use of sedation andparalytic agents in ventilated patients and, at times, by fluctuations of theneurologic examination. Models based on electrophysiologic function arenot only operator-dependent but also often confounded by the electrical‘‘noise’’ present in the background of most ICUs. Many hospitals do nothave easily available electroencephalography or evoked-potential capabil-ities for assessment of postanoxic patients. Not all facilities have MRI avail-able within the hospital. Therefore, valid bedside tests with readily availabletools for the early prediction of poor outcome following postanoxic comaare necessary. A quantitative prognostic model of outcome that can be ap-plied early on to a patient who has been sedated following CA has becomealmost a Holy Grail for many intensive care practitioners, not only to assist


E-mail address: [email protected] (M.T. Torbey).




108 GERAGHTY & TORBEY

the grieving families in making difficult decisions, but also to ensure efficientuse of precious limited resources.

The ideal prognostic test is readily available, easily reproducible, and as-sociated with a high degree of specificity for poor outcome. The goal is notto define which patients may recover, but which patients have no likelihoodof meaningful neurologic recovery at all to justify early withdrawal of sup-port. It is acceptable to identify only a subset of the patients who will notrecover neurologically to avoid giving a falsely pessimistic prediction. Onemust be careful when prognosticating to clarify that it is purely prognosisof neurologic recovery; these patients remain critically ill and may die fromnonneurologic causes even if their neurologic outcome seems good.

Despite the many studies evaluating a wide range of prognostic variablesfollowing anoxic insult, it has been difficult to compare the variables directly.Studies have been limited by different causes of cardiopulmonary arrest,different outcomes scales, different time frames for prognostication, andsmall numbers of patients. This article presents an objective assessmentof serum and radiologic markers of recovery following CA and a proposedalgorithm for prognostication in CA patients.

Serologic markers

The most convenient markers of recovery are those that (1) can be ob-tained easily at the bedside and (2) measure enzymes in the bloodstreamor the cerebrospinal fluid (CSF). There have been contradictory reportsabout the development of elevated intracranial pressure following CA[5,6], suggesting that lumbar puncture in the setting of post-anoxic comamay increase the risk for herniation. None of these studies reported hernia-tion as a direct result of lumbar puncture. Nonetheless, elevated intracranialpressure remains at least a theoretic concern; and the search continues forprognostic peripheral markers in the bloodstream.

Table 1 summarizes theperformanceof various biochemicalmarkers in pre-dicting poor outcome. In these studies, poor outcome is variably defined asdeath, poor functional recovery, or both. If the predictive indiceswere not pro-vided in the articles, the authors calculated themde novo from the data specificfor global anoxia caused by CA, if the data were available for analysis.

Glucose

Glucose is the main source of energy for the brain. Several investigatorsreported a significant association between elevated blood glucose followingcardiac resuscitation and poor neurologic recovery [3,7–9,10]. The Long-streth Awakening Scale [3] went so far as to include admission blood glucoselevels as one of the four main variables in the model, where blood glucoselevels less than 300 mg/dL predicted awakening. It is unclear if postanoxichyperglycemia is causative or simply correlative with poor outcomes, be-cause the relationship between glucose levels and duration of resuscitation

109NEUROIMAGING AND SEROLOGIC MARKERS AFTER CARDIAC ARREST

is inconsistent across studies. If causative, the hypothesis is that hyperglyce-mia during cerebral ischemia drives anaerobic glycolysis and thus leads toincreased lactate production and an increase in intracellular acidosis.

In 1997, Mullner and colleagues [11] evaluated the role of postischemicblood glucose levels in a carefully defined subset of 145 patients in whomventricular fibrillatory arrest had been witnessed. They excluded patientswho had diabetes and those who received either insulin or glucose withinthe first 24 hours. They found that in CA survivors, high blood glucose lev-els over the first 24 hours after return of spontaneous circulation were inde-pendently associated with unfavorable functional neurologic recovery(146 G 39 mg/dL compared with 184 G 88 mg/dL). It is unfortunate, butbecause of overlap between the values for the groups, no ‘‘cutoff’’ point wasestablished by this study; and emphasis was placed instead on the trend to-ward poorer outcomes with higher glucose levels.

If attempting to use glucose as a prognostic indicator, the practitionermust take into account multiple variables, such as the presence of preexist-ing diabetes mellitus, the use of glucose or insulin during the periarrest pe-riod, and the total dosage of epinephrine used during the resuscitation [7].There does not appear to be support for using glucose as a primary variablein predicting outcome following CA. However, the admission glucose andmedian glucose over the first 24 hours can be used as an adjunctive measure-ment supporting the gestalt of the individual patient’s prognosis.

Lactate

The oxygen deficiency inherent in anoxic injury leads to anaerobic glycol-ysis and therefore to lactate overproduction. At the same time, the profound

Table 1

Predictive indices of biomarkers by timeframe

Marker

Cutoff values

(ng/mL)

Time Specificity

(%)

Sensitivity

(%)

PPV

(%)

NPV

(%)

NSE

Serum 13.3–20.0 24 h 89–100 51–59 86–100 65–70

8.8–25.0 24–48 h 100 59–76 100 10–80

16.4–100.0 72 h 100 70–100 100 72–100

CSF 24.0 24 h 100 74 100 89

50.0 48 h 83 89 96 63

S-100

Serum 0.2–0.7 24 h 80–96 55–100 71–95 63–100

0.2 48 h 70–100 79–100 75–100 89–100

CSF 6.0 48 h 60 93 93 63

CK-BB

CSF 17.0 24 h 50–98 40–52 66–93 25–77

Abbreviations: CK-BB, brain-type isoenzyme of creatine kinase; CSF, cerebrospinal fluid;

NPV, negative predictive value; NSE, neuron specific enolase; PPV, positive predictive value;

S-100, an astroglial protein established as a marker for cerebral injury.


ischemic state may impair liver function, leading to reduced lactate elimina-tion [12]. These factors suggest that lactate levels may be a useful marker forsevere anoxic injury [13]. Measurements of serum lactate levels on admissionhave not been particularly accurate at predicting prognosis in postanoxic pa-tients, requiring excessively high levels to achieve 100% specificity for poorneurologic outcome [14,11]. Serial lactate measurements fared slightly better,with levels higher than 2 mmol/L after 48 hours correlating well with mortal-ity or severe neurologic disability. This elevation, although statistically sig-nificant, was low in absolute numbers; and the specificity was less thanideal at 86%. A series of five patients in anoxic coma following CA suggestedthat perhaps it is not the absolute elevation of lactate that predicts poor out-come, but the failure of the lactate levels to decline [15]. One concern withusing serum lactate that has not been addressed by the prior studies is con-trolling for the use of lactated intravenous fluid during the resuscitation. El-evated lactate levels in the CSF have been reported to provide more accurateinformation regarding the extent of brain damage caused by CA. A smallcase series of seven postanoxic patients revealed a mean CSF lactate of2.5G 0.5 mmol/L over the first 28 hours in patients who regained conscious-ness versus 3.8 G 0.9 mmol/L in patients who remained comatose and died.No cutoff value was determined [16]. Much like the use of glucose, serum andCSF levels of lactate can be used as supporting data; but neither should beused as the primary method of prognostication in an individual patient.

S-100

S-100, an astroglial protein, has been established as a marker for cerebralinjury. It may also be released from cardiac tissues [17] and from tumors,such as glioma, schwannoma, neuroblastoma, and melanoma [18–20]. In-creased levels of serum S-100 have been correlated with cerebral injury[21–23]. Global cerebral circulatory arrest causes diffuse brain edema andselective neuronal death in vulnerable areas of the brain, and extended an-oxia time leads to infarctions in cortical and subcortical regions [24,25]. Theinitial elevation of S-100 may represent blood–brain barrier breakdown inconjunction with astroglial damage and early brain edema.

Rosen and colleagues [26] found that comatose CA patients who had se-rum S-100 levels persistently greater than 0.2 mcg/L over 48 hours diedwithin 14 days of CA. They controlled for the possible release of cardiacS-100 by comparing levels to those of patients who experienced myocardialinfarction but did not undergo resuscitation. Martens and colleagues [27]achieved the same results with a post hoc cutoff value of 0.7 mcg/L over48 hours; and they found that this cutoff value was useful for predictionat 24 hours as well, although the specificity was not as good comparedthe 48-hour S-100 levels. The cutoff value of 0.7 mcg/L has been supportedrecently by Hachimi-Idrissi and colleagues [28] with good specificity, evenusing the admission levels of S-100 protein.


Creatine kinase brain-type isoenzyme

The brain-type isoenzyme of creatine kinase (CK-BB) is primarily locatedin neurons and astrocytes, and it comprises 95% of the total creatine kinaseactivity of the brain [29]. Normal serum contains little, if any, CK-BB. Lev-els of CK-BB have been noted to be elevated following global cerebral ische-mic injury [30]. There have been several difficulties noted in the use of serumCK-BB as a prognostic indicator. Many patients suffering CA will alsodevelop myocardial infarctions with resultant release of the CK-MB isoen-zyme into the serum. Many of the available assays have some cross-reactivitybetween the MB and BB isoenzymes, making a serum test less accurate. Inaddition, the CK-BB fraction peaks rapidly in the serum and is rapidly in-activated, with individual variation as to the speed of inactivation [31].Neither the optimal time for testing nor the need for serial testing have yetbeen determined.

Roine and colleagues [32] compared the levels of CK-BB in the serumand in the CFS of survivors of out-of-hospital ventricular fibrillatory arrestat approximately 24 hours postinjury. Although they found that patientswho remained comatose had higher levels of CK-BB in the serum thandid patients who regained consciousness, there was significant overlapbetween the values; and no cutoff value was established. CFS levels ofCK-BB at 24 hours were a better predictor of outcome. Using a cutoff valueof 17 ng/mL, established post hoc, the CSF CK-BB measurement had a spec-ificity of 98% and a sensitivity of 52%. A smaller study by Clemmensen andcolleagues [33] was not able to support using CSF CK-BB because of poorspecificity (54%), but this study was underpowered to detect large differen-ces. They did, however, notice that patients who had low levels of CK-BB intheir CSF were more likely to make a good neurologic recovery.

The serum CK-BB and the CSF CK-BB levels did not correlate well witheach other. These results have been duplicated, with CSF levels of CK-BBmeasured between 28 and 76 hours following injury being useful. Again, se-rum CK-BB was not found to have prognostic value [16,34,35].

As with previous biomarker studies, it seems that elevated CK-BB levelsin the CSF at 24 hours postinjury might be useful as adjunctive prognosticindicators for poor neurologic outcome but would not be recommended asthe primary evidence on which to base a decision to withdraw care. A pro-spective predictive study with preestablished cutoff values would be re-quired to elucidate further the usefulness of CSF CK-BB isoenzymesfollowing CA.

Neuron-specific enolase

Neuron-specific enolase (NSE) is a glycolytic enzyme present almost ex-clusively in neurons and neuroendocrine cells. Of all the biochemicalmarkers available for prognosis following CA, NSE seems to be the mostsubstantiated. NSE levels in serum and CSF have been investigated several


times for possible use in estimating neuronal injury and predicting the clin-ical outcome of patients who have anoxic encephalopathy.

In 1989, Roine and colleagues [32] published the first trial investigatingthe levels of NSE in serum and CSF of survivors of out-of-hospital ventric-ular fibrillatory arrest at approximately 24 hours postinjury. Using a cutoffvalue of 24 ng/mLdestablished post hocdthe CSF NSE measurement hada specificity of 100% and a sensitivity of 74% for predicting death. Multiplestudies have followed evaluating the prognostic value of serum NSE, CSFNSE, and serial NSE measurements [27,35–41]. Several trials thus far havenoted an association between decreasing serum NSE levels and good neuro-logic outcome [39,41,42], so perhaps the rate of decrease may be used toprognosticate in addition to the absolute level of NSE. Although one study[43] failed to show a correlation between serum NSE and neurologic out-come, the remaining studies have shown generally good predictive powerusing serial serum NSE measurements over the first 72 hours.

One of the provisos with the use of NSE levels is that platelets and redblood cells contain small amounts of this enzyme, so hemolysis can increasethe total level of NSE [44]. Accordingly, CSF samples from traumatic lum-bar puncture and hemolysed blood samples should not be used for progno-sis. Serum NSE is easier to collect than CSF and allows for serial sampling.The coincidental occurrences of other conditions known to elevate NSE lev-els, such as neuroblastoma and small-cell lung cancer, are rare enough notto interfere with prognostication. However, stroke can and does occur in thesetting of CA and elevates the serum and CSF levels of NSE [45,46]. It is notknown whether the elevation of NSE secondary to stroke should be includedor excluded from the mortality prediction. Ultimately, a cutoff value shouldbe set prospectively and followed to determine specificity, sensitivity, andpositive predictive value before the use of NSE as a primary prognosticindicator.

Radiologic markers

Gray matter (GM), with emphasis on the hippocampus, caudate/puta-men, thalamus, large cell layers of the neocortex, and Purkinje cells of thecerebellar cortex, has long been considered to be selectively vulnerable tohypoxic injury. This susceptibility does not imply that white matter (WM)is invulnerable to hypoxia, and the different reactions of the two tissue typesto hypoxic injury hold potential in the quest for reliable objective measure-ments for prognosis following CA. White matter injury, a delayed leukoen-cephalopathy, has been described weeks to months following global anoxicinjury [47,48]. CT and conventional MRI imaging in the acute period fol-lowing diffuse cerebral anoxia is often read as normal or with subtle find-ings only. The imaging findings of diffuse anoxia include diffuse cerebraledema, obscuration of gray–white matter borders, so-called ‘‘watershed’’ in-farctions, and selective neuronal necrosis affecting the deep gray nuclei and


of the cortical layers III, IV, and V [25,49–51]. The advent of diffusion-weighted (DWI) MRI enhanced the ability to detect these early changesof cerebral anoxia.

MRI

Although earlier studies focused primarily on GM injury following cere-bral ischemia, DWI opened the door for evaluating WM injury also. Arbe-laez and colleagues [52] evaluated DWI findings in ten patients who hadglobal cerebral anoxia. Only 6 of 10 patients had primary CA. Using low-bvalue (B ¼ 30 s/mm2) and high-b value (B ¼ 1100 s/mm2) diffusion imaging,investigators were able to find abnormal signal in the cerebellum, basal gan-glia, and cortex in less than 24 hours and WM changes in the late subacuteperiod of 14 to 20 days following injury. All of the patients demonstratingDWI changes had a poor neurologic outcome (death or vegetative state).This study demonstrated that, although GM has traditionally been consid-ered as more vulnerable to hypoxic injury, WM is also susceptible to injuryearlier than previously postulated [47] and that DWI holds potential forprognostication.

Chalela and colleagues [53] presented a case series in which MRIs withT2-weighted imaging and DWI were obtained on seven comatose patientsbetween one and six days following severe anoxic injury. Only two of theseven patients had primary CA. Restricted diffusion was prominentthroughout the WM in the periventricular region, corpus callosum, and in-ternal capsule. These findings were corroborated with decreased apparentdiffusion coefficient mapping. GM damage was demonstrated more readilyon T2-weighted imaging than by DWI. This study demonstrated that mye-linopathy can occur early following anoxic injury and holds potential forevaluation as a prognostic indicator.

Similarly, Wijdicks and colleagues [54] evaluated the MRI features in tencomatose male survivors of CA. All ten of the patients had intact brainstemreflexes on physical examination. Five patients had electroencephalography,none of which demonstrated alpha-coma or burst-suppression patterns. Sixpatients received median somatosensory evoked potentials, which were ei-ther normal (four patients) or indeterminate (two patients). The outcomesof these patients ran the gamut from good neurologic function to death.MR imaging was obtained between 1 and 15 days following the anoxicinsult. Eight of the patients showed diffuse GM signal abnormalities onfluid-attenuated inversion recovery (FLAIR) sequences and DWI. Signalabnormalities were specifically noted in cerebral and cerebellar cortices, cau-date nuclei, putamen, globus pallidi, and thalami. Eight of the ten patientshad diffuse abnormalities in these regions on FLAIR and DWI, and thesepatients either died or recovered to a vegetative state. The two patientswho made good recoveries had minimal (no cortical involvement) or no ab-normalities seen on MRI. Although the numbers are small, this study


supported earlier findings that early MRI changes may be reliable predictorsof poor outcome.

Fig. 1 demonstrates some of the MRI features of global cerebral anoxia.This 55-year-old man presented in CA with a duration of more than 20 mi-nutes. He was comatose, sedated, and intubated following the arrest witha Glasgow Coma Scale score of 3. DWI and corresponding apparent diffu-sion coefficient map show diffuse laminar restriction of diffusion. These find-ings are subtle but definite using low b-value diffusion.

The advantage of the increased sensitivity of MRI is offset by patient in-stability, need for monitoring, and cost. Many hospitals have limited MRresources; and the MR scanner may not be located within the hospital build-ing, necessitating a ‘‘road trip’’ for a critically ill patient. Transportation re-quires monitoring by critical-care-trained staff and often also may requirean anesthesiologist. Many critically ill patients are not stable enough tobe moved, and there is some evidence that the mere act of moving someof these patients may be subtly detrimental [54–56]. The logistical difficultiesinherent in obtaining an MRI on a critically ill patient currently limit theapplication of this technology in obtaining outcomes data.

Magnetic resonance spectroscopy

Magnetic resonance spectroscopy (MRS) was developed in 1988. Earlystudies of cerebral ischemia demonstrated an increase in the lactate spec-trum as a consequence of anaerobic glycolysis, cellular necrosis, and hypo-perfusion, with accompanying disturbances of lactate removal [57]. Berekand colleagues [24] evaluated the use of MRS to detect elevated cerebral

Fig. 1. (A) The axial diffusion-weighted image (low b-value, b ¼ 30 s/mm2) shows diffusely re-

stricted diffusion in a laminar fashion along the cortex, most prominent along the bilateral pa-

rieto–occipital cortices (large arrows). More subtle regions of restricted diffusion are noted at

the right insular cortex and bilateral frontal cortices (small arrows). (B) The apparent diffusion

coefficient map shows subtle focal areas of decreased signal in locations corresponding to the

restricted diffusion image on the left.


lactate resonances as a prognostic indicator in 30 patients who had sufferedout-of-hospital CA. The presence of a lactate peak on MRS correlated withdeath or severe disability at one month. Patients who lacked a lactate peakeither had good neurologic recovery or else died from extracranial causes,typically cardiac. Further quantification beyond presence or absence of a lac-tate peak was not attempted.

Positron emission tomography

Positron emission and single-photon emission tomography may be usefulin the acute phase of postanoxic injury, but these tests are not widely avail-able and are difficult to obtain in patients who are critically ill [58]. Onestudy has shown a nearly 50% reduction in cerebral glucose metabolism fol-lowing CA, but was unable to identify any specific patterns of metabolicchange that would be useful for prognostication of outcome [59,60].

Computerized tomography

The most commonly ordered and most widely available imaging modalityfor comatose CA patients is the CT scan. A normal CT scan of the brainshows a clear difference between the WM with its high lipid content andthe GM with its high water content [61]. Following severe global anoxia,there is a loss of distinction between the GM and WM in the brain asseen on head CT [62]. This difference is not reliably noted on visual assess-ment. However, measuring the Hounsfield unit (HU) density may allowa reliable quantifiable comparison between GM and WM. HUs are mea-surements of x-ray attenuation used in CT scanning where each pixel is as-signed a value on a scale on which air is �1000 HU, water is 0 HU, and boneis þ1000 HU [63]. Torbey and colleagues [64] compared the HU density ofGM of the caudate nucleus and the HU of the WM of the posterior limb ofthe internal capsule (Fig. 2). On noncontrast CT scans performed within 48hours of CA, the GM/WM ratios were significantly lower than those of con-trol patients. This loss of differentiation was due more to a loss of GM den-sity than to an increase in WM density. Using a receiver operatingcharacteristic curve analysis, a GM/WM ratio of less than 1.18 was 100%predictive of death, whereas the survival rate was 46% among patientswho had a GM/WM ratio of greater than or equal to 1.18. Of note, thenoncontrast CT scans were obtained primarily between 24 and 48 hours.It is unclear if CT scans obtained in less than 24 hours might miss someabnormal findings.

Summary

The current serum or radiologic markers have their pros and cons. Clearly,what is needed to prognosticate accurately following CA is a multimodal


scale or algorithm that incorporates multiple parameters, specifically serummarkers, radiologic markers, and the neurologic examination. Zingler andcolleagues [65] found that combining NSE and S-100 could markedly in-crease the sensitivity of prediction from 75% to 87.5% on day 3 without sac-rificing specificity. Continuing the multimodal approach, the Brain ArrestNeurologic Outcome Scale (BrANOS) was developed in 2004 as a way of en-compassing clinical and radiologic markers of brain injury into a predictivemodel in the setting of global anoxic injury following CA. The scoring sys-tem is outlined in Table 2. In 32 patients who had a witnessed CA, survivorshad a significantly lower BrANOS score (8 G 2 points) compared with non-survivors (13 G 1 points). The scale predicted death within 2 weeks with100% positive predictive value (PPV) and 100% specificity for a score greater

Table 2

Brain arrest neurologic outcome scale (BrANOS)

Scoring system Points

Duration of arrest (DAR)

0–5 min 1

6–15 min 2

O15 min 3

Best reverse Glasgow Coma Scale score (GCS)

(15-GCS) in 24 h 0–12

Hounsfield unit ratio (HUR)

!1.18 1

R1.18 0

Total 1–16

Fig. 2. Hounsfield unit density of a 10 mm2 region of (1) caudate nucleus and Hounsfield unit

density of a 10 mm2 region of the posterior limb of the (2) internal capsule. The measurements

should be taken from the same axial head CT slice at the basal ganglia level, defined as the

level in which the caudate nucleus, internal capsule, third ventricles, and sylvian fissures are

all visualized.


than or equal to 14. Using a perhaps more realistic measurement of outcome,the scale also predicted the combined outcome of death and severe disability(Glasgow Outcome Score ¼ 3) with 100% PPV and specificity for a scoregreater than or equal to 10 [66]. This scale provides another method to predictoutcome quantitatively, although it will need to be validated prospectively be-fore general application. The authors further recommend a prospective eval-uation of a ‘‘cardiac arrest prognosis protocol,’’ which would include serialmeasurements of serum NSE and serum S-100 protein over 72 hours in addi-tion to incorporating imaging between 24 and 48 hours. The authors believethat the specificity and sensitivity of serumNSE are highest over a 72-hour pe-riod. Although the S-100 protein levels may have already peaked by day 3,testing over 72 hours allows capture of any late-peaking levels. Althoughthe other biomarkers, such as lactate, glucose, and CSF CK-BB, also allowfor some prognostication, they either lack ease of use or enough specificityto make reliable markers at this time. For such a prospective evaluation, theauthors recommend the protocol represented in Table 3. Until these parame-ters have been evaluated prospectively, the treating physicianmay alsowish torely on the ‘‘standard’’ clinical and electrophysiologic parameters.

Future directions

The ability to determine a reliable prognosis in CA survivors continues toimprove as more advanced methods of data collection are developed. Fiber-optic jugular venous oximetry is a newer technique for assessing globalcerebral oxygen use. Monitoring the jugular venous oxygen saturation pro-vides an early diagnosis of global ischemia by venous desaturation [67]. Thecatheter can be placed rapidly at the bedside, and there may be potential todetect ultra-early cerebral anoxic injury even before changes can be detectedon imaging or with serum markers.

Table 3

Cardiac arrest prognosis protocol

Tests/Scores to obtain Poor prognostic indicators

Admission Glasgow Coma Scale score GCS ! 8

Serial NSE levels on days 0–3 Peak value O 35 ng/mL

Serial S-100 levels on days 0–3 Peak value O 0.2 ng/mL

Admission glucose level Initial value O 300 mg/dL

Head CT at 24–48 h GM/WM ratio ! 1.18

BrANOS score R10

Head MRI at 72 h Restricted diffusion in the cerebral and cerebellar

cortices, caudate nuclei, putamen, globus

pallidi, and/or thalami.

Abbreviations: BrANOS, Brain Arrest Neurologic Outcome Scale; GM, gray matter; GCS,

Glasgow Coma Scale; NSE, neuron-specific enolase; S-100, an astroglial protein established as

a marker for cerebral injury; WM, white matter.


The recent development of cerebral microdialysis catheters may be ableto collectdalbeit in an invasive fashiondfar more detailed quantitative in-formation on neurochemical markers of cerebral metabolism in the immedi-ate postresuscitation period [55,68]. Animal studies of microdialysismonitoring in the setting of induced hypothermic circulatory arrest havedemonstrated high cerebral lactate/glucose ratios are predictive of postoper-ative death [69]. The reevaluation of glucose and lactate as biochemicalmarkers for prognosis in the setting of cerebral microdialysis will be mostinteresting in the upcoming years. Used in conjunction with jugular bulboxygen monitoring in the early phase following global anoxic injury, thesetechniques may be able to guide therapy (increasing cerebral oxygen deliv-ery) as well as predict outcomes.

Portable CT scanners are now available, and their cost is roughly half thatof a ‘‘fixed’’ helical CT scanner [70]. The patient is scanned without strippingan ICU of skilled caregivers to accompany a patient during transport. Pa-tients who are considered too unstable from either a neurologic or cardiologicstandpoint will benefit as these portable modules become more widely usedand as more ICUs install dedicated CT scanners directly within the units.

As these techniques are being evaluated more closely and as imaging mo-dalities increase in sensitivity and portability, physicians will continue to as-sist families by providing some guidance as to which patients have no chanceof meaningful recovery. In carefully applied settings, serum NSE levels,S-100 protein levels, and multimodal scales such as BrANOS may help inthe making of these decisions.

References

[1] Cobbe SM,DalzielK, Ford I, et al. Survival of 1476 patients initially resuscitated fromout of

hospital cardiac arrest. BMJ 1996;312(7047):1633–7.


JAMA 1985;253(10):1420–6.

[3] Longstreth WT Jr, Diehr P, Inui TS. Prediction of awakening after out-of-hospital cardiac

arrest. N Engl J Med 1983;308(23):1378–82.

[4] Gray WA, Capone RJ, Most AS. Unsuccessful emergency medical resuscitationdare con-

tinued efforts in the emergency department justified? N Engl J Med 1991;325(20):1393–8.

[5] Gueugniaud PY, Garcia-Darennes F, Gaussorgues P, et al. Prognostic significance of early

intracranial and cerebral perfusion pressures in post-cardiac arrest anoxic coma. Intensive

Care Med 1991;17(7):392–8.

[6] Gustafson I, Edgren E, Hulting J. Brain-oriented intensive care after resuscitation from car-

diac arrest. Resuscitation 1992;24(3):245–61.

[7] Calle PA, Buylaert WA, Vanhaute OA. Glycemia in the post-resuscitation period. The

Cerebral Resuscitation Study Group. Resuscitation 1989;17(Suppl):S181–8 [discussion

S199–206].

[8] Fang JF, Chen RJ, Lin BC, et al. Prognosis in presumptive hypoxic-ischemic coma in non-

neurologic trauma. Journal of Trauma-Injury Infection & Critical Care 1999;47(6):1122–5.

[9] Longstreth WT Jr, Diehr P, Cobb LA, et al. Neurologic outcome and blood glucose levels

during out-of-hospital cardiopulmonary resuscitation. Neurology 1986;36(9):1186–91.


[10] Longstreth WT Jr, Inui TS, Cobb LA, et al. Neurologic recovery after out-of-hospital car-

diac arrest. Ann Intern Med 1983;98(5 Pt 1):588–92.

[11] Mullner M, Sterz F, Domanovits H, et al. The association between blood lactate concentra-

tion on admission, duration of cardiac arrest, and functional neurological recovery in pa-

tients resuscitated from ventricular fibrillation. Intensive Care Med 1997;23(11):1138–43.

[12] AbramsonD, ScaleaTM,HitchcockR, et al. Lactate clearance and survival following injury.

Journal of Trauma-Injury Infection & Critical Care 1993;35(4):584–8 [discussion 588–9].

[13] Tuchschmidt JA, Mecher CE. Predictors of outcome from critical illness, shock and cardio-

pulmonary resuscitation. Crit Care Clin 1994;10(1):179–95.

[14] Edgren E, Hedstrand U, Nordin M, et al. Prediction of outcome after cardiac arrest.

Crit Care Med 1987;15(9):820–5.

[15] Vincent JL, Dufaye P, Berre J, et al. Serial lactate determinations during circulatory shock.

Crit Care Med 1983;11(6):449–51.

[16] Karkela J, PasanenM,Kaukinen S, et al. Evaluation of hypoxic brain injurywith spinal fluid

enzymes, lactate, and pyruvate. Crit Care Med 1992;20(3):378–86.

[17] Anderson RE, Hansson LO, Nilsson O, et al. Increase in serum S100A1-B and S100BB

during cardiac surgery arises from extracerebral sources. Ann Thorac Surg 2001;71(5):

1512–7.

[18] Isobe T, Takahashi K, OkuyamaT. S100a0 (alpha alpha) protein is present in neurons of the

central and peripheral nervous system. J Neurochem 1984;43(5):1494–6.

[19] Takahashi K, Isobe T, Ohtsuki Y, et al. Immunohistochemical localization and distribution

of S-100 proteins in the human lymphoreticular system. Am J Pathol 1984;116(3):497–503.

[20] Yambe H, Ishii Y, Tabuchi K, et al. Detection of S-100 protein in melanocytic and neuro-

genic cutaneous tumors. Arch Dermatol Res 1984;276(3):160–4.

[21] Astudillo R, Van der Linden J, Radegran K, et al. Elevated serum levels of S-100 after deep

hypothermic arrest correlate with duration of circulatory arrest. Eur J Cardiothorac Surg

1996;10(12):1107–12 [discussion 1113].

[22] Johnsson P, Lundqvist C, Lindgren A, et al. Cerebral complications after cardiac surgery as-

sessed by S-100 and NSE levels in blood. J Cardiothorac Vasc Anesth 1995;9(6):694–9.

[23] Kim JS, Yoon SS, KimYH, et al. Serial measurement of interleukin-6, transforming growth

factor-beta, and S-100 protein in patients with acute stroke. Stroke 1996;27(9):1553–7.

[24] Berek K, Lechleitner P, Luef G, et al. Early determination of neurological outcome after

prehospital cardiopulmonary resuscitation. Stroke 1995;26(4):543–9.

[25] Roine RO,RaininkoR, Erkinjuntti T, et al.Magnetic resonance imaging findings associated

with cardiac arrest. Stroke 1993;24(7):1005–14.

[26] Rosen H, Rosengren L, Herlitz J, et al. Increased serum levels of the S-100 protein are asso-

ciated with hypoxic brain damage after cardiac arrest. Stroke 1998;29(2):473–7.

[27] Martens P. Serum neuron-specific enolase as a prognostic marker for irreversible brain dam-

age in comatose cardiac arrest survivors. Acad Emerg Med 1996;3(2):126–31.

[28] Hachimi-Idrissi S, Van der AuweraM, Schiettecatte J, et al. S-100 protein as early predictor

of regaining consciousness after out of hospital cardiac arrest. Resuscitation 2002;53(3):

251–7.

[29] Chandler WL, Clayson KJ, Longstreth WT Jr, et al. Creatine kinase isoenzymes in human

cerebrospinal fluid and brain. Clin Chem 1984;30(11):1804–6.

[30] Longstreth WT Jr, Clayson KJ, Chandler WL, et al. Cerebrospinal fluid creatine kinase

activity and neurologic recovery after cardiac arrest. Neurology 1984;34(6):834–7.

[31] Goe MR, Massey TH. Assessment of neurologic damage: creatine kinase-BB assay after

cardiac arrest. Heart Lung 1988;17(3):247–53.

[32] Roine RO, Somer H, Kaste M, et al. Neurological outcome after out-of-hospital cardiac

arrest. Prediction by cerebrospinal fluid enzyme analysis. Arch Neurol 1989;46(7):753–6.

[33] Clemmensen P, Strandgaard S, Rasmussen S, et al. Cerebrospinal fluid creatine kinase

isoenzyme BB levels do not predict the clinical outcome in patients unconscious following

cardiac resuscitation. Clin Cardiol 1987;10(4):235–6.


[34] Bodvarsson A, Franzson L, Briem H. Creatine kinase isoenzyme BB in the cerebrospinal

fluid of patients with acute neurological diseases. J Intern Med 1990;227(1):5–9.

[35] Karkela J, Bock E, Kaukinen S. CSF and serum brain-specific creatine kinase isoenzyme

(CK-BB), neuron-specific enolase (NSE) and neural cell adhesion molecule (NCAM) as

prognostic markers for hypoxic brain injury after cardiac arrest in man. J Neurol Sci

1993;116(1):100–9.

[36] Dauberschmidt R, Zinsmeyer J, Mrochen H, et al. Changes of neuron-specific enolase con-

centration in plasma after cardiac arrest and resuscitation. Mol Chem Neuropathol 1991;

14(3):237–45.

[37] Fogel W, Krieger D, VeithM, et al. Serum neuron-specific enolase as early predictor of out-

come after cardiac arrest. Crit Care Med 1997;25(7):1133–8.

[38] SchaarschmidtH, PrangeHW,ReiberH.Neuron-specific enolase concentrations in blood as

a prognostic parameter in cerebrovascular diseases. Stroke 1994;25(3):558–65.

[39] Schoerkhuber W, Kittler H, Sterz F, et al. Time course of serum neuron-specific enolase.

A predictor of neurological outcome in patients resuscitated from cardiac arrest. Stroke

1999;30(8):1598–603.

[40] Stelzl T, von BoseMJ, Hogl B, et al. A comparison of the prognostic value of neuron-specific

enolase serum levels and somatosensory evoked potentials in 13 reanimated patients. Eur J

Emerg Med 1995;2(1):24–7.

[41] Tiainen M, Roine RO, Pettila V, et al. Serum neuron-specific enolase and S-100B protein in

cardiac arrest patients treated with hypothermia. Stroke 2003;34(12):2881–6.

[42] Rosen H, Sunnerhagen KS, Herlitz J, et al. Serum levels of the brain-derived proteins S-100

and NSE predict long-term outcome after cardiac arrest. Resuscitation 2001;49(2):183–91.

[43] Lima JE, TakayanaguiOM,Garcia LV, et al.Use of neuron-specific enolase for assessing the

severity and outcome in patients with neurological disorders. Braz J Med Biol Res 2004;

37(1):19–26.

[44] Marangos PJ, Campbell IC, Schmechel DE, et al. Blood platelets contain a neuron-specific

enolase subunit. J Neurochem 1980;34(5):1254–8.

[45] Butterworth RJ, Wassif WS, Sherwood RA, et al. Serum neuron-specific enolase, carnosi-

nase, and their ratio in acute stroke. An enzymatic test for predicting outcome? Stroke

1996;27(11):2064–8.

[46] Persson L, Hardemark HG, Gustafsson J, et al. S-100 protein and neuron-specific enolase in

cerebrospinal fluid and serum: markers of cell damage in human central nervous system.

Stroke 1987;18(5):911–8.

[47] Ginsberg MD, Hedley-Whyte ET, Richardson EP Jr. Hypoxic-ischemic leukoencephalop-

athy in man. Arch Neurol 1976;33(1):5–14.

[48] SawadaH, Udaka F, Seriu N, et al. MRI demonstration of cortical laminar necrosis and de-

layed white matter injury in anoxic encephalopathy. Neuroradiology 1990;32(4):319–21.

[49] Castillo M. Imaging of cerebral infarction. Curr Probl Diagn Radiol 1998;27(4):105–31.

[50] Osborn A. Diagnostic neuroradiology. St. Louis (MO): Mosby; 1994. p. 330–98.

[51] Singhal AB, TopcuogluMA,KoroshetzWJ. DiffusionMRI in three types of anoxic enceph-

alopathy. J Neurol Sci 2002;196(1–2):37–40.

[52] Arbelaez A, Castillo M, Mukherji SK. Diffusion-weighted MR imaging of global cerebral

anoxia. AJNR Am J Neuroradiol 1999;20(6):999–1007.

[53] Chalela JA, Wolf RL, Maldjian JA, et al. MRI identification of early white matter injury in

anoxic-ischemic encephalopathy. Neurology 2001;56(4):481–5.

[54] Wijdicks EF, Campeau NG, Miller GM. MR imaging in comatose survivors of cardiac re-

suscitation. AJNR Am J Neuroradiol 2001;22(8):1561–5.

[55] Peerdeman SM, Girbes AR, VandertopWP. Cerebral microdialysis as a new tool for neuro-

metabolic monitoring. Intensive Care Med 2000;26(6):662–9.

[56] Peerdeman SM, Girbes AR, Vandertop WP. Changes in cerebral glycolytic activity during

transport of critically ill neurotrauma patients measured with microdialysis. J Neurol

2002;249(6):676–9.


[57] Raichle ME. The pathophysiology of brain ischemia. Ann Neurol 1983;13(1):2–10.

[58] SchaafsmaA, de JongBM,Bams JL, et al. Cerebral perfusion andmetabolism in resuscitated

patients with severe post-hypoxic encephalopathy. J Neurol Sci 2003;210(1–2):23–30.

[59] De Volder AG, Michel C, Guerit JM, et al. Brain glucose metabolism in postanoxic syn-

drome due to cardiac arrest. Acta Neurol Belg 1994;94(3):183–9.

[60] DeVolderAG,GoffinetAM,BolA, et al. Brain glucosemetabolism in postanoxic syndrome.

Positron emission tomographic study. Arch Neurol 1990;47(2):197–204.

[61] BrooksRA,DiChiroG,KellerMR.Explanationof cerebralwhite–gray contrast in computed

tomography. J Comput Assist Tomogr 1980;4(4):489–91.

[62] Kjos BO, Brant-Zawadzki M, Young RG. Early CT findings of global central nervous sys-

tem hypoperfusion. AJR Am J Roentgenol 1983;141(6):1227–32.

[63] Brooks RA. A quantitative theory of the Hounsfield unit and its application to dual energy

scanning. J Comput Assist Tomogr 1977;1(4):487–93.

[64] TorbeyMT, SelimM, Knorr J, et al. Quantitative analysis of the loss of distinction between

gray and white matter in comatose patients after cardiac arrest. Stroke 2000;31(9):2163–7.

[65] Zingler VC, Krumm B, Bertsch T, et al. Early prediction of neurological outcome after car-

diopulmonary resuscitation: a multimodal approach combining neurobiochemical and elec-

trophysiological investigations may provide high prognostic certainty in patients after

cardiac arrest. Eur Neurol 2003;49(2):79–84.

[66] Torbey MT, Geocadin R, Bhardwaj A. Brain arrest neurological outcome scale (BrANOS):

predicting mortality and severe disability following cardiac arrest. Resuscitation 2004;63(1):

55–63.

[67] Schell RM, Cole DJ. Cerebral monitoring: jugular venous oximetry. Anesth Analg 2000;

90(3):559–66.

[68] Bauer R, GablM, Obwegeser A, et al. Neurochemical monitoring using intracerebral micro-

dialysis during cardiac resuscitation. Intensive Care Med 2004;30(1):159–61.

[69] Pokela M, Biancari F, Rimpilainen J, et al. The role of cerebral microdialysis in predicting

the outcome after experimental hypothermic circulatory arrest. Scand Cardiovasc J 2001;

35(6):395–402.

[70] McCunn M, Mirvis S, Reynolds N, et al. Physician utilization of a portable computed

tomography scanner in the intensive care unit. Crit Care Med 2000;28(12):3808–13.

Neurol Clin 24 (2006) 123–132

Movement Disorders after Resuscitationfrom Cardiac Arrest

Arun Venkatesan, MD, PhDa,*, Steven Frucht, MDb

aDepartment of Neurology, Johns Hopkins University School of Medicine,

Baltimore, MD, USAbNeurological Institute, Columbia University College of Physicians and Surgeons,

New York, NY, USA

Those who survive cardiac arrest often experience significant neurologicimpairment. A rare, but often debilitating, consequence of cardiac arrest isthe development of movement disorders. A wide range of movement dis-orders, with many different causes, is observed after cardiac arrest. Cardiacarrest survivors may develop movement disorders from metabolic distur-bances resulting from hypoxic-ischemic damage to the liver or kidney,frommedications administered to treat other complications of cardiac arrest,or from cardioembolic ischemic stroke as a result of impaired myocardium orcardiac valves. This review focuses on movement disorders caused by cere-bral hypoxia after cardiac arrest. Many different movement disorders are de-scribed after hypoxic-ischemic brain injury, including parkinsonism,dystonia, chorea, tics, athetosis, tremor, and myoclonus [1–5]. Of thesemovement disorders, the one reported and investigated most extensively isposthypoxic myoclonus (PHM). Hence, this article describes the clinicalspectrum, pathophysiology, and treatment of PHM before briefly discussingother posthypoxic movement disorders.

Posthypoxic myoclonus

Myoclonus refers to sudden, shock-like, involuntary movements that canmanifest in various patterns. Myoclonus may be focal, where a few adjacentmuscles are involved; multifocal, where many muscles jerk asynchronously;or generalized, where most of the muscles of the body are involved in

* Corresponding author. Department of Neurology, Johns Hopkins Hospital Pathology,

509 600 North Wolfe Street, Baltimore, MD 21287.

E-mail address: [email protected] (A. Venkatesan).




124 VENKATESAN & FRUCHT

synchronized fashion. Additionally, myoclonic movements may be sponta-neous or they may be activated by either movement or sensory stimulation.Finally, myoclonus may be comprised of ‘‘positive’’ movements, in whicha burst of electromyographic activity is associated with the movement, ornegative movements, in which a brief pause of tonic muscular activity leadsto a jerk [6,7].

The causes of myoclonus are many. In 1963, however, Lance and Adamsdescribed four patients who developed severe myoclonus after surviving car-diac arrest [8]. These patients initially developed a generalized myoclonusaccompanied by dysmetria, dysarthria, and ataxia. Over time, the myoclo-nus persisted but its character changed to a predominantly action myoclo-nus involving the limbs. Lance and Adams hypothesized that thisparticular constellation of symptoms observed after cardiac arrest was theresult of cerebral hypoxia [8]. Since this initial description, more than 40years ago, more than 100 patients who have had PHM have been reportedin the medical literature [9–11], many of whom suffered hypoxia from car-diac arrest. Concordant with the initial descriptions by Lance and Adams,it is recognized that there are two types of PHM: acute PHM, which occurssoon after a hypoxic insult and is characterized by generalized myoclonus;and chronic PHM (Lance-Adams syndrome), which begins after a periodof delay and is manifested predominantly by action myoclonus.

Acute posthypoxic myoclonus

Acute PHM occurs soon after a hypoxic episode and is characterized bysevere, generalized myoclonic jerks in patients who are deeply comatose [7].The jerks begin typically within the first 24 hours after hypoxia and oftenare characterized by violent flexion movements. When they persist formore than 30 minutes or occur for most of the first postresuscitation day,some term the abnormal movements, myoclonic status epilepticus (MSE), de-spite the lack of definitive evidence that these movements represent epilepticactivity. PosthypoxicMSE occurs in approximately 30% to 40% of comatoseadult survivors of cardiopulmonary resuscitation and is difficult to controland associated with a poor prognosis. In the largest published series of post-hypoxic MSE, Wijdicks and colleagues find that all 40 patients had intermit-tent generalized myoclonus involving both face and limb muscles. Stimuli,such as touch, tracheal suctioning, and loud handclaps, triggered myoclonicjerks in most of the patients. None of the 40 patients who had acute posthyp-oxic MSE awakened, improved in motor response, or survived [12]. Inanother review of 18 patients who had posthypoxic MSE, 14 patients diedwithin 2 weeks and the two patients who survived were left with profound dis-ability [13]. A meta-analysis of patients who had posthypoxic MSE paintsa similarly grim picture: of 134 pooled cases, 119 (88.8%) died, 11 (8.2%)remained in a persistent vegetative state, and 4 (3.0%) survived. Of the fourpatients who survived, two were described as having a good outcome [13].

125MOVEMENT DISORDERS AFTER CARDIAC ARREST RESUSCITATION

Clues to the pathophysiology of acute PHM arise from electrophysiologicand histopathologic studies. The electroencephalograms (EEGs) in these pa-tients are variable, but often display bursts of generalized spikes and poly-spikes or a burst suppression pattern, believed to be consistent with severeneuronal injury. On autopsy, patients who have acute posthypoxic MSEhave evidence of neuronal ischemia and cell death in the cerebral cortex,deep gray nuclei (ie, basal ganglia and thalamus), hippocampus, and cere-bellum [13]. Cortical damage is more severe in patients who have MSEthan in those who do not have myoclonus, however [14]. Because the sever-ity of cortical damage implies that the cortex may not be capable of gener-ating any activity, including myoclonic activity, it is postulated that acutePHM arises from a brainstem generator [7].

Treatment of myoclonic jerks in acute PHM is difficult and of question-able usefulness, particularly in the setting of posthypoxic MSE. Multiplemedications often are used, the most common of which are phenytoin, val-proate, and benzodiazepines. In one study of 18 patients who had posthyp-oxic MSE, 13 patients required intravenous anesthesetic agents, includingpropofol and midazolam [13]. Despite aggressive treatment of the myoclonicjerks, poor prognosis resulting from the severity of underlying brain injury isthe rule rather than the exception.

Chronic posthypoxic myoclonusdLance-Adams syndrome

Chronic PHM, also known as Lance-Adams syndrome, typically occurswithin a few days to a few weeks after hypoxic injury. In a series of 14 pa-tients who had chronic PHM, all but one were noted to have the onset ofmyoclonus while still in coma [9]. The myoclonus has several characteristicfeatures. Patients have an action myoclonus involving predominantly thelimbs. Myoclonic jerks commonly appear immediately on attempting tomove or position a limb and occasionally spread to other portions of thebody. The jerks generally disappear with relaxation of the limb. The preci-sion of the motor task seems to be proportional to the severity of myoclo-nus, making everyday tasks, such as bringing a cup to the mouth orgrasping a small object, extremely difficult [8,10,11]. In addition, the myoc-lonus of chronic PHM has several other distinctive characteristics. In theirseries of 14 patients, Werhahn and colleagues report that 11 had stimulus-sensitive myoclonus [9]. Negative myoclonic jerks also contribute signifi-cantly to morbidity in patients who have chronic PHM. Postural lapses re-sulting from negative myoclonus predispose patients to frequent falls andoften result in patients being confined to wheelchairs [10].

Chronic PHM is a syndrome with diverse clinical, electrophysiologic, andneurochemical abnormalities [7,10], and the pathophysiology of this condi-tion is poorly understood. An example of the diversity of the disorder lies inthe nature of the myoclonus itself. The myoclonus in chronic PHM mayhave a cortical or subcortical origin. Clinical clues suggestive of a cortical


origin include myoclonus that is distally predominant, highly action in-duced, and stimulus sensitive. A cortical origin for myoclonus also is sup-ported by electrophysiologic measures, such as enlarged somatosensoryevoked potentials and the gold standard, back-averaged EEG. Using suchmeasures, it seems that cortical myoclonus is much more common in chronicPHM than subcortical myoclonus, the latter of which tends to cause violentjerks of the proximal limbs and trunk [10]. Patients who have chronic PHM,however, may suffer from cortical myoclonus, subcortical myoclonus, ora combination of the two, and there are no prognostic factors that can pre-dict which subtype of myoclonus a patient will develop.

The neurochemical and anatomic bases of chronic PHM remain unclear.Several lines of evidence suggest that specific neurotransmitter abnormalitiesare involved in the pathogenesis of chronic PHM. Isolated reports demon-strate that low levels of 5-hydroxyindole amino acid (5-HIAA), a serotoninmetabolite, are present in the cerebrospinal fluid of patients who have PHM[15]. Some patients who have chronic PHM improve with administration ofthe serotonin precursor, 5-hydroxytryptophan (5-HTP) [16]. Further sup-port for a role of the serotonergic system in chronic PHM comes from an-imal models. Several groups demonstrate that, in rat models of PHM,myoclonus improves with 5-HTP treatment and severity of myoclonus cor-relates inversely with levels of striatal serotonin and cortical 5-HIAA. Inaddition, direct modulation of serotonin receptors affects PHM, as demon-strated in animal models in which the administration of agonists and certainantagonists of serotonin receptors can ameliorate PHM [17–20].

Another clue that serotonin signaling may be involved in the pathophys-iology of PHM lies in the finding that the majority of patients in the largestpublished series on PHM are female. A recent study examines the role of es-trogen in PHM and finds that estrogen treatment of female rats that wereovariectomized resulted in a significant increase in intensity and durationof PHM [21]. It is hypothesized that estrogen, through its regulation of se-rotonergic activity, may influence the clinical course of PHM. Thus, it seemsthat although serotonergic modulation affects PHM, the relative contribu-tions of serotonin metabolism and of specific serotonin receptors to theoverall mechanisms of PHM remain unclear.

A recent report that describes marked exacerbation of chronic PHM ina patient administered trimethoprim-sulfamethoxazole (TMP-SMX) mayimplicate phenylalanine, a neurotransmitter precursor, in the pathogenesisof PHM. A patient who had high-grade non-Hodgkin’s lymphoma andchronic PHM whose myoclonus was well treated with oral piracetam devel-oped marked worsening of myoclonus on exposure to high-dose intravenousTMP-SMX. A reduction of the dosage of TMP-SMX resulted in a dramaticand rapid improvement of themyoclonic jerks [22]. The investigators believedit unlikely that alterations of renal or hepatic metabolism, or pharmacokine-tic or pharmacodynamic interactions between TMP-SMX and piracetam,accounted for the worsening of myoclonus; rather, they hypothesized that


TMP-SMX may have resulted in increased myoclonus by its well-describedaction of impairing phenylalaninemetabolism and, thereby, elevating phenyl-alanine levels. Elevated phenylalanine levels, in turn, are linked to severalneurologic conditions. Unfortunately, neither serum nor cerebrospinal fluidphenylalanine levels were obtained to support this hypothesis.

Recently, radiologic tools have been used in an attempt to understand theanatomic and pathophysiologic basis of PHM. Seven patients who hadchronic PHM and EEG back averaging that demonstrated cortical myoclo-nus underwent fluorodeoxyglucose–positron emission tomographic scan-ning. Comparedwith control subjects, patients who had PHM exhibitedsignificant increases in glucose metabolism in several brain regions, includingthe ventrolateral thalamus [23]. Such findings may be compatible with the ratPHM model, in which the ventrolateral thalamic nucleus is implicated indi-rectly in the pathogenesis of PHM; Purkinje cell death occurs selectively inthe paravermal and vermal areas, which project mainly to the dorsolateralprotuberance of the fastigial nucleus and, in turn, to the ventrolateral thalamicnucleus [24]. Thus, there may be a link between the Purkinje cell death seen inPHM rats and the increased metabolic uptake observed by PET scanning inthe ventrolateral thalamic nucleus of humans who have PHM.

The treatment of chronic PHMcan pose a challenge for several reasons. Aspreviously alluded to, chronic PHM is a syndrome with diverse clinical pre-sentations likely stemming from varied pathophysiologies. In addition, the ef-fect of many drugs reported in the literature is given in qualitative fashion,providing only an approximation of an agent’s efficacy. Also, the small num-ber of patients who develop this syndrome obviates the possibility of large-scale clinical trials to evaluate the efficacy of antimyoclonic agents. Despitethese difficulties, a recent review of the literature, which includes more than100 cases of PHM, supports several important conclusions. Clonazepam, val-proate, and piracetam demonstrate significant efficacy in approximately 50%of patients in whom these agents were instituted. Therefore, the authors con-sider these to be first-line agents in the treatment of chronic PHM. A role for5-HTP also is supported by the authors’ recent literature review. 5-HTP re-sulted in marked or full improvement in 40% of patients in whom the effectwas reported; however, concomitant treatment with carbidopa often is neces-sary to prevent the severe nausea that accompanies administration of thisdrug. Several other drugs, including baclofen, diazepam, ethanol, and meth-ysergide, also are reported as efficacious in a limited number of patients. Ananalysis of these 122 cases reveals that there are several drugs that were re-ported to be not significantly efficacious in any case of chronic PHM; theseinclude phenytoin, primidone, phenobarbitol, and tetrabenazine [10].

Several newer agents also are reported to have efficacy in the treatment ofchronic PHM. Levetiracetam, chemically related to piracetam, has beenstudied by several groups of investigators. Krauss and colleagues notethat one of two patients who had chronic PHM experienced substantial im-provement in myoclonus, whereas the other experienced some improvement


with doses of 500 to 750 mg twice daily [25]. Another study of the effects oflevetiracetam on myoclonus of different causes notes that rapid and sus-tained symptomatic improvement occurred only in a single patient whohad chronic PHM [26]. The authors also have studied the efficacy of levetir-acetam in PHM. They conducted an open-label, dose-escalation trial of thismedication in seven patients who had chronic myoclonus (including threewho had posthypoxic myoclonus) and showed that the mean Unified Myoc-lonus Rating Scale scores trended downward in every section after adminis-tration of levetiracetam, with significant decreases in the sections addressingpatient self-assessment and physician assessment of global disability [27].

Several reports describe patients who have PHM experiencing improve-mentofmyoclonuswith administrationof alcohol [28,29]. Basedonapreviousreport that a patient who had alcohol-responsive myoclonus-dystonia hadsignificant improvement when treated with g-hydroxybutyric acid (GHB)[30], the authors recently studied the efficacy of GHB in one patient whohad alcohol-responsive chronic PHM. Using an open-label, dose-finding,blinded-rater approach, they found that oral GHB was markedly effectivein ameliorating severe alcohol-sensitive PHM in this single patient [31].

Regardless of treatment, the majority of patients who have chronic PHMimprove over time. Myoclonus, ataxia, and speech all tend to improve overseveral years, and disability scores reflective of the ability to ambulate, com-municate, and take care of themselves also improve [8,9]. In one series of 14patients who had chronic PHM, only four patients did not have an improve-ment in global disability score at a mean follow-up of 3.7 years [9]. Althoughmany patients do experience significant improvement in symptoms, somewho have chronic PHM remain significantly disabled despite medical ther-apy. Currently, there are no well-accepted surgical options for such patients.The authors hypothesize, however, that, given the ventral thalamic hyper-metabolism observed on PET scan in these patients and the fact that thala-motomy and thalamic stimulation are applied successfully to single patientswho have intractable myoclonus, stereotactic targeting of the ventrolateralthalamus using deep brain stimulation may be appropriate in some patientswho have severe, medication-refractory chronic PHM [23].

Other posthypoxic movement disorders

Avariety of othermovement disorders are observed after cerebral hypoxia,including parkinsonism, dystonia, chorea, athetosis, and tremor [1–5]. Al-though PHM may result from injury to the cerebellum or thalamus, manyof these other movement disorders are caused by damage of the basal ganglia.Dystonia is one of the more common movement disorders to occur aftercerebral hypoxia and may develop in combination with an akinetic-rigid(parkinsonian) syndrome. This article discusses the clinical spectrum of post-hypoxic dystonic and akinetic-rigid syndromes and focuses on the pathophys-iology of basal ganglia dysfunction in these conditions.


Dystonic and akinetic-rigid syndromes, alone or in combination, repre-sent a sizeable proportion of the posthypoxic movement disorders describedin the literature. Many of these cases are reported in patients surviving car-diac arrest. These syndromes may occur acutely, either at the time the hyp-oxic insult occurs or shortly thereafter, or more commonly in delayedfashion, months to years after the initial hypoxic insult [32]. The posthyp-oxic akinetic-rigid syndrome usually is a symmetric condition characterizedby various combinations of bradykinesia, micrographia, axial and appendic-ular rigidity, resting or postural tremor, and marked postural instability[3,33]. Posthypoxic dystonia can affect the limbs and face and often is asym-metric at onset with progression to a symmetric, generalized dystonia. Ina review of 12 patients who previously were normal and who suffered hyp-oxic ischemic insults of various causes, including cardiac arrest, Marsdenand colleagues note that six of the patients developed a pure dystonic syn-drome, two developed a pure akinetic-rigid syndrome, and four initially de-veloped an akinetic-rigid syndrome followed later by a dystonic syndrome[34]. The akinetic-rigid syndrome developed typically within 3 months ofthe hypoxic event; after a rapid evolution, the majority of patients remainedclinically stable for many subsequent years. In contrast, the pure dystonicsyndrome developed, on average, 10 months after the hypoxic event, andprogressed gradually over several years. The majority of patients had visiblelesions in the basal ganglia on brain CT or MR imaging. Treatment ofakinetic-rigid symptoms with levodopa or dopamine agonists and adminis-tration of high-dose anticholinergic drugs for dystonic symptoms conferredlittle benefit to these patients.

Why do some patients who have basal ganglia lesions after cerebral hyp-oxia develop an akinetic-rigid syndrome, whereas others develop a predom-inantly dystonic syndrome? Marsden and colleagues note that the mean ageof the akinetic-rigid group at the time of anoxia was 41 years, whereas thatof the pure dystonic group was 13.5 years. Indeed, all six patients who de-veloped a pure dystonic syndrome were ages 21 years or less. This observa-tion led to the hypothesis that an age-dependent difference in the clinicalmanifestations of hypoxia exists, with younger people more prone to dysto-nia and older individuals more prone to an akinetic-rigid state [33]. Suchage-dependent differences also are observed in conditions, such as Parkinson’sdisease, in which early-onset patients are predisposed to dystonia. A mecha-nistic understanding of these observations, however, remains elusive.

The location of brain injury within the basal ganglia seems to be anotherfactor that governs whether or not patients develop dystonia versus anakinetic-rigid state after cerebral hypoxia. Hawker and Lang, in their case se-ries of three patients who had suffered cerebral hypoxia, note that the two pa-tients who had primarily dystonic syndromes had lesions of the putamina onhead CT, whereas the patients who had an akinetic-rigid syndrome hadmarked bilateral lesions of the globus pallidus. Similar clinicoanatomicassociations are found in a wide variety of insults to the basal ganglia,


including trauma, neurodegenerative diseases, and encephalitides [4]. Thishas led to the proposal that in the setting of hypoxia, lesions of the globuspallidus are responsible for the akinetic-rigid syndrome, whereas lesions ofthe putamen account for dystonia [2]. Further support for this hypothesisis provided by Marsden and colleagues, who find that dystonia is associatedwith putaminal injury in 10 of 14 cases and an akinetic-rigid syndrome is as-sociated with globus pallidus lesions in 11 of 14 cases. This association is notabsolute, however, as several examples of dystonia associated with pallidallesions and parkinsonism associated with putaminal lesions are noted [33].

Two other questions arise with respect to the pathophysiology of hypoxia-induced basal ganglia lesions. First, why are the basal ganglia so vulnerableto hypoxic insults? Two main hypotheses are put forth to explain this selec-tive vulnerability. The ‘‘vascular hypothesis’’ states that selective hypoperfu-sion results from the vascular supply of the basal ganglia and, in particular,the globus pallidus, underlies its susceptibility to hypoxic injury. The secondtheory is the ‘‘metabolic hypothesis,’’ which postulates that factors intrinsicto the striatum, such as intrinsically high oxidative metabolism or high den-sity of excitatory amino acid receptors, results in hypoxic damage [2,34].Further studies are needed to determine whether or not vascular, metabolic,or other factors underlie the susceptibility of the basal ganglia to hypoxic in-jury. The second main question that arises is, how can a single exposure tocerebral hypoxia lead to the delayed onset and progression of symptomsyears later? Several mechanisms, including aberrant sprouting, synaptic reor-ganization, ephaptic transmission, and inflammatory changes, are suggestedas possible mechanisms of delayed symptomatology. Parallels with otherneurodegenerative diseases, in which excitotoxicity is followed by mitochon-drial dysfunction, oxidative stress, and eventual neuronal apoptosis, also arespeculated to play a role [34,35]. Regardless of the mechanism, it seems thatdamage to the basal ganglia with preservation of the pyramidal system isa pathologic correlate of delayed posthypoxic dystonia or akinetic-rigidsyndromes [36].

Summary

It is difficult to predict precisely the final neurologic outcome from car-diac arrest and accompanying cerebral hypoxia. Although rare, severalmovement disorders may arise as a consequence of hypoxic injury, includingmyoclonus, dystonia, akinetic-rigid syndromes, tremor, and chorea. Dys-function of various portions of the central nervous system, includingthe basal ganglia, thalamus, midbrain, and cerebellum, is implicated in thepathogenesis of these posthypoxic movement disorders. The developmentof animal models of posthypoxic movement disorders and of newer imagingtechniques applied to human patients who have movement disorders afterhypoxic episodes has improved understanding of the pathophysiology ofposthypoxic movement disorders and has suggested newer treatments.


Many outstanding questions remain, however. What factors promote sus-ceptibility to the development of posthypoxic movement disorders? Whydo patients who have similar clinical hypoxic insults develop markedly dis-similar movement disorders? Why are the basal ganglia especially vulnerableto cerebral hypoxia? Why do some movement disorders occur in delayedfashion and progress for years after the hypoxic insult? Is the pathogenesisof progressive posthypoxic movement disorders related to that of neurode-generative diseases? What are the most effective medications for the variousposthypoxic movement disorders? Is there a role for deep brain stimulationin the treatment of posthypoxic movement disorders? We anticipate thatcurrent and future research in the area of posthypoxic movement disorderswill reveal answers to some of these important questions.

References

[1] Fahn S. Posthypoxic action myoclonus: literature review update. Adv Neurol 1986;43:

157–69.

[2] Hawker K, Lang AE. Hypoxic-ischemic damage of the basal ganglia. Case reports and a re-

view of the literature. Mov Disord 1990;5:219–24.

[3] Feve AP, Fenelon G, Wallays C, et al. Axial motor disturbances after hypoxic lesions of the

globus pallidus. Mov Disord 1993;8:321–6.

[4] Bhatia KP, Marsden CD. The behavioral and motor consequences of focal lesions of the

basal ganglia in man. Brain 1994;117:859–76.

[5] Govaerts A, Zandijcke MV, Dehaene I, et al. Posthypoxic midbrain tremor. Mov Disord

1998;13:359–61.

[6] Fahn S,MarsdenCD,WoertMHV.Definition and classification of myoclonus. AdvNeurol

1986;43:1–5.

[7] Hallett M. Physiology of human posthypoxic myoclonus. Mov Disord 2000;15(Suppl 1):

8–13.

[8] Lance JW, Adams RD. The syndrome of intention or action myoclonus as a sequel to

hypoxic encephalopathy. Brain 1963;86:111–36.

[9] Werhahn KJ, Brown P, Thompson PD, et al. The clinical features and prognosis of chronic

posthypoxic myoclonus. Mov Disord 1997;12:216–20.

[10] Frucht S, Fahn S. The clinical spectrum of posthypoxic myoclonus. Mov Disord 2000;

15(Suppl 1):2–7.

[11] Frucht SJ. The clinical challenge of posthypoxic myoclonus. Adv Neurol 2002;89:85–8.

[12] Wijdicks EFM, Prisi JE, Sharbrough FW. Prognostic value of myoclonus status in comatose

survivors of cardiac arrest. Ann Neurol 1994;35:239–43.

[13] Hui ACF, Cheng C, Lam A, et al. Prognosis following postanoxic myoclonus status epilep-

ticus. Eur Neurol 2005;54:10–3.

[14] Young GB, Gilbert JJ, Zochodne DW. The significance of myoclonic status epilepticus in

postanoxic coma. Neurology 1990;40:1843–8.

[15] Fahn S. Posthypoxic action myoclonus: review of the literature and report of two new cases

with response to valproate and estrogen. Adv Neurol 1979;26:49–82.

[16] Woert MHV, Rosenbaum D. L-5-hydroxytryptophan therapy in myoclonus. Adv Neurol

1979;26:107–15.

[17] TruongDD,MatsumotoRR, Schwartz PH, et al. Novel rat cardiac arrestmodel of posthyp-

oxic moclonus. Mov Disord 1994;9:201–6.

[18] Matsumoto RR, Aziz N, Truong DD. Association between brain indole levels and severity

of posthypoxic myoclonus in rats. Pharmacol Biochem Behav 1995;50:553–8.


[19] Goetz CG, Vu TQ, Carvey PM, et al. Posthypoxic myoclonus in the rat: natural history, sta-

bility, and serotonergic influences. Mov Disord 2000;15(Suppl 1):39–46.

[20] Truong DD, Kirby M, Kanthasamy A, et al. Posthypoxic myoclonus animal models. Adv

Neurol 2002;89:295–306.

[21] Kompoliti K, Goetz CG, Vu TQ, et al. Estrogen supplementation in the posthypoxic myoc-

lonus rat model. Clin Neuropharm 2001;24:58–61.

[22] Jundt F, Lempert T, Dorken B, et al. Trimethoprim-sulfamethoxazole exacerbates post-

hypoxic action myoclonus in a patient with suspicion of Pneumocystis jiroveci infection.

Infection 2004;32:176–8.

[23] Frucht SJ, TrostM,MaY, et al. Themetabolic topography of posthypoxicmyoclonus.Neu-

rology 2004;62:1879–81.

[24] Welsh JP, Yeun G, Placantonakis DG, et al. Why do Purkinje cells die so easily after global

brain ischemia? Aldolase C, EAAT4, and the cerebellar contribution to posthypoxic myoc-

lonus. Adv Neurol 2002;89:331–59.

[25] Krauss GL, Bergin A, Kramer RE, et al. Suppression of posthypoxic and postencephalitic

myoclonus with levetiracetam. Neurology 2001;56:411–2.

[26] Lim LL, Ahmed A. Limited efficacy of levetiracetam on myoclonus of different etiologies.

Park Rel Dis 2005;11:135–7.

[27] Frucht SJ, Louis ED, Chuang C, et al. A pilot tolerability and efficacy study of levetiracetam

in patients with chronic myoclonus. Neurology 2001;57:1112–4.

[28] Genton P, Guerrini R. Effect of alcohol on action myoclonus in Lance-Adams syndrome

and progressive myoclonic epilepsy. Mov Disord 1992;7:92.

[29] Jain S, JainM. Action myoclonus (Lance-Adam syndrome) secondary to strangulation with

dramatic response to alcohol. Mov Disord 1992;6:183.

[30] Priori A, Bertolasi L, Pesenti A, et al. Gammahydroxybutyric acid for alcohol-sensitive my-

oclonus with dystonia. Neurology 2000;54:1706.

[31] FruchtSJ,BordelonY,HoughtonWH.Markedameliorationof alcohol-responsiveposthyp-

oxic myoclonus by gamma-hydroxybutyric acid (Xyrem). Mov Disord 2005;20:745–51.

[32] Burke RE, Fahn S, Gold AP. ‘‘Delayed-onset dystonia in patients with ‘‘static’’ encephalop-

athy. J Neurol Neurosurg Psychiatry 1980;43:789–97.

[33] Bhatt MH, Obeso JA, Marsden CD. Time course of postanoxic akinetic-rigid and dystonic

syndromes. Neurology 1993;43:314–7.

[34] Kuoppamaki M, Bhatia K, Quinn N. Progressive delayed-onset dystonia after cerebral an-

oxic insult in adults. Mov Disord 2002;17:1345–9.

[35] Scott BL, Jankovic J. Delayed-onset progressive movement disorders after static brain

lesions. Neurology 1996;46:68–74.

[36] Boylan KB, Chin JH, DeArmond SJ. Progressive dystonia following resuscitation from car-

diac arrest. Neurology 1990;40:1458–61.

Neurol Clin 24 (2006) 133–145

Cognitive and NeurobehavioralDysfunction after Cardiac Bypass

Procedures

Ola A. Selnes, PhDa,*, Guy M. McKhann, MDa,b,Louis M. Borowicz, Jr, MSb,Maura A. Grega, RN, MSNc

aDepartment of Neurology, Division of Cognitive Neuroscience,

Johns Hopkins University School of Medicine, Reed Hall East–2,

1620 McElderry Street, Baltimore, MD 21287, USAbZanvyl Krieger Mind/Brain Institute, The Johns Hopkins University, 3400 N. Charles Street,

338 Krieger Hall, Baltimore, MD 21218, USAcDepartment of Surgery, Johns Hopkins University School of Medicine, 600 N. Wolfe Street,

Blalock 618, Baltimore, MD 21287, USA

Since the introduction of cardiopulmonary bypass nearly 5 decades ago,there has been a significant reduction in the morbidity and mortality associ-ated with open heart surgery. Nonetheless, adverse neurologic outcomes af-ter cardiac surgery continue to remain a significant concern. The spectrumof these outcomes ranges from coma and debilitating stroke to encephalop-athy, delirium, and cognitive impairment. Recent studies suggest that theincidence of these adverse outcomes may be closely related to the status ofthe patient’s brain before surgery. Patients who have had transient ischemicattacks or stroke or who have a history of risk factors for cerebrovasculardisease appear to be at greater risk for postoperative neurologic complica-tions. This article focuses on the short- and long-term cognitive changes af-ter coronary artery bypass grafting (CABG).

This research was supported by grant 35610 from the National Institutes of Health

National Institute of Neurological Disorders and Stroke, the Dana Foundation, and the

Johns Hopkins Medical Institutions GCRC grant RR 00052.


E-mail address: [email protected] (O.A. Selnes).




134 SELNES et al

Short-term cognitive changes

Numerous studies have examined the incidence of cognitive declineafter bypass surgery, but several questions regarding the specificity, epide-miology, and pathophysiology of these changes remain unanswered. Esti-mates of cognitive change after CABG have been highly variable becauseof differences in study exclusion criteria, choice of time points for measuringfollow-up, and statistical criteria used for defining decline. Although earlystudies focused almost exclusively on the role of surgery-related factors, re-cent studies have attempted to take into account patient-related variables aswell. The patients undergoing CABG today not only are older but they alsohave a greater prevalence of comorbid diseases, particularly those that areknown risk factors for cerebrovascular disease.

Methodologic issues

The incidence of short-term cognitive changes varies according to the in-terval between surgery and follow-up testing. Neurocognitive test perfor-mance at the time of hospital discharge after CABG may be subject topotential confounders relating to the pharmacologic effects of anestheticdrugs or other clinical issues [1]. Therefore, some investigators have chosento defer follow-up testing until 3 to 4 weeks after surgery. Although thisstrategy may yield follow-up data that are less contaminated by nonspecificsurgical factors, extending the follow-up time may also mask transientchanges in cognition. Not only may recognition of such cognitive outcomesbe important for understanding the pathophysiology of early post-CABGcognitive changes but some researchers have suggested that early, transientpostoperative cognitive changes may also be predictive of late cognitive de-cline up to 5 years after surgery [2].

A second factor of critical importance for accurately estimating the inci-dence of postoperative cognitive decline after CABG is the inclusion ofa control group. Some studies have controlled for the effect of age on cog-nitive performance [3], whereas others have chosen control patients whohave been diagnosed with coronary artery disease [4] or have chosen hospi-talized inpatients undergoing noncardiac procedures [5]. Because of the im-portance of controls for interpreting cognitive changes after CABG, thisreview emphasizes studies that included a control group.

Controlled studies

In a recent study of acute postoperative cognitive outcomes from Ger-many, 67 CABG patients were examined before surgery and at days 3, 6,and 9 after surgery [6]. The study group was relatively healthy, in that pa-tients who had a history of previous stroke, carotid stenosis, or general med-ical disorders and those who had low baseline cognitive test scores wereexcluded. Hospitalized patients who had peripheral neuropathy served as

135NEUROCOGNITIVE DYSFUNCTION AFTER CABG

control subjects. Although there was a significant decline in neuropsycho-logic test performance at days 3 and 6, return to baseline levels of per-formance or above was observed by day 9. Thus, in a relatively low-riskgroup of CABG patients, postoperative cognitive decline appeared to bemild and reversible within a period of less than 2 weeks.

In a study comparing the neuropsychologic test performance of 57CABG patients with 55 control subjects from a senior citizen wellness pro-gram, 37 of the CABG patients completed follow-up testing 3 to 4 weeks af-ter surgery. Patients who had previous CABG and those who had a historyof visual impairments were excluded. The CABG patients had significantlylower baseline performance than the control subjects for some tests. Al-though the neuropsychologic test performance of the CABG group didnot decline for any of the neuropsychologic tests at follow-up, the CABGgroup showed a reduced practice effect compared with the control subjects.Thus, in this relatively unselected group of CABG patients, lower than ex-pected cognitive performance was observed preoperatively, but no signifi-cant decline was observed compared with healthy control subjects at 3 to4 weeks after surgery [3].

In a recent study, the authors [4] evaluated 140 CABG patients and 92demographically similar control subjects who had diagnosed coronary ar-tery disease but no surgery. Both groups improved from baseline to 12weeks, and apart from an unexpected greater improvement in verbal mem-ory among the CABG patients, there were no statistically significant differ-ences between the two groups. Together with the findings from other studiesof earlier postoperative outcomes, this study demonstrates that cognitive de-cline after CABG is transient and reversible and that most patients return totheir baseline cognitive performance between 3 to 12 weeks after surgery.

There is little consensus regarding which specific cognitive functions aremost vulnerable during the immediate postoperative period. Some studieshave reported early decline in several cognitive domains including memory,psychomotor speed, executive functions, and visuoconstructional abilities,suggesting that multiple brain regions may be involved [7]. From the per-spective of the patient and family members, the most frequent complaintinvolves changes in concentration and memory [8]. Because subjective cog-nitive complaints do not always correlate well with objective neuropsycho-logic test performance, some investigators have dismissed such subjectivecomplaints as being secondary to depression. There is some evidence fromnoncardiac populations, however, that subjective memory complaints mayreflect changes in memory that may not be captured by standardized testsof new verbal memory and delayed recall [9].

Pathophysiology of early cognitive changes

No single factor that can account for the early postoperative cognitivechanges has yet been identified. The focus of most investigations has been

136 SELNES et al

on neural injury secondary to surgery-related factors, including microem-boli, hypoperfusion, and the systemic inflammatory response. It has provedsurprisingly difficult, however, to find direct evidence that any of these var-iables, individually or in combination with other risk factors, can accountfor the short-term cognitive changes.

Microemboli

Patients who have atherosclerosis, especially of the carotid arteries andaortic arch, are known to be at increased risk for cerebral microemboli dur-ing the surgery [10]. Several studies using transcranial Doppler have demon-strated that showers of emboli are common during cardiac surgery [11],particularly during cannulation and clamping/unclamping of the aorta.These emboli vary in size and composition. Some studies have reportedthat most of these emboli appear to be gaseous rather than solid [12]; how-ever, the clinical significance of the emboli remains to be determined. Someearlier studies demonstrated a modest association between embolic countsand short-term cognitive outcomes [13,14], but others have not found anystatistically significant correlations [15–17]. It is unclear whether these dis-crepant findings are due to technical aspects such as problems distinguishingbetween solid versus gaseous emboli or to other factors.

It is possible that the cognitive manifestations of microemboli may de-pend as much on patient- related risk factors (such as the degree of pre-existing cerebrovascular disease) as on the number and size of the embolicload [18]. Patients who do not have significant pre-existing cerebrovasculardisease may have a higher tolerance for embolic injury than those who havesuch disease. Consistent with this, the predictors of cognitive decline ina large multicenter Department of Veterans Affairs study included cerebro-vascular disease, peripheral vascular disease, and a history of chronic dis-abling neurologic illness [19].

It is also expected that the cognitive implications of embolic injury de-pend on which parts of the brain are most heavily exposed. Special stainingtechniques have demonstrated numerous capillary and arteriolar dilitationsin the brains of patients who die shortly after their surgery, but the regionaldistribution of these emboli has not been described. Most of these presumedembolic changes disappear over time and are not seen in patients who cometo autopsy a week or longer after the surgery [20]. Other studies have re-ported low numbers of emboli in the brains of patients who die shortly afterCABG, but cerebral microbleeds are relatively common [21].

Atrial fibrillation

Atrial fibrillation is a common complication after CABG surgery andmay be associated with hemodynamic changes, increased risk for adverseneurologic outcomes, and prolonged hospitalization. The only variable


that has been consistently associated with the development of postoperativeatrial fibrillation is older age [22]. Approximately one third of patientsundergoing cardiac surgery with cardiopulmonary bypass have new-onsetpostoperative atrial fibrillation [23,24]. Those who have recurrent episodesof atrial fibrillation are at increased risk of stroke [25] and neurocognitivedecline [26].

Hypoperfusion

Longstanding hypertension and aging are associated with morphologicchanges of the brain vascular supply that may predispose the elderly tothe effects of hypoperfusion. Certain regions of the brain, including the hip-pocampus, periventricular white matter areas, and watershed areas, may bemore susceptible to the effects of hypoperfusion. Abildstrom and colleagues[27] found that candidates for CABG had lower global cerebral blood flowpreoperatively than controls, but there was no correlation between neuro-psychologic test performance and postoperative global or regional bloodflow. Caplan and Hennerici [28] proposed that emboli and hypoperfusionmay play a synergistic role (ie, decreased flow during the surgery may resultin reduced washout of embolic materials from the brain) and that the water-shed areas are particularly susceptible to this combination.

Anesthesia

A significant percentage of elderly patients undergoing major noncardiacsurgery with general anesthesia also suffer short- or long-term cognitive dys-function. Although increasing age appears to be the principal risk factor,postoperative cognitive decline has also been reported in younger patients.In a study of patients in the 40- to 60-year age range, 19% were found tohave cognitive decline 7 days after surgery with general anesthesia, whichis comparable to the incidence reported for patients 60 years or older [29].Studies comparing regional and general anesthesia have not found any dif-ference in the incidence of cognitive decline 3 months after surgery, thusquestioning a direct causal relationship between general anesthesia andpostoperative cognitive dysfunction [30]. Regardless of the specific etiologyof postoperative cognitive impairment after general anesthesia, there is evi-dence of some degree of short-term cognitive decline even after major non-cardiac surgery with general anesthesia.

Depression

Mild to moderate depression is common after CABG, but one of the bestpredictors of postoperative depression is being depressed preoperatively,thus suggesting that postoperative depression is not caused by CABG. Anec-dotally, short-term cognitive decline after CABG was often attributed to de-pression, and some cross-sectional investigations in noncardiac populations

138 SELNES et al

have reported an association between depression and performance on neuro-psychologic tests [31]. In prospective studies, however, there is no evidencethat new-onset depression after CABG correlates with short- or long-termchanges in cognitive performance [32].

Genetic factors

The possibility of a link between genetic factors and risk of cognitive de-cline after CABG was suggested by a study that found a greater likelihoodof decline in patients who had the apolipoprotein (apo) E epsilon4 allele[33]. Subsequent studies, however, have not been able to replicate these find-ings [34]. Steed and colleagues [35] evaluated 111 CABG patients preopera-tively and at 4 to 7 weeks postoperatively and reported no relationshipbetween change in neuropsychologic test performance and apo E allele status.In a study of longer-term cognitive outcomes after CABG, no significant as-sociation was found between apoE status and cognitive decline at 5 years [36].

Long-term changes

Although most studies have focused on short-term cognitive decline afterCABG, a study from Duke University raised the possibility of late or de-layed cognitive decline after CABG. Newman and colleagues [2] studied261 patients before surgery and followed them prospectively before dis-charge and at 6 weeks, 6 months, and 5 years after CABG surgery. The in-cidence of decline at the time of discharge was 53%, dropping to 24% at 6months. At 5 years, 66% of the patients were available for follow-up testing,and an unexpected 42% of these patients performed below their baselineperformance on a global measure of cognition. Predictors of late cognitivedecline included older age, fewer years of education, higher baseline score,and cognitive decline at the time of discharge.

Selnes and colleagues [34] evaluated 172 patients before and after CABGand followed them prospectively for up to 5 years. Similar to the findingsfrom the Duke study, a statistically significant decline in performance wasobserved for most cognitive domains between 1 and 5 years. Comparingbaseline to performance at 5 years, decline was observed for only two cog-nitive domains (psychomotor speed and visuoconstruction).

In a more recent study, Stygall and coworkers [36] obtained 5-year neuro-psychologic follow-up on 107 of 171 CABG patients who were evaluated be-fore surgery. Relative to baseline performance, they found decline in theoverall neuropsychologic change z-score score at 6 days, followed by im-provement at 8 weeks and decline at 5 years. As in previous studies oflate cognitive decline, the greatest change was observed for measures of mo-tor and psychomotor speed. Surprisingly, no decline was observed for mostmeasures of verbal learning and memory. The number of emboli during thesurgery, the degree of decline at the time of discharge, and the degree of


recovery between discharge and follow-up testing at 6 to 8 weeks were amongthe predictors of late decline. The investigators suggested that the number ofemboli during surgery might be a surrogate marker for severity of cerebro-vascular disease and that patients who have more severe pre-existing cerebro-vascular disease might be more likely to have late decline.

Some studies, however, have not found evidence of late cognitive declineafter CABG. In a small study from Germany, 52 patients were followed upto 5 years after CABG [5]. None of these patients had what was consideredclinically significant cognitive decline and only a small subset had mild de-cline. The investigators suggested that better control of hypertension, hyper-cholesterolemia, and other risk factors for cerebrovascular disease duringthe 5-year follow-up period was a possible explanation for the lack of latedecline in their study. If these findings are replicated in larger cohorts, itwould suggest that late cognitive decline after CABG could be avoided byimproved postoperative control of risk factors that accelerate progressionof underlying cerebrovascular disease [37].

Only two long-term follow-up studies published to date have includeda control group. Hlatky and colleagues [38] obtained cross-sectional neuro-psychologic test performance data for patients who had been randomized tostandard coronary artery bypass (N ¼ 125) or angioplasty (N ¼ 64) 5 yearsearlier. In an intention-to-treat analysis, there were no significant differencesin the 5-year cognitive test scores for these two groups. Although this studydid not evaluate cognitive changes prospectively, the findings nonethelessconfirm that 5 years after the procedure, the cognitive performance of pa-tients who had coronary artery bypass surgery did not differ from that ofpatients who had angioplasty.

In a study of twins, the postoperative cognitive performance of 232 CABGpatients, stratified across three age categories, was compared with that of theirtwins who did not undergo CABG. Surprisingly, CABG patients who hadtheir surgery at a relatively young age (between age 63 and 70 years) had bettercognitive performance 1 to 2 years postoperatively than the co-twin who didnot have surgery. No significant differences in cognitive performance werefound for the twin pairs in the older age groups [39].

In summary, there is evidence from several studies that late cognitive de-cline occurs between baseline and 5 years after surgery. The degree of de-cline relative to baseline performance appears to be relatively minor andis observed mainly in the domains of motor or psychomotor speed, withno significant decline in memory performance. The pattern of these late cog-nitive changes is similar to what is seen in patients who have mild subcorti-cal vascular disease.

Etiology of late cognitive changes

Of the prospective studies that found evidence of late cognitive declineafter CABG, none included a control group. It remains unclear, therefore,

140 SELNES et al

whether the late cognitive changes are causally related to cardiopulmonarybypass with general anesthesia 5 years earlier, normal aging, development ofAlzheimer’s disease during the follow-up period, or other causes. The selec-tion of an appropriate control group for patients undergoing CABG hasbeen controversial. Patients undergoing CABG today are known to havea high prevalence of hypertension, diabetes, and other risk factors for cere-brovascular disease. Because these factors by themselves are associated withmild cognitive decline over time, the control group should include subjectswho have a similar profile of risk factors for cerebrovascular disease. Dura-tion or severity of these risk factors cannot easily be measured, however, andhow these risk factors translate into eventual vascular disease of the brainremains unknown.

Cerebrovascular disease risk factors

There is accumulating evidence from several epidemiologic studies dem-onstrating that a history of one or more risk factors for cerebrovascular dis-ease may be associated with accelerated cognitive decline even withoutcardiac surgery [40–43]. In a study of community-dwelling individuals,Knopman and colleagues [44] reported that participants who had a historyof diabetes or hypertension at baseline had greater cognitive decline overa 4- to 6-year follow-up period than participants who did not have suchrisk factors. Other studies have found that diabetes alone may be associatedwith cognitive decline over time. Fontbonne and colleagues [45] reportedthat subjects who had diabetes had lower cognitive test scores at 4-year fol-low-up testing than subjects who had normal glucose levels. There is evi-dence that longer duration of diabetes is associated with worse cognitiveperformance [46], and data suggest that individuals who have diabetesand hypertension have greater cognitive decline in late life [47]. Finally,there is evidence that treatment of these risk factors for vascular diseasemay prevent any late cognitive consequences [46,48]. These findings are con-sistent with the findings from the 5-year CABG follow-up study by Mullgesand colleagues [5]. These investigators hypothesized that better control ofrisk factors for vascular disease during the 5 years after surgery may haveaccounted for the lack of late decline seen in their study.

In summary, there is now considerable evidence from epidemiologic stud-ies demonstrating an association between the duration and degree of vascu-lar disease and the risk of cognitive decline during the later years of life, evenin community-dwelling elderly individuals who have not undergone CABG.

Pre-existing MRI abnormalities

Some studies have found that nearly one third of otherwise asymptomaticindividuals have silent brain infarcts on MRI. Several studies have concludedthat older age and hypertension are significant risk factors for having such


MRI abnormalities [49]. The presence of such lesions may be associated withprogressive cognitive decline or late dementia [49,50]. The neuropsychologicprofile of patients who have subcortical silent infarcts depends on the overallvolume and location of the lesions, but there may be additional hemody-namic factors that contribute to the clinical expression of such lesions [51].Several studies have reported that motor and psychomotor speed are typi-cally slowed in patients who have silent subcortical infarcts [49,52].

Given the high prevalence of silent infarcts in community-dwellingindividuals, one would expect such MRI findings to be even more com-mon among candidates for CABG. Because of the relatively short time be-tween admission and surgery, however, preoperative brain imaging has beendifficult to obtain. In a study from Japan, preoperative MRI scans wereperformed in 421 candidates for CABG [53]. Of these patients, 30% werefound to have single, small brain infarctions and 20% had multiple infarc-tions. Thus, an unexpected half of this group had evidence of silent brainabnormalities before surgery. Patients who had single or multiple infarc-tions had lower baseline cognitive performance and were more likely tohave decline in cognitive test performance postoperatively.

A more recent study confirmed a high frequency of cerebrovascular dis-ease in candidates for CABG. Nakamura and colleagues [54] obtained pre-operative MRI scans in 91 patients and found that 33 had small infarctions,38 had multiple small infarctions, and 8 had infarctions greater than 15 mm.Thus, silent ischemic cerebral disease is commonly seen preoperatively in pa-tients undergoing CABG and is associated with an increased risk of postop-erative cognitive decline in the short-term. Whether pre-existing silent MRIabnormalities are also predictive of late cognitive changes has not yet beenexamined.

Off-pump coronary artery bypass graft

The recent development of techniques for performing CABG surgery with-out the use of cardiopulmonary bypass has generated considerable optimismthat such techniques might lead to a reduction of adverse neurologic and cog-nitive outcomes. Direct comparison with outcomes in CABGpopulations hasbeen difficult because most off-pump CABG studies included relativelyhealthy, low-risk patients [55]. Although the incidence of postoperative atrialfibrillationmay be lower after off-pumpCABG, it nonetheless remains a com-mon postoperative complication [22]. Several studies have demonstrated thatthe use of off-pump surgery is associated with a reduction in the number ofemboli to the brain [56,57], but clear-cut benefits in terms of neurocognitiveoutcomes are less obvious. In the only large-scale prospective randomizedstudy to date, there was no significant difference in the incidence of declinebetween patients having conventional on-pump versus off-pump surgery at3 or 12 months [58]. Prospective long-term follow-up studies of cognition in

142 SELNES et al

off-pump patients have not yet been reported; therefore, it is not knownwhether the degree of late decline in these patients differs from that observedin CABG patients.

Summary

From a cognitive standpoint, CABG as currently practiced appears to besafe for the great majority of patients, but transient changes involving mem-ory, executive functions, and motor speed may still occur in a subset of pa-tients during the first few days to weeks after CABG. The etiology mostlikely is multifactorial and includes a synergistic effect of microemboli, hypo-perfusion, and other variables associated with major surgery. Older age anddegree of pre-existing cerebrovascular disease have been identified as impor-tant risk factors. The short-term cognitive changes appear to be reversible by3 months after surgery for most patients. Late cognitive decline after CABG,occurring between 1 and 5 years after the surgery, has been well documented,but controlled studies demonstrating that this decline is specifically attribut-able to the use of cardiopulmonary bypass itself rather than to progression ofunderlying cerebrovascular disease or other age-related changes are pending.

Acknowledgments

The authors thank the staff and participants in this study and are partic-ularly grateful for the ongoing collaboration with Dr. W. Baumgartner andother participating cardiac surgeons and cardiologists.

References

[1] Johnson RG. Abnormal neuropsychometrics early after coronary artery bypass grafting.

Crit Care Med 2000;28(6):2142–3.

[2] NewmanMF,Kirchner JL, Phillips-ButeB, et al. Longitudinal assessment of neurocognitive

function after coronary artery bypass surgery. N Engl J Med 2001;344:395–402.

[3] Keith JR, Puente AE,MalcolmsonKL, et al. Assessing postoperative cognitive change after

cardiopulmonary bypass surgery. Neuropsychology 2002;16(3):411–21.

[4] Selnes OA, Grega MA, Borowicz LM Jr, et al. Cognitive changes with coronary artery dis-

ease: a prospective study of coronary artery bypass graft patients and nonsurgical controls.

Ann Thorac Surg 2003;75(5):1377–84.

[5] Mullges W, Babin-Ebell J, Reents W, et al. Cognitive performance after coronary artery

bypass grafting: a follow-up study. Neurology 2002;59:741–3.

[6] Mullges W, Berg D, Schmidtke A, et al. Early natural course of transient encephalopathy

after coronary artery bypass grafting. Crit Care Med 2000;28(6):1808–11.

[7] Selnes OA, GoldsboroughMA, Borowicz LM, et al. Determinants of cognitive change after

coronary artery bypass surgery: a multifactorial problem. Ann Thorac Surg 1999;67:

1669–76.

[8] Bergh C, Backstrom M, Jonsson H, et al. In the eye of both patient and spouse: memory is

poor 1 to 2 years after coronary bypass and angioplasty. Ann Thorac Surg 2002;74(3):

689–93.


[9] Bassel C, Rourke SB, HalmanMH, et al. Working memory performance predicts subjective

cognitive complaints in HIV infection. Neuropsychology 2002;16(3):400–10.

[10] Goto T, Baba T,MatsuyamaK, et al. Aortic atherosclerosis and postoperative neurological

dysfunction in elderly coronary surgical patients. Ann Thorac Surg 2003;75(6):1912–8.

[11] Mullges W, Franke D, Reents W, et al. Brain microembolic counts during extracorporeal

circulation depend on aortic cannula position. Ultrasound Med Biol 2001;27(7):933–6.

[12] Abu-Omar Y, Balacumaraswami L, Pigott DW, et al. Solid and gaseous cerebral microem-

bolization during off-pump, on-pump, and open cardiac surgery procedures. J Thorac Car-

diovasc Surg 2004;127(6):1759–65.

[13] Clark RE, Brillman J, Davis DA, et al. Microemboli during coronary artery bypass grafting.

Genesis and effect on outcome. J Thorac Cardiovasc Surg 1995;109:249–57 [discussion:

257–8].

[14] Fearn SJ, Pole R, Wesnes K, et al. Cerebral injury during cardiopulmonary bypass: emboli

impair memory. J Thorac Cardiovasc Surg 2001;121(6):1150–60.

[15] Braekken SK, Reinvang I, Russell D, et al. Association between intraoperative cerebral mi-

croembolic signals and postoperative neuropsychological deficit: comparison between pa-

tients with cardiac valve replacement and patients with coronary artery bypass grafting.

J Neurol Neurosurg Psychiatry 1998;65(4):573–6.

[16] Neville MJ, Butterworth J, James RL, et al. Similar neurobehavioral outcome after valve or

coronary artery operations despite differing carotid embolic counts. J Thorac Cardiovasc

Surg 2001;121(1):125–36.

[17] Browndyke JN, Moser DJ, Cohen RA, et al. Acute neuropsychological functioning follow-

ing cardiosurgical interventions associated with the production of intraoperative cerebral

microemboli. Clin Neuropsychol 2002;16(4):463–71.

[18] Andrell P, Jensen C, Norrsell H, et al. White matter disease in magnetic resonance imaging

predicts cerebral complications after coronary artery bypass grafting. Ann Thorac Surg

2005;79(1):74–9.

[19] Ho PM, Arciniegas DB, Grigsby J, et al. Predictors of cognitive decline following coronary

artery bypass graft surgery. Ann Thorac Surg 2004;77(2):597–603.

[20] BrownWR,MoodyDM, Challa VR. Cerebral fat embolism from cardiopulmonary bypass.

J Neuropathol Exp Neurol 1999;58(2):109–19.

[21] Emmrich P, Hahn J, Ogunlade V, et al. Neuropathological findings after cardiac surgeryd

retrospective study over 6 years. Z Kardiol 2003;92(11):925–37.

[22] Zangrillo A, Landoni G, SparicioD, et al. Predictors of atrial fibrillation after off-pump cor-

onary artery bypass graft surgery. J Cardiothorac Vasc Anesth 2004;18(6):704–8.

[23] Mathew JP, Fontes ML, Tudor IC, et al. A multicenter risk index for atrial fibrillation after

cardiac surgery. JAMA 2004;291(14):1720–9.

[24] Crystal E, Connolly SJ. Atrial fibrillation: guiding lessons from epidemiology. Cardiol Clin

2004;22(1):1–8.

[25] Lahtinen J, Biancari F, Salmela E, et al. Postoperative atrial fibrillation is a major cause of

stroke after on-pump coronary artery bypass surgery. Ann Thorac Surg 2004;77(4):1241–4.

[26] Stanley TO, Mackensen GB, Grocott HP, et al. The impact of postoperative atrial fibrilla-

tion on neurocognitive outcome after coronary artery bypass graft surgery. Anesth Analg

2002;94(2):290–5.

[27] AbildstromH,Hogh P, Sperling B, et al. Cerebral blood flow and cognitive dysfunction after

coronary surgery. Ann Thorac Surg 2002;73(4):1174–8.

[28] Caplan LR, Hennerici M. Impaired clearance of emboli (washout) is an important link be-

tween hypoperfusion, embolism, and ischemic stroke. Arch Neurol 1998;55(11):1475–82.

[29] Johnson T, Monk T, Rasmussen LS, et al. Postoperative cognitive dysfunction in middle-

aged patients. Anesthesiology 2002;96(6):1351–7.

[30] Rasmussen LS, Johnson T, Kuipers HM, et al. Does anaesthesia cause postoperative cogni-

tive dysfunction? A randomised study of regional versus general anaesthesia in 438 elderly

patients. Acta Anaesthesiol Scand 2003;47(3):260–6.

144 SELNES et al

[31] Vinkers DJ, Gussekloo J, Stek ML, et al. Temporal relation between depression and cogni-

tive impairment in old age: prospective population based study. BMJ 2004;329(7471):881.

[32] McKhann GM, Borowicz LM, Goldsborough MA, et al. Depression and cognitive decline

after coronary artery bypass grafting. Lancet 1997;349:1282–4.

[33] Tardiff BE, NewmanMF, Saunders AM, et al. Preliminary report of a genetic basis for cog-

nitive decline after cardiac operations. The Neurologic Outcome Research Group of the

Duke Heart Center. Ann Thorac Surg 1997;64:715–20.

[34] Selnes OA, Royall RM, Grega MA, et al. Cognitive changes 5 years after coronary artery

bypass grafting: is there evidence of late decline? Arch Neurol 2001;58(4):598–604.

[35] Steed L, Kong R, Stygall J, et al. The role of apolipoprotein E in cognitive decline after car-

diac operation. Ann Thorac Surg 2001;71(3):823–6.

[36] Stygall J, Newman SP, Fitzgerald G, et al. Cognitive change 5 years after coronary artery

bypass surgery. Health Psychol 2003;22(6):579–86.

[37] Denton TA, FonarowGC, LaBresh KA, et al. Secondary prevention after coronary bypass:

the American Heart Association ‘‘get with the guidelines’’ program. Ann Thorac Surg 2003;

75(3):758–60.

[38] HlatkyMA, BaconC, BoothroydD, et al. Cognitive function 5 years after randomization to

coronary angioplasty or coronary artery bypass graft surgery. Circulation 1999;96(Suppl II):

11–5.

[39] Potter GG, Plassman BL, Helms MJ, et al. Age effects of coronary artery bypass graft on

cognitive status change among elderly male twins. Neurology 2004;63(12):2245–9.

[40] BennettHP, Piguet O,GraysonDA, et al. A 6-year study of cognition and spatial function in

the demented and non-demented elderly: the Sydney Older Persons Study. Dement Geriatr

Cogn Disord 2003;16(4):181–6.

[41] Piguet O, Grayson DA, Creasey H, et al. Vascular risk factors, cognition and dementia in-

cidence over 6 years in the Sydney Older Persons Study. Neuroepidemiology 2003;22(3):

165–71.

[42] Saxton J, Ratcliff G, Newman A, et al. Cognitive test performance and presence of subclin-

ical cardiovascular disease in the cardiovascular health study. Neuroepidemiology 2000;

19(6):312–9.

[43] Elwood PC, Pickering J, Bayer A, et al. Vascular disease and cognitive function in older men

in the Caerphilly cohort. Age Ageing 2002;31(1):43–8.

[44] Knopman D, Boland LL, Mosley T, et al. Cardiovascular risk factors and cognitive decline

in middle-aged adults. Neurology 2001;56(1):42–8.

[45] FontbonneA, BerrC,Ducimetiere P, et al. Changes in cognitive abilities over a 4-year period

are unfavorably affected in elderly diabetic subjects: results of the Epidemiology of Vascular

Aging Study. Diabetes Care 2001;24(2):366–70.

[46] Logroscino G, Kang JH, Grodstein F. Prospective study of type 2 diabetes and cognitive

decline in women aged 70–81 years. BMJ 2004;328(7439):548.

[47] Hassing LB, Hofer SM, Nilsson SE, et al. Comorbid type 2 diabetes mellitus and hyperten-

sion exacerbates cognitive decline: evidence from a longitudinal study. Age Ageing 2004;

33(4):355–61.

[48] Hebert LE, Scherr PA, Bennett DA, et al. Blood pressure and late-life cognitive function

change: a biracial longitudinal population study. Neurology 2004;62(11):2021–4.

[49] Vermeer SE, Prins ND, DenHeijer T, et al. Silent brain infarcts and the risk of dementia and

cognitive decline. N Engl J Med 2003;348(13):1215–22.

[50] Schneider JA,WilsonRS,CochranEJ, et al. Relation of cerebral infarctions to dementia and

cognitive function in older persons. Neurology 2003;60(7):1082–8.

[51] Mathiesen EB, Waterloo K, Joakimsen O, et al. Reduced neuropsychological test perfor-

mance in asymptomatic carotid stenosis: the Tromso Study. Neurology 2004;62(5):695–701.

[52] de Groot JC, de Leeuw FE, Oudkerk M, et al. Cerebral white matter lesions and cognitive

function: the Rotterdam Scan Study. Ann Neurol 2000;47(2):145–51.


[53] Goto T, Baba T, Honma K, et al. Magnetic resonance imaging findings and postoperative

neurologic dysfunction in elderly patients undergoing coronary artery bypass grafting.

Ann Thorac Surg 2001;72(1):137–42.

[54] NakamuraY,Kawachi K, ImagawaH, et al. The prevalence and severity of cerebrovascular

disease in patients undergoing cardiovascular surgery. Ann Thorac Cardiovasc Surg 2004;

10(2):81–4.

[55] Bainbridge D, Martin J, Cheng D. Off pump coronary artery bypass graft surgery versus

conventional coronary artery bypass graft surgery: a systematic review of the literature.

Semin Cardiothorac Vasc Anesth 2005;9(1):105–11.

[56] Lund C, Hol PK, Lundblad R, et al. Comparison of cerebral embolization during off-pump

and on-pump coronary artery bypass surgery. Ann Thorac Surg 2003;76(3):765–70.

[57] Bowles BJ, Lee JD, Dang CR, et al. Coronary artery bypass performed without the use of

cardiopulmonary bypass is associatedwith reduced cerebralmicroemboli and improved clin-

ical results. Chest 2001;119(1):25–30.

[58] van Dijk D, Jansen EW, Hijman R, et al. Cognitive outcome after off-pump and on-pump

coronary artery bypass graft surgery: a randomized trial. JAMA 2002;287(11):1405–12.

Neurol Clin 24 (2006) 147–158

Brain Injury from CardiacArrest in Children

Robert W. Hickey, MDa,*, Michael J. Painter, MDb

aDivision of Pediatric Emergency Medicine, Department of Pediatrics,

University of Pittsburgh, Children’s Hospital of Pittsburgh, 3705 Fifth Avenue,

Pittsburgh, PA 15213, USAbDivision of Child Neurology, Department of Pediatrics, University of Pittsburgh,

Children’s Hospital of Pittsburgh, 3705 Fifth Avenue, Pittsburgh, PA 15213, USA

Many of the features of postischemic brain injury in children are similarto injury in adults; thus, much of this issue of the Neurologic Clinics of NorthAmerica applies to children and adults. There are two important differences,however, that merit a separate section focused on pediatric injury. First, themechanism of cardiac arrest in children differs, with respiratory causes faroutnumbering cardiac causes (Tables 1 and 2). Second, the developing brainhas different vulnerability and potential for repair compared with the ma-ture brain. This article reviews these differences and the available clinicaldata relevant to pediatric brain injury following cardiac arrest.

Asphyxial cardiac arrest

The most common cause of nontraumatic cardiopulmonary arrest in chil-dren is airway compromise [1–3]. Although ventricular fibrillation (VF) orventricular tachycardia (VT) occurs less commonly in children than inadults, it is not rare: approximately 5% to 15% of children with prehospitalarrest have VF/VT [4–6].

Asphyxia can be clinically defined as airway obstruction or inadequateventilation leading to hypoxemia and hypercarbia. Examples includedrowning, choking, and coma accompanied by loss of airway patency.The typical progression of untreated asphyxia is hypertension and increased

This work was supported by the National Institutes of Health grant NIH KO8 HD40848

to Robert W. Hickey.


E-mail address: [email protected] (R.W. Hickey).




148 HICKEY & PAINTER

work of breathing (where possible), followed by bradycardia, hypotension,pulseless electrical activity, and eventually, asystole.

Although both VF and asphyxial cardiac arrest result in global brainischemia, the pattern of ischemia differs (Table 3). VF causes an abrupt ces-sation of cardiac output, whereas asphyxia causes an initial hypertension,followed by a gradual decrease in flow until pulseless electrical activityand, finally, asystole occur. Paradoxically, although low cerebral bloodflow is better than no flow, a ‘‘trickle’’ of flow can be worse than no flow.This phenomenon was demonstrated in a study by Bottiger and colleagues[7] that showed worse postresuscitation cerebral reperfusion in rats thathad 12-minute untreated VF plus 5-minute VF treated with cardiopulmo-nary resuscitation compared with rats subjected to 17-minute untreatedVF. Theories for the damaging effect of trickle flow include (1) the contin-ued delivery of substrate during conditions of anaerobic metabolism, caus-ing worse tissue acidosis; and (2) the continued delivery of platelets andcoagulation factors, causing worse microvascular plugging that would

Table 1

Etiology of out-of-hospital cardiac arrest in children

Cause of arrest n (%)

Sudden infant death syndrome 136 (23)

Trauma 118 (20)

Respiratory 96 (16)

Submersion 73 (12)

Cardiac 48 (8)

Central nervous system 35 (6)

Burn 6 (1)

Poisoning 6 (1)

Other 63 (10)

Unknown 20 (3)

Data from Young KD, Gausche-Hill M, McClung CD, et al. A prospective, population-

based study of the epidemiology and outcome of out-of-hospital pediatric cardiopulmonary

arrest. Pediatrics 2004;114(1):157–64.

Table 2

Characteristics of children with in-hospital cardiac arrest

Patient characteristic n (%)

Cardiac arrest 176 (100)

No CPR (terminal phase of chronic disease) 47 (27)

CPR performed 129 (73)

Chronic disease in subset with CPR 92 (71)

Respiratory failure in subset with CPR 79 (61)

Circulatory shock in subset with CPR 37 (29)

Abbreviation: CPR, cardiopulmonary resusciation.

Data from Reis AG, Nadkarni V, Perondi MB, et al. A prospective investigation into the

epidemiology of in-hospital pediatric cardiopulmonary resuscitation using the international

Utstein reporting style. Pediatrics 2002;109(2):200–9.

149BRAIN INJURY IN CHILDREN

impair reperfusion during resuscitation. Asphyxia, but not VF/VT, has aninterval of trickle cerebral blood flow accompanied by profound hypoxemia.

The histology of cerebral injury following asphyxia differs from that seenin VF. Safar and colleagues [8,9] showed that brain damage from asphyxialcardiac arrest in dogs is worse than the damage found after equivalentperiods of circulatory arrest from VF. In addition, asphyxiated brains hadscattered microinfarcts and hemorrhage not seen in VF animals. Thus, lab-oratory experiments demonstrate that the severity and pattern of cerebralinjury following asphyxial cardiac arrest differs from VF arrest. Clinical ev-idence of a difference in injury patterns is suggested by a report from Mor-imoto and colleagues [10] that described increased prevalence of brainedema (diagnosed by head CT) in adults remaining comatose followingrespiratory-induced cardiac arrest compared with cardiac arrhythmia–induced cardiac arrest.

Do the differences between asphyxial brain injury and ‘‘cardiac-mediated’’brain injury have clinical relevance? They do to the extent that asphyxialinjuries are more severe. Both injuries, however, demonstrate selective vul-nerability and delayed neuronal death. Specifically, both mechanisms causecell death that is ‘‘delayed’’ and is first seen on histology at 24 to 72 hoursfollowing reperfusion. The most prominent of these ‘‘selectively vulnera-ble’’ regions are the hippocampus and reticular thalamus. Thus, althoughan asphyxial injury may be more severe than a cardiac-mediated injury foran equivalent period of ischemia, asphyxial injuries should respond simi-larly to neuroprotective therapies.

The developing brain

Brain maturation entails a complex coordination of neuronal prolifera-tion, migration, synaptic overgrowth, pruning, and myelination. Althoughproliferation and migration are complete in humans at birth, the remainingprocesses continue into early adulthood, with completion of the myelination

Table 3

Comparison of injury from ventricular fibrillation versus asphyxial cardiac arrest

Injury Ventricular fibrillation Asphyxial cardiac arrest

Postresuscitation cardiac injury Relatively more Relatively less

Postresuscitation cerebral injury Relatively less Relatively more

Cerebral blood flow Sudden complete

ischemia

Trickle flow prior to

complete ischemia

Scattered microinfarcts No Yes

Injury to basal ganglia Relatively less Relatively more

Selective vulnerability of CA1

hippocampus

Yes Yes

Selective vulnerability of

cerebellar Purkinje’s cells

Yes Yes

Comparisons are extrapolated from animal experiments.


of long association pathways occurring in the third decade of life. Accord-ingly, the brain’s vulnerability to injury is not constant across different ageranges [11]. For example, rodents are extraordinarily resistant to ischemicinjury when first born and then go through a period of increased sensitivityto injury, followed by intermediate sensitivity that lasts into adulthood[12–15]. The period of increased vulnerability correlates with maturationof receptors for excitatory neurotransmitters and maximal synaptogenesis.Laboratory data also show that immature neurons and oligodendrocyteshave a lower threshold for initiating programmed cell death (apoptosis)compared with mature cells [16–19]. These developmental events that deter-mine susceptibility to brain injury in laboratory models also occur duringnormal human development [20]. Synaptogenesis in human striatal cortexaccelerates between age 2 and 4 months, creating a condition of exuberantconnectivity that is subsequently pruned by 40% between age 8 monthsand 11 years [21]. A second wave of synaptic formation and pruning, pri-marily in the frontal cortex, has now been identified in adolescence [22].In addition, 31P magnetic resonance spectroscopy shows that the metabolicrate for local cerebral regions is 190% to 226% of adult levels between age3 and 8 years, and there is a peak in the phosphomonoester spectrum that isindicative of active myelination just before age 2 years [23].

Functional development

Coincident tomaturation seen at the tissue and cellular layer, the functionalcapabilities of the brain also mature over time. The acquisitions of grossmotor, fine motor, and cognitive skills during youth are well-recognizedphenomena. Beyond early youth, brain growth continues into the twentiesand is associated with maturation of executive functions. Perhaps not surpris-ingly, functional imaging studies show that risk-taking during adolescence isassociated with an immature pattern of cerebral activity compared with adultactivity.

Some skills can only be acquired during specific periods of development(‘‘use it or lose it’’). For example, temporary monocular occlusion in childrenat the time when visual cortical pathways are being established can result inpermanent cortical blindness in the occluded eye. Likewise, patching the‘‘good eye’’ in patients with strabismus improves vision in the ‘‘weak eye’’only when patching is initiated at a young age. A more prosaic example isthe ability of young children to speak a second language without accent.

There is also an age-dependent capacity for repair [24]. A dramatic exampleof plasticity is the recovery that can occur in young children, but not adults,undergoing hemispherectomy for intractable seizures [25,26]. Children whohave had hemispherectomy are more capable than adults of reallocatingfunctions from the removed hemisphere to the remaining hemisphere.Similarly, language deficits are less common in children suffering injury tothe dominant hemisphere before age 8 years.


The impact of global brain ischemia on functional outcome during ‘‘de-velopmentally sensitive’’ periods of skill acquisition and the role of plasticityin recovery from global ischemia has not been systematically studied. Onechallenge is that injury obtained in young patients may not become apparentuntil the functional correlates are required for normal behavior and are test-able. For example, children who have learning disabilities are often not di-agnosed until the requisite learning skills become necessary for schoolperformance. A greater understanding of the dynamic between injury anddevelopment has considerable potential benefit for children who have ische-mic brain injury.

Clinical experience

Young and Seidel [27] recently summarized the results from 44 studies re-porting on 3094 pediatric patients with cardiopulmonary arrest. The datashowed an overall survival rate from cardiopulmonary arrest of 13%,with in-hospital arrest rates being higher than out-of-hospital arrest rates(24% versus 9%). Most of the reviewed studies reported good neurologicoutcome in approximately 60% of survivors; however, comparison betweenstudies is difficult because of the differences in inclusion criteria and the def-inition of ‘‘good’’ neurologic outcome. Neurologic outcome assessmentsthat target motor function and rudimentary life-skill tasks suggest thatmost patients have full recovery or severe disability [28–35]. Patients whohave poor outcome have generally suffered a severe, acute asphyxial event.Among children with good neurologic outcome, assessments that measureIQ or psychocognitive function often reveal impaired performance [36–40]. Robertson and colleagues [39] recently published data on a cohort of53 children younger than age 3 years admitted to an ICU who had traumaticbrain injury (n ¼ 26) or hypoxic ischemic (HI) brain injury (n ¼ 27) and aninitial Glasgow Coma Scale of 8 or less. Of the 23 children identified as hav-ing ‘‘good recovery’’ based on the Glasgow Outcome Scale, 15 (65%) hadbelow-average scores on the Mental Developmental Index or PerformanceDevelopmental Index. These indices, however, have imperfect correlationwith later cognitive assessments and it is possible that additional recoverymay occur or that deficits may remain ‘‘dormant’’ until uncovered by in-creasingly complex cognitive demands associated with maturation. Addi-tional studies are desirable to better define long-term psychocognitiveoutcome.

Specific patterns of functional deficits have been described; notably,memory deficits in adults who have global brain injury [41,42] and cerebralpalsy in neonatal asphyxia [43,44]. Similarly, the neurodevelopmental out-comes of premature infants have been well characterized in a meta-analysisof 227 studies by Bhutta and colleages [45]. The infrastructure that directshigh-risk neonatal ICU graduates into comprehensive assessment/treatmentprograms, however, does not exist for older children who have HI injury


and, thus, the functional outcome of children who have HI injury is less wellcharacterized.

Prognosis

Recently, Mandel and colleagues [46] reported on clinical and electro-physiologic predictors of outcome in 42 pediatric patients who remainedcomatose or had impaired consciousness at 24 hours following an HI injury.Twelve patients had an eventual good outcome, 4 had mild to moderatedisability, 7 had severe disability or survival in a persistent vegetative state,and 19 ultimately died (9 with brain death, 2 after failed repeated resuscita-tion attempts, 8 after withdrawal of therapy). The positive predictive valuefor poor outcome (severe disability, persistent vegetative state, or death;n ¼ 26) was 91% for duration of initial cardiopulmonary resuscitation ex-ceeding 10 minutes and 100% for (1) Glasgow Coma Scale scores lessthan 5, (2) absence of spontaneous respirations, or (3) absence of pupillaryreflex at 24 hours. The positive predictive value for poor outcome was 100%for discontinuous electroencephalographic activity, epileptiform electroen-cephalographic activity, or bilateral absent N20 latency on sensory evokedpotential. This study is in agreement with other studies that have found neu-rologic examination and electrophysiology studies to be good predictors ofoutcome [47]. The main limitations of many of these studies are the smallsample sizes and the post hoc derivation of decision rules.

Imaging

Neuroimaging techniques have identified specific patterns of cerebral in-jury in adults and newborns with ischemic injury. Data from human neonatesand experimental primate models of neonatal asphyxia reveal that imagingabnormalities correlate with the nature and duration of the insult and thematurational stage of the brain at time of injury [48]. The immature brain ofpremature neonates is more vulnerable to white matter injury, whereas thebrain of term neonates is more vulnerable to gray matter injury [49]. Acute,total asphyxia tends to result in greater injury to the brainstem and thalamus,whereas prolonged, partial asphyxia results in greater injury to the cortex andsubcortical regions [50]. Even so, full-term infants who have prolonged, par-tial ischemia have a different pattern of injury compared with prematureinfants who have prolonged, partial ischemia. Thus, there are differentpatterns of HI brain injury relative to gestational age in newborns. What isnot known is whether there are different patterns of injury relative to age inchildren outside of the newborn period suffering ischemia.

MRI can be used to measure regional volume (morphometrics) and thuscharacterize brain injury/recovery during follow-up of HI injury. A series of17 adults surviving cardiac arrest studied by MRI at 6 months following re-suscitation reported reduced hippocampal volume and a global reduction in


brain volume [51]. The reduction in hippocampal volume is consistent withthe specific cognitive memory impairments commonly documented in adultsfollowing cardiac arrest [52]. The reduction in global brain volume is consis-tent with the more widespread deficits in memory, visuospatial, and executivefunctions that have also been documented. Similarly, follow-upMRI for for-mer low birth weight preterm infants demonstrates smaller regional corticalvolumes [53] and selective loss of hippocampal volume [54] compared withcontrol subjects. The volume losses in these former preterm infants is corre-lated with memory performance and full-scale, verbal, and performance IQscores. Again, information on the predictive value and long-term changesseen in MRI imaging of older children who have HI is limited [55].

Kreis and colleagues [56] used proton spectroscopy to study 16 childrensuffering near-drowning from age 7 months to 6 years. Loss of N-acetylas-partate from gray matter preceded the loss observed in white matter and wasmore severe. There was a delayed second peak of lactate, similar to the de-layed secondary energy failure documented in neonatal HI. A spectroscopicindex was derived that predicted neurologic outcome in this small series withgreater accuracy than published clinical criteria.

A contemporary variation of diffusion-weighted MRI (diffusion tensorimaging; DTI) analyzes vector forces of diffusion patterns. Diffusion patternsare highly dependent on development and orientation of axonal fibers andoligodendroglia; thus, DTI is a sensitive tool for detecting white matter devel-opment (myelination) and injury. DTI scans obtained from premature infantswho have white matter injury demonstrate disorganized vector forces consis-tent with disrupted white matter tract development [57]. Furthermore, whitematter injury has been shown to correlate with diminished volume in the asso-ciated graymatter. DTI studies on older childrenwho haveHI injury have notbeen reported; however, DTI andMRI studies of normal children at differentages confirm that myelination continues through the second decade of life inan age-dependent, region-specific fashion [58–62]. Thus, it is likely that pat-terns of white matter injury vary by age. If so, DTI may be useful for identify-ing white matter injury and directing rehabilitative therapy to the associatedcortical (and functional) brain regions.

Treatment

Following resuscitation from cardiac arrest, there is a period of increasedsensitivity of the brain to secondary injury. A review by Kochanek and col-leagues [63] provided the known precipitants of secondary injury, which in-clude hypotension, hypoxia, hyperglycemia, and hyperthermia. Earlypostresuscitative care should focus on avoiding these causes of secondaryinjury (Box 1). Detailed discussion of intensive care therapy is beyond thescope of this article. Instead, discussion is limited to hyperventilation andhypothermia (hyperventilation because it is a pervasive problem and hypo-thermia because it is the most promising neuroprotective strategy).


One preventable cause of secondary injury is iatrogenic hyperventilation.Hyperventilation has been shown to cause vasoconstriction and significantlydecreased cerebral blood flow in children following traumatic brain injury [64]and in adults recovering from cardiac arrest [65]. Hyperventilation can alsodecrease cerebral blood flow by increasing intrathoracic pressure, causinga decrease in cardiac output and cerebral venous return. In addition, respira-tory alkalosis shifts the oxygen hemoglobin dissociation curve to the left,reducing oxygen delivery to tissue. These alterations are particularly danger-ous early after resuscitation when there is prolonged, multifocal decreasedcerebral blood flow [66]. Avoidance of hyperventilation is challengingdcaregivers under stressful circumstances unintentionally but predictablyhyperventilate patients [67,68]. Tobias and colleagues [69] published a studyon pediatric patients transported from the ICU to the radiology suite bynurses and respiratory therapists blinded to end tidal CO2 values: 23% ofreadings were less than 20 torr. Increased use of quantitative continuousCO2monitors throughout the health care systemwould decrease the potentialfor harm secondary to inadvertent hyperventilation.

Measurement and control of temperature following cardiac arrest is animportant part of patient management. After arrest, children commonlyhave an initial period of spontaneous hypothermia followed by a delayed(approximately 24 hours) development of fever [69a]. These temperaturechanges are relevant because hypothermia is neuroprotective, whereas hy-perthermia can exacerbate brain injury. Accordingly, routine warming ofpatients during initial hypothermia is no longer recommended. Rewarmingcan negate the neuroprotective effects of hypothermia and may cause an‘‘overshoot’’ of temperature that contributes to subsequent fever. Intentionalinduction or maintenance of hypothermia (therapeutic hypothermia) hasrecently been shown to be beneficial in adults recovering from cardiac ar-rest and in newborns recovering from birth asphyxia [70–73]. Although thestudies in adults excluded asphyxia (enrollment was limited to patientswho had VF/VT), there are significant animal data to support the use ofhypothermia in asphyxial arrest [74]. Thus, consideration should be givento actively cooling children who remain comatose following resuscitation

Box 1. Postresuscitation treatment priorities

� Avoid hypotension� Maintain normoxia (avoid hypoxia and prolonged hyperoxia)� Maintain euglycemia (avoid hyperglycemia and hypoglycemia)� Avoid hyperventilation� Avoid hyperthermia� Avoid rewarming� Consider induced hypothermia


from cardiac arrest. In addition, temperature should be monitored closelyand fever should be treated aggressively.

Knowledge gaps and future directions

There is an accumulating literature on neurologic outcome in adultsresuscitated from cardiac arrest and newborns recovering from perinatal as-phyxia. In contrast, there is very little information on children resuscitatedfrom cardiac arrest. Animal models showing age-dependent susceptibility toinjury and clinical data showing age-dependent windows for learning andplasticity suggest that extrapolating from neonatal or adult experience willbe imperfect. Thus, there is a critical need for studies targeting the pediatricage range between these populations. Important areas of inquiry include

� Age-dependent susceptibilities for injury and repair� Contemporary imaging strategies targeting white matter development,morphometric measurements, and functional imaging

� Clinical or laboratory markers for severity of the initial event� Role of antiapoptotic neuroprotective strategies in children� Induced hypothermia� Rehabilitation strategies (eg, enriched environment, forced use) thattarget age-dependent injury/repair susceptibilities

Because of the infrequent occurrence of pediatric cardiac arrest and thenumber of confounding variables, advances in understanding will likely re-quire multicenter and interdisciplinary collaborations. Although studies ofbrain injury in children across a range of developmental stages will be chal-lenging, they will also be unique opportunities to increase our understandingof brain development, learning, and plasticity.

References

[1] KuismaM,SuominenP,KorpelaR.Paediatric out-of-hospital cardiac arrestsdepidemiology

and outcome. Resuscitation 1995;30(2):141–50.

[2] Sirbaugh PE, Pepe PE, Shook JE, et al. A prospective, population-based study of the demo-

graphics, epidemiology, management, and outcome of out-of-hospital pediatric cardiopul-

monary arrest. Ann Emerg Med 1999;33(2):174–84.

[3] Young KD, Gausche-Hill M, McClung CD, et al. A prospective, population-based study of

the epidemiology and outcome of out-of-hospital pediatric cardiopulmonary arrest. Pediat-

rics 2004;114(1):157–64.

[4] Hickey RW, Cohen DM, Strausbaugh S, et al. Pediatric patients requiring CPR in the pre-

hospital setting. Ann Emerg Med 1995;25:495–501.

[5] Appleton GO, Cummins RO, Larson MP, et al. CPR and the single rescuer: at what age

should you ‘‘call first’’ rather than ‘‘call fast?’’. Ann Emerg Med 1995;25(4):492–4.

[6] Mogayzel C, Quan L, Graves JR, et al. Out-of-hospital ventricular fibrillation in children

and adolescents: causes and outcomes. Ann Emerg Med 1995;25(4):484–91.

[7] Bottiger BW, Krumnikl JJ, Gass P, et al. The cerebral ‘no-reflow’ phenomenon after cardiac

arrest in ratsdinfluence of low-flow reperfusion. Resuscitation 1997;34(1):79–87.


[8] Vaagenes P, Safar P, Diven W, et al. Brain enzyme levels in CSF after cardiac arrest and re-

suscitation in dogs: markers of damage and predictors of outcome. J Cereb Blood Flow

Metab 1988;8(2):262–75.

[9] Vaagenes P, Safar P, Moossy J, et al. Asphyxiation versus ventricular fibrillation cardiac ar-

rest in dogs. Differences in cerebral resuscitation effectsda preliminary study. Resuscitation

1997;35(1):41–52.

[10] Morimoto Y, Kemmotsu O, Kitami K, et al. Acute brain swelling after out-of-hospital car-

diac arrest: pathogenesis and outcome. Crit Care Med 1993;21:104–10.

[11] Rice D, Barone S Jr. Critical periods of vulnerability for the developing nervous system:

evidence from humans and animal models. Environ Health Perspect 2000;108(Suppl 3):

511–33.

[12] Ikonomidou C, Mosinger JL, Salles KS, et al. Sensitivity of the developing rat brain to

hypobaric/ischemic damage parallels sensitivity to N-methyl-aspartate neurotoxicity.

J Neurosci 1989;9(8):2809–18.

[13] Yager JY, Shuaib A, Thornhill J. The effect of age on susceptibility to brain damage in a

model of global hemispheric hypoxia-ischemia. Brain Res Dev Brain Res 1996;93(1–2):

143–54.

[14] Grafe MR. Developmental changes in the sensitivity of the neonatal rat brain to hypoxic/

ischemic injury. Brain Res 1994;653(1–2):161–6.

[15] Towfighi J, Mauger D, Vannucci RC, et al. Influence of age on the cerebral lesions in an im-

mature rat model of cerebral hypoxia-ischemia: a light microscopic study. Brain Res Dev

Brain Res 1997;100(2):149–60.

[16] McDonald JW, Behrens MI, Chung C, et al. Susceptibility to apoptosis is enhanced in im-

mature cortical neurons. Brain Res 1997;759(2):228–32.

[17] Bittigau P, Sifringer M, Pohl D, et al. Apoptotic neurodegeneration following trauma is

markedly enhanced in the immature brain. Ann Neurol 1999;45(6):724–35.

[18] Hu BR, Liu CL, Ouyang Y, et al. Involvement of caspase-3 in cell death after hypoxia-

ischemia declines during brainmaturation. J CerebBlood FlowMetab 2000;20(9):1294–300.

[19] Ness JK, Romanko MJ, Rothstein RP, et al. Perinatal hypoxia-ischemia induces apoptotic

and excitotoxic death of periventricular whitematter oligodendrocyte progenitors.DevNeu-

rosci 2001;23(3):203–8.

[20] Greenamyre T, Penney JB, Young AB, et al. Evidence for transient perinatal glutamatergic

innervation of globus pallidus. J Neurosci 1987;7(4):1022–30.

[21] Huttenlocher PR, de Courten C. The development of synapses in striate cortex of man. Hum

Neurobiol 1987;6(1):1–9.

[22] Giedd JN, Blumenthal J, Jeffries NO, et al. Brain development during childhood and adoles-

cence: a longitudinal MRI study. Nat Neurosci 1999;2(10):861–3.

[23] Hanaoka S, Takashima S, Morooka K. Study of the maturation of the child’s brain using

31P-MRS. Pediatr Neurol 1998;18(4):305–10.

[24] Stiles J. Neural plasticity and cognitive development. Dev Neuropsychol 2000;18(2):237–72.

[25] Bittar RG, Rosenfeld JV, Klug GL, et al. Resective surgery in infants and young children

with intractable epilepsy. J Clin Neurosci 2002;9(2):142–6.

[26] Vining EP, Freeman JM, Pillas DJ, et al. Why would you remove half a brain? The outcome

of 58 children after hemispherectomydthe Johns Hopkins experience: 1968 to 1996. Pediat-

rics 1997;100(2 Pt 1):163–71.

[27] Young KD, Seidel JS. Pediatric cardiopulmonary resuscitation: a collective review. Ann

Emerg Med 1999;33(2):195–205.

[28] Dean JM, Kaufman ND. Prognostic indicators in pediatric near-drowning: the Glasgow

coma scale. Crit Care Med 1981;9(7):536–9.

[29] O’Rourke PP. Out-of-hospital cardiac arrest in pediatric patients. Crit Care Med 1984;12:

283.

[30] O’Rourke PP.Outcome of childrenwho are apneic and pulseless in the emergency room.Crit

Care Med 1986;14:466–8.


[31] Davies CR, Carrigan T, Wright JA. Neurologic outcome following pediatric resuscitation.

Neuroscience 1987;19:205–10.

[32] Quan L,WentzKR,Gore EJ, et al. Outcome and predictors of outcome in pediatric submer-

sion victims receiving prehospital care in King County, Washington. Pediatrics 1990;86(4):

586–93.

[33] Kriel RL, Krach LE, Luxenberg MG, et al. Outcome of severe anoxic/ischemic brain injury

in children. Pediatri Neurol 1994;10:207–12.

[34] Torres A Jr, Pickert CB, Firestone J, et al. Long-term functional outcome of inpatient pedi-

atric cardiopulmonary resuscitation. Pediatr Emerg Care 1997;13(6):369–73.

[35] Spack L, Gedeit R, Splaingard M, et al. Failure of aggressive therapy to alter outcome in

pediatric near-drowning. Pediatr Emerg Care 1997;13(2):98–102.

[36] Pearn J. Neurological and psychometric studies in children surviving freshwater immersion

accidents. Lancet 1977;1:7–9.

[37] Bell TS, Ellenberg L,McComb JG.Neuropsychological outcome after severe pediatric near-

drowning. Neurosurgery 1985;17:604–8.

[38] Morris RD, Krawiecki NS, Wright JA. Neuropsychological, academic, and adaptive func-

tioning in children who survive in-hospital cardiac arrest and resuscitation. J Learn Disabil

1993;26:46–51.

[39] Robertson CMT, Joffe AR, Moore AJ, et al. Neurodevelopmental outcome of young pedi-

atric intensive care survivors of serious brain injury. Pediatr Crit Care Med 2002;3(4):

345–50.

[40] ClarkRSB.Out of the frying pan and into the fire: neurodevelopmental outcome in pediatric

intensive care unit survivors of serious brain injury. Pediatr Crit Care Med 2002;3(4):384–5.

[41] Longstreth WT, Dikmen SS. Outcomes after cardiac arrest. Ann Emerg Med 1993;22:64–9.

[42] Grubb NR, O’Carroll R, Cobbe SM, et al. Chronic memory impairment after cardiac arrest

outside hospital. BMJ 1996;313:143–6.

[43] Maneru C, Junque C, Botet F, et al. Neuropsychological long-term sequelae of perinatal

asphyxia. Brain Inj 2001;15(12):1029–39.

[44] Dixon G, Badawi N, Kurinczuk JJ, et al. Early developmental outcomes after newborn

encephalopathy. Pediatrics 2002;109(1):26–33.

[45] Bhutta AT, Cleves MA, Casey PH, et al. Cognitive and behavioral outcomes of school-aged

children who were born preterm: a meta-analysis. JAMA 2002;288(6):728–37.

[46] Mandel R, Martinot A, Delepoulle F, et al. Prediction of outcome after hypoxic-ischemic

encephalopathy: a prospective clinical and electrophysiologic study. J Pediatr 2002;141(1):

45–50.

[47] Jacinto SJ, Gieron-Korthals M, Ferreira JA. Predicting outcome in hypoxic-ischemic brain

injury. Pediatr Clin North Am 2001;48(3):647–60.

[48] Painter MJ. Animal models of perinatal asphyxia: contributions, contradictions, clinical

relevance. Semin Pediatr Neurol 1995;2(1):37–56.

[49] Johnston MV, Trescher WH, Taylor GA. Hypoxic and ischemic central nervous system

disorders in infants and children. Adv Pediatr 1995;42:1–45.

[50] Barkovich AJ. MR and CT evaluation of profound neonatal and infantile asphyxia. Am J

Neuroradiol 1992;13(13):973–5.

[51] Grubb NR, Fox KA, Smith K, et al. Memory impairment in out-of-hospital cardiac arrest

survivors is associated with global reduction in brain volume, not focal hippocampal injury.

Stroke 2000;31(7):1509–14.

[52] Ng T, GrahamDI, Adams JH, et al. Changes in the hippocampus and the cerebellum result-

ing from hypoxic insults: frequency and distribution. Acta Neuropathol (Berl) 1989;78:

438–43.

[53] Peterson BS, Vohr B, Staib LH, et al. Regional brain volume abnormalities and long-term

cognitive outcome in preterm infants. JAMA 2000;284(15):1939–47.

[54] Isaacs EB, Lucas A, Chong WK, et al. Hippocampal volume and everyday memory in chil-

dren of very low birth weight. Pediatr Res 2000;47(6):713–20.


[55] Sie LT, van der KnaapMS, Oosting J, et al. MR patterns of hypoxic-ischemic brain damage

after prenatal, perinatal or postnatal asphyxia. Neuropediatrics 2000;31(3):128–36.

[56] Kreis R, Arcinue E, Ernst T, et al. Hypoxic encephalopathy after near-drowning studied by

quantitative 1H-magnetic resonance spectroscopy. J Clin Invest 1996;97:1142–54.

[57] Huppi PS, Murphy B, Maier SE, et al. Microstructural brain development after perinatal

cerebral white matter injury assessed by diffusion tensor magnetic resonance imaging. Pedi-

atrics 2001;107(3):455–60.

[58] Klingberg T, Vaidya CJ, Gabrieli JD, et al. Myelination and organization of the frontal

white matter in children: a diffusion tensor MRI study. Neuroreport 1999;10(13):2817–21.

[59] PausT,CollinsDL,EvansAC, et al.Maturationofwhitematter in the humanbrain: a review

of magnetic resonance studies. Brain Res Bull 2001;54(3):255–66.

[60] Mukherjee P, Miller JH, Shimony JS, et al. Normal brain maturation during childhood:

developmental trends characterized with diffusion-tensor MR imaging. Radiology 2001;

221(2):349–58.

[61] Boujraf S, Luypaert R, Shabana W, et al. Study of pediatric brain development using mag-

netic resonance imaging of anisotropic diffusion. Magn Reson Imaging 2002;20(4):327–36.

[62] Schmithorst VJ, Wilke M, Dardzinski BJ, et al. Correlation of white matter diffusivity and

anisotropy with age during childhood and adolescence: a cross-sectional diffusion-tensor

MR imaging study. Radiology 2002;222(1):212–8.

[63] Kochanek PM, Clark RS, Ruppel RA, et al. Cerebral resuscitation after traumatic brain

injury and cardiopulmonary arrest in infants and children in the new millennium. Pediatr

Clin North Am 2001;48(3):661–81.

[64] Skippen P, Seear M, Poskitt K, et al. Effect of hyperventilation on regional cerebral blood

flow in head-injured children. Crit Care Med 1997;25(8):1402–9.

[65] Buunk G, van der Hoeven JG, Meinders AE. Cerebrovascular reactivity in comatose

patients resuscitated from a cardiac arrest. Stroke 1997;28(8):1569–73.

[66] Vaagenes P, GinsbergM, Ebmeyer U, et al. Cerebral resuscitation from cardiac arrest: path-

ophysiologic mechanisms. Crit Care Med 1996;24:S57–68.

[67] Aufderheide TP, Sigurdsson G, Pirrallo RG, et al. Hyperventilation-induced hypotension

during cardiopulmonary resuscitation. Circulation 2004;109(16):1960–5.

[68] Aufderheide TP, Lurie KG. Death by hyperventilation: a common and life-threatening

problem during cardiopulmonary resuscitation. Crit Care Med 2004;32(Suppl 9):S345–51.

[69] Tobias JD, Lynch A, Garrett J. Alterations of end-tidal carbon dioxide during the intraho-

spital transport of children. Pediatr Emerg Care 1996;12(4):249–51.

[69a] Hickey RW, Kochanek PM, Ferimer H, et al. Hypothermia and hyperthermia in children

after resuscitation from cardiac arrest. Pediatrics 2000;106(1Pt 1):118–22.


cardiac arrest with induced hypothermia. N Engl J Med 2002;346(8):557–63.

[71] Hypothermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve

the neurologic outcome after cardiac arrest. N Engl J Med 2002;346(8):549–56.

[72] Nolan JP, Morley PT, Hoek TL, et al. Therapeutic hypothermia after cardiac arrest. An ad-

visory statement by the Advancement Life support Task Force of the International Liaison

Committee on Resuscitation. Resuscitation 2003;57(3):231–5.

[73] Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head cooling with mild systemic

hypothermia after neonatal encephalopathy: multicentre randomised trial. Lancet 2005;

365(9460):663–70.

[74] Hickey RW, Callaway CW. Asphyxia. In: Tisherman SA, Sterz F, editors. Therapeutic

hypothermia. New York: Springer; 2005. p. 119–34.

Neurol Clin 24 (2006) 159–169

The Declaration of Deathand the Withdrawal of Carein the Neurologic Patient

Edward M. Manno, MD*, Eelco F.M. Wijdicks, MDDivision of Critical Care Neurology, Department of Neurology W8B,

Mayo Clinic College of Medicine, 200 First Street Southwest,

Rochester, MN 55905, USA

Advances in critical care medicine and neurology have led to an increasein survival of critically ill patients who previously would have succumbed totheir illnesses. Whereas some may progress to lose all brain function, manywill survive with severe neurologic impairment. Under these circumstances,many patients or families may choose to withdraw medical care. Physiciansand staff caring for these patients are often presented with various complexmedical, moral, and ethical issues. This article discusses some of the issuesinvolved in the transitional period around the time of withdrawal of careor the declaration of death by neurologic criteria.

Ethical and legal issues involved in the withdrawal of care

Considerable confusion and misconceptions exist about withdrawal ofcare in a patient who is neurologically impaired. In a recent survey, manyneurologists believed that they needed legal counsel to withdraw therapyin the patient who is neurologically compromised and that withdrawal ofcare in these instances was tantamount to killing the patient [1]. This surveyhighlighted the limitations of many physicians’ knowledge and experience indealing with end-of-life issues.

Physician misunderstanding may lead to overaggressive treatment of pa-tients beyond their wishes [2]. Some patient and family surveys have sug-gested that patients who are dying often receive unwanted procedures andinterventions [2–4].


E-mail address: [email protected] (E.M. Manno).




160 MANNO & WIJDICKS

A patient’s right to informed consent and right to refuse treatment havebeen basic tenets of modern western medicine [5]. The right to informed con-sent is based on the ethical concept of self-autonomy and the legal conceptof self-determination [5]. Valid informed consent requires that the treatingphysician provides adequate information about the risks and benefits ofall therapeutic options. The patient also must be competent to make medicaldecisions and must not be coerced [5].

Competence is often the key feature that needs to be evaluated in the pa-tient who is neurologically impaired to ensure that a request for withdrawalof care is valid. Usually, competence can be evaluated adequately as part ofthe neurologic examination. Care must be taken not only to evaluate orien-tation and reasoning, but also to look for signs of depression or adjustmentdisorder that may influence a patient’s decision. In some circumstances, psy-chiatric input is needed to assess competence. In its strictest sense, compe-tence is a legal determination that can be made only in a court of law. Inrare instances, such as family disputes, court evaluation may be requiredto determine competence [5].

When a patient is unable to participate in medical decisions, a surrogateis assigned. The surrogate is typically a family member or members actingjointly. Most hospitals have an order as to who can make decisions forthe patient. The sequence can vary according to different hospital or lo-cal jurisdictions. In general, the typical stratification of surrogate decision-makers starts with the appointed legal guardian and runs in descendingorder to spouse, adult children, parents, siblings, and then more distant rel-atives. The role of surrogate decision-making has been clearly delineated.The surrogate should make decisions according to the patient’s previouslystated wishes. If these wishes are not known, then the surrogate makes de-cisions based on the patient’s cultural and belief systems. This concept isbest expressed as the surrogate attempting to reproduce the decision the pa-tient would make if he or she were able to participate in the discussion con-cerning continued care. If this aspect is also unknown, then the surrogateshould make decisions based on a patient’s perceived best interest [5,6].

There is no ethical or legal difference between withdrawal of therapy andrefusal of treatment; therefore, there is no difference between stopping a treat-ment that has been initiated and never starting the treatment. There is also eth-ical and legal consensus that artificial hydration and nutrition are forms ofmedical therapy that can be refused by a patient or surrogate [5].

Withdrawal of care: the decision

Whereas much has been written about the legal and ethical aspects of thewithdrawal of care, little has been written about how the process should ac-tually occur. In particular, the withdrawal of care in the patient who is neu-rologically devastated raises specific issues that may not occur with otherpatients.

161DECLARATION OF DEATH AND WITHDRAWAL OF CARE

The first issue in the process of withdrawing care from a patient who isneurologically devastated is to clarify the diagnosis and prognosis ‘‘sinceall subsequent decision making must start from this’’ [7]. Questions thatare often encountered include: ‘‘What are the chances of recovery?’’‘‘What tests need to be performed?’’ ‘‘Has anyone ever improved?’’ ‘‘Couldyour diagnosis or prognosis be wrong?’’ The answers to these questions arenot often clear-cut. Retrospective evaluations of patients who have globalcerebral anoxic injury have provided some guidelines in assessing patientoutcome [8–11]. The most consistent predictors of prognosis after intracere-bral hemorrhage include the size of the initial hemorrhage, the amount ofintraventricular blood, and the initial clinical presentation of the patient[12]. Prognosis after subarachnoid hemorrhage is based largely on theamount of blood and initial presentation of the patient [13]. Multiple studieshave evaluated prognosis after ischemic stroke [14–16]. The outcome afterhead trauma may be hard to predict initially, particularly in the patientwho is young. In addition, the action of withdrawal of care may actually in-fluence prognostic models, leading to a self-fulfilled prophecy [17]. Thus theability to predict early in the course of a neurologic disease may be limited.In general, prognosis often becomes clear with extended and repeated obser-vations and examinations.

Ancillary testing may serve several roles. Neuroimaging is often requiredfor making an accurate diagnosis. It can also be helpful to show to familiesthe significance of the neurologic injury. Electroencephalography (EEG) candetect subclinical seizure activity. EEG evaluation after global anoxic injurymay have limited prognostic use; however, specific EEG coma patterns mayportend a poor outcome [9]. A lack of a cortical signal during somatosensory-evoked potentials seems to have the most significance for predicting pooroutcome after anoxic injury [10].

Family discussion and decision about continued medical support are of-ten addressed at the time the prognosis for a functional recovery seems un-likely. These discussions need to be accurate, honest, and direct about theprognosis and chance of recovery. They commonly occur at critical decisionpoints along the continuum of care for the patient. It may be at the timea tracheostomy or feeding tube needs to be placed. Several important issuesneed to be explored at this time if these have not been previously discussed.Is there an advanced directive or any written instructions to provide someguidelines about potential withdrawal of support? Did the patient expressany feelings, wishes, or desires about end-of-life care? Who and how will de-cisions be made about continued care? In family disagreements, social serv-ices may be able to provide some support. In rare instances, the legaldepartment may need to evaluate and address some of the above questions.

It is important to elucidate the physicians’ and families’ understanding ofwhat represents a functional neurologic recovery and what conditionwould be acceptable to the patient. What types of abilities would the patientneed to regain in order for the family to view its loved one’s recovery as


functional? Does that mean a certain level of consciousness or physicalcapability, or an ability to interact with family briskly and genuinely? Isa long-term care facility acceptable to the family or consistent with the pa-tient’s previously expressed wishes? The physician must also be cognizant ofhis or her own prejudices, because it has been implied in several studies thatphysicians may apply their personal biases when entertaining decisionsabout withdrawal of care [18–20]. Physicians need to be sensitive to variouscultural, family, and individual preferences about end-of-life care becausesome families may have limited ability to understand an oftentimes confus-ing and complicated process.

Withdrawal of care: the process

At the time death seems imminent or if a patient is so severely impairedthat there is little chance of recovery, many families may elect not to continuewith medical care. When a surrogate or family decides to withdraw treat-ment from a patient, the plan of care for the patient proceeds from a tac-tical approach designed to identify and treat a specific illness to a strategicplan with a newly defined goal [2].

The new goal and strategic plan need to be defined clearly. The goal ofmedical care after the decision for withdrawal of support is to provide pal-liative care. The family discussion needs to describe exactly what is to bedone. Issues that need to be clarified include what, when, and how specificinterventions will be discontinued. The process may occur in a stepwise fash-ion or move directly to comfort measures only.

The discussion and the plan of care need to be documented in the medicalrecord. When advanced directives are not available, phrases such as ‘‘theseplans are consistent with the patient’s previously expressed wishes or de-sires’’ provides documentation for the family or legal guardian that the de-cisions made represented substituted judgment based on the patient’s valuesystem. Further discussions with the family, plans for withdrawal, and seda-tion should be carefully noted.

Once the plan for withdrawal of care has been developed, it is the job ofthe treating physician and team to implement the plan. Physicians will haveto consider the timing of and which life support measures will be discontin-ued [18]. Many families or physicians may have reservations about specifictherapies that can or cannot be withdrawn. In practice, all measures can bewithdrawn, including vasopressors, drugs, mechanical ventilation, and nu-tritional support [18]. Most physicians, however, withdraw care in a stepwisefashion [21], often based on cost or medical or ethical misconceptions [22].

The purpose of withdrawal of support is to maximize comfort for the pa-tient. To that end, if a stepwise withdrawal meets this purpose, it may bebest to proceed in that manner. For instance, if a patient depends on vaso-pressors for blood pressure support, withdrawal of these medications maylead predictably and expeditiously to death without the need to withdraw


mechanical ventilation. Similarly, some physicians have recommended a ter-minal weaning protocol for mechanical ventilation rather than proceedingdirectly to extubation [23]. In many cases, however, a stepwise withdrawalof care may not lead predictably to death. In these situations, a slow with-drawal of treatments may actually prolong the dying process and potentiallyincrease suffering.

The patient who is neurologically devastated differs in many respectsfrom other terminally ill patients. In most cases, medical decision willneed to be made by legal guardians or families. Many patients whohave neurologic injury will be intubated for concerns over airway protec-tion and not because of a primary pulmonary process [24]. Graduatedwithdrawal of ventilatory rates or oxygen levels may not be necessary be-cause pulmonary function may be normal in these individuals. The needfor sedation and analgesia for the patient who is neurologically devastatedis unknown; however, cerebral blood flow patterns in patients in a vegeta-tive state resemble patterns of patients under anesthesia, suggesting thatpatients with diffuse neocortic damage do not perceive pain [25,26]. Gri-macing and many facial expressions are mediated at a brainstem level;and although they may not denote pain for the patient, it may be distress-ing for the family to observe.

Routine blood draws, medications, and monitoring should be discontin-ued once the decision has been made to withdraw care. The room should befree of distractions and noisy alarms. If the goal of palliative care is terminalextubation, then plans must be made to assure the patient has adequate air-way management postextubation. Usually, careful head positioning accom-panied with nasal or oral airways are sufficient to maintain airway patency.If the patient seems to be in pain or is agitated, sedation and analgesiashould be given. Typically, morphine, in 5 mg intravenous infusions, maybe given initially. A morphine or fentanyl drip is preferred to intermittentintravenous bolus administration to allow for better titration for effect.

Morphine administration should be titrated to ensure comfort for the pa-tient. Nurses or physicians may be reluctant to increase medication infusionsbecause of concerns of inducing active euthanasia through respiratory de-pression. The concern is probably overstated because actual respiratorydepression, even with large doses of morphine, is rare [18]. In addition,physicians and nurses are covered medically and legally under the principleof double effect, which ensures that staff is not liable for the untoward sec-ondary effects of a medication if the primary reason a medication is given isto initiate another effect [2,5,18]. In this situation, the reason the medicationis given is to provide comfort. Some of this anxiety can be alleviated by writ-ing orders to titrate a medication to a specific effect (ie, a respiratory rate orlevel of sedation).

The use of neuromuscular blockers should be avoided or discontinued ifwithdrawal is considered because it may not only directly hasten a patient’sdeath but also may mask any signs of distress that a patient may have [5].


Predicting the timing of death in a neurologic patient after withdrawal ofcare can be problematic. In some instances, such as patients who are wors-ening acutely or patients who have severe brainstem damage, death can beexpected in a short time. Other patients may tolerate extubation without in-cident. Death under these circumstances may not be imminent.

It is important to prepare families for a potential variety of scenarios. Ex-plaining to families what they may expect to see may relieve some anxiety.Also, if there is some uncertainty over the timing of death, arrangementsshould be made to move the patient from an ICU setting to a private or hos-pice room to allow for a more peaceful environment. Care and counseling ofthe family should continue as part of care for the patient. Clergy and othersupport staff should be actively engaged during this period. Most delayedcriticisms from families occur during this time, if this period is not carefullyand sensitively managed [27,28].

For many, it is important to be present at the time of cardiopulmonaryarrest, and some families may request to be present at the time of extubationif death is deemed to be imminent. Families should be counseled about pos-sible terminal limb or body movements that can occur. It is prudent to askthat all present be sitting to avoid any possible vasovagal syncope that couldoccur to family members during this period of psychosocial stress.

Neurologic criteria for the declaration of death

Many patients will succumb to their neurologic injuries and subsequentlybe declared dead by neurologic criteria. Modern critical care techniques haveled to the development of a population of patients who have lost all neuro-logic function but continue to have their respirations supported and sustaina heartbeat. A redefinition of death resulted, initially by the Harvard AdHoc committee criteria defining irreversible coma, which were later adaptedby the Uniform Determination of Death Act in 1981. This act defined deathas either the irreversible cessation of whole brain function or the irreversibleloss of circulatory and respiratory function [29–31].

In the United States and in most other countries, the concept of declaringdeath through neurologic criteria uses the idea of ‘whole brain death.’ Inwhole brain death, death is declared when there is demonstrated loss of cor-tical and brainstem reflexes. In Britain, only the loss of brainstem function isrequired. The concept of ‘whole brain death’ has survived decades of moral,ethical, and legal inspection [32]; however, the implementation of the criteriaestablished to determine brain death has come under scrutiny [33–36].

The American Academy of Neurology [37] established a set of practice pa-rameters for the determination of brain death in 1995; it has been reviewedpreviously by Wijdicks [38]. In general, the diagnosis of death by neurologiccriteria requires the clinical or radiographic evidence for catastrophic and ir-reversible brain injury, the exclusion of confounding factors, and a neurologicassessment to evaluate cortical and cranial nerve function [36].


Potential pitfalls in the declaration of death by neurologic criteria

There are several potential problems in the process of declaring death byneurologic criteria. The identification of a neurologically devastating injuryon neuroimaging provides assurance that a confounding diagnosis is unlikelyto account for the neurologic findings. In particular, locked-in syndrome andGuillain-Barre syndrome need to be ruled out as possible diagnoses [38].Severe electrolyte, acid-base, and endocrine disturbances need to be evalu-ated, although the exact level at which these disturbances would precludeneurologic testing has not been clarified. Drug intoxication needs to be ex-cluded before neurologic testing. Barbiturate levels should be subtherapeu-tic. Wijdicks has suggested a period of observation for at least four timesthe elimination half-life of a known intoxicant and a period of observationof 48 hours for an unknown or nonquantifiable substance [38].

The patient may become unstable hemodynamically at the time the lastfew brain stem reflexes are lost. Most patients can be treated successfullywith intravenous fluids and vasopressors. Diabetes insipidus can occurquickly and with dramatic hemodynamic effect. The authors prefer intrave-nous vasopressin over intranasal or subcutaneous administration because ofits more rapid and predictable absorption.

The neurologic examination of the patient who is deceased must be per-formed and carefully documented. Two common pitfalls are to examine pa-tients who are hypothermic or hypotensive. The American AcademyPractice Parameters recommend a core temperature of greater than 32�Cand a systolic blood pressure of greater than 90 mm/Hg for the clinical dec-laration of death. Apnea testing requires a temperature of 36.5�C [37,38].

Examination of each cranial nerve may have some potential for error.The patient’s head should be elevated 30�, and tympanic membranes shouldbe identifiable and intact before testing cold caloric responses. Atropine ormydriatics can confound pupillary responses and should be excluded. Bron-chial suctioning rather than endotracheal tube movement should be used toassess for a cough reflex [38]. Painful stimuli to the extremities should be sig-nificant because this examination must not be misinterpreted.

The apnea test is an evaluation of a brainstem reflex. Apnea is the mosttolerant reflex to hypoxia, and the absence of breathing is the most associ-ated with death by the public. Special care needs to be taken when perform-ing an apnea test. Apnea tests are not benign and can be associated withinjury or cardiopulmonary death if not performed properly [39].

Preoxygenation and normalization of the partial pressure of carbon diox-ide (pCO2) is important before starting the test. The examiner must be awareof any reflexive-type movements, which can mimic respiration. The exami-nation should be discontinued if there is significant hypotension or hypoxiaduring the testing. Vasopressors may be used to support a patient’s bloodpressure as needed during the test. If the patient is unable to tolerate or com-plete the examination, a confirmatory test is required.


Confirmatory tests are recommended when some portion of the neurol-ogic history or assessment is not known or cannot be performed. Thereare several tests that are commonly used, including cerebral angiography,EEG, transcranial Doppler ultrasonography, or technetium-99m hexame-thylpropyleneamineoxime brain scans. Each test has specific guidelinesand limitations to its use [37].

It is unfortunate, but guidelines for the declaration of brain death haveonly been variably applied. There remains considerable inconsistency inthe determination of death by neurologic criteria nationally and internation-ally [36,40]. Guidelines for the neurologic declaration of death are often de-termined by individual hospitals [33]. It is important that a physician befamiliar with the guidelines at each hospital where he or she practices.

The declaration of death and organ donation

Once a patient has failed the set of established neurologic criteria, he orshe is declared dead. All paperwork or legal documents that need to be com-pleted should list the time of death as the time of death by neurologic criteriaand not the time of circulatory or cardiopulmonary arrest. Discussions withfamilies at this time need to be caring but unambiguous and direct in statingthat the patient has died. In rare instances, families may not accept this factor may continue to grieve. Usually, with patience and time, most conflictscan be resolved. The physician is under no moral or legal obligation tocontinue medical support for a cadaver. In general, family requests to spendtime with the body before withdrawal of the ventilator can beaccommodated.

The acceptance of death by neurologic criteria is accepted by every majorreligion, with the exceptions of gypsies and selected orthodox Jewish sects[41]. In New York and New Jersey, state laws do not allow the declarationof death by neurologic criteria if the family or individual previously objectedto the concept of brain death based on religious beliefs [42,43]. Under thesecircumstances, the physician is required to continue medical support.

The request for organ donation after the declaration of death by neu-rologic criteria is often a difficult and delicate situation. The structure,environment, and sequence of how organ donation is approached have sig-nificant impact on the rates of organ donation [44–46]. In 1998, the centersfor Medicare and Medicaid services required that hospitals notify their localorgan procurement organization about patients whose death appeared im-minent. The federal conditions of participation mandated that any requestfor donation be performed by a person trained specifically in these matters.This mandate does not exclude physicians or staff, but it does require themto undergo training [47].

Consent for organ donation is highest when the discussion of brain deathis separated from the discussion about organ donation, when the requestis made by a trained individual, and when the ICU staff and organ


procurement organizations coordinate their efforts [46,47]. In the authors’practice, the ICU physicians declare and discuss death with the family.They allow time for questions, grieving, and consolation. They leave thediscussion of organ donation to the organ procurement organization butinform patients that someone will inquire about this issue. This ‘‘decou-pling’’ of conversations has led to an improvement in the ability to obtainconsent for organ donation.

Summary

Intensive care technologies have led to an increase in patients who areneurologically devastated and deceased. The practical, moral, and ethicalsituations encountered can be varied and challenging to manage. Decisionsand discussions surrounding withdrawal of care, death by neurologic crite-ria, and organ donation require significant knowledge of the prognosis, an-cillary testing, and definitions of these processes. Experience and skill areoften required on the part of physicians and staff to guide families throughthese most difficult of circumstances.

References

[1] Carver AC, Vickrey BG, Bernat JL, et al. End-of-life care: a survey of US neurologists’ atti-

tudes, behavior, and knowledge. Neurology 1999;53(2):284–93.

[2] Brody H, Campbell ML, Faber-Langendoen K, et al. Withdrawing intensive life-sustaining

treatmentdrecommendations for compassionate clinical management. N Engl J Med 1997;

336(9):652–7.

[3] Solomon MZ, O’Donnell L, Jennings B, et al. Decisions near the end of life: professional

views on life-sustaining treatments. Am J Public Health 1993;83(1):14–23.

[4] Anonymous. A controlled trial to improve care for seriously ill hospitalized patients.

The study to understand prognoses and preferences for outcomes and risks of treatments

(SUPPORT). The SUPPORT Principal Investigators. JAMA 1995;274(20):1591–8.

[5] Bernat JL. Ethical and legal issues in palliative care. Neurol Clin 2001;19(4):969–87.

[6] Bernat JL, Goldstein ML, Viste KM Jr. The neurologist and the dying patient. Neurology

1996;46(3):598–9.

[7] Anonymous. Persistent vegetative state: report of the American Neurological Association

Committee on Ethical Affairs. ANA Committee on Ethical Affairs. Ann Neurol 1993;

33(4):386–90.

[8] Attia J, Cook DJ. Prognosis in anoxic and traumatic coma. Crit Care Clin 1998;14(3):

497–511.

[9] ChenR, BoltonCF, Young B. Prediction of outcome in patients with anoxic coma: a clinical

and electrophysiologic study. Crit Care Med 1996;24(4):672–8.

[10] ZandbergenEG, deHaanRJ, StoutenbeekCP, et al. Systematic review of early prediction of

poor outcome in anoxic-ischaemic coma. Lancet 1998;352(9143):1808–12.


JAMA 1985;253(10):1420–6.

[12] Manno EM, Atkinson JL, Fulgham JR, et al. Emerging medical and surgical management

strategies in the evaluation and treatment of intracerebral hemorrhage. Mayo Clin Proc

2005;80(3):420–33.

[13] Manno EM. Subarachnoid hemorrhage. Neurol Clin 2004;22(2):347–66.


[14] Wijdicks EF, Rabinstein AA. Absolutely no hope? Some ambiguity of futility of care in dev-

astating acute stroke. Crit Care Med 2004;32(11):2332–42.

[15] Gujjar AR, Deibert E, Manno EM, et al. Mechanical ventilation for ischemic stroke and in-

tracerebral hemorrhage: indications, timing, and outcome. Neurology 1998;51(2):447–51.

[16] Toni D, Fiorelli M, Bastianello S, et al. Acute ischemic strokes improving during the first 48

hours of onset: predictability, outcome, and possible mechanisms. A comparison with early

deteriorating strokes. Stroke 1997;28(1):10–4.

[17] BeckerKJ, BaxterAB,CohenWA, et al.Withdrawal of support in intracerebral hemorrhage

may lead to self-fulfilling prophecies. Neurology 2001;56(6):766–72.

[18] Rubenfeld GD. Principles and practice of withdrawing life-sustaining treatments. Crit Care

Clin 2004;20(3):435–51.

[19] Wachter RM, Luce JM, Hearst N, et al. Decisions about resuscitation: inequities among pa-

tients with different diseases but similar prognoses. Ann Intern Med 1989;111(6):525–32.

[20] Hanson LC, Danis M, Garrett JM, et al. Who decides? Physicians’ willingness to use life-

sustaining treatment. Arch Intern Med 1996;156(7):785–9.

[21] Faber-Langendoen K. A multi-institutional study of care given to patients dying in hospi-

tals. Ethical and practice implications. Arch Intern Med 1996;156(18):2130–6.

[22] Asch DA, Christakis NA. Why do physicians prefer to withdraw some forms of life support

over others? Intrinsic attributes of life-sustaining treatments are associated with physicians’

preferences. Med Care 1996;34(2):103–11.

[23] Gianakos D. Terminal weaning. Chest 1995;108(5):1405–6.

[24] CoplinWM, PiersonDJ, CooleyKD, et al. Implications of extubation delay in brain-injured

patients meeting standard weaning criteria. Am J Respir Crit CareMed 2000;161(5):1530–6.

[25] Anonymous. Medical aspects of the persistent vegetative state (1). The Multi-Society Task

Force on PVS. N Engl J Med 1994;330(21):1499–508.

[26] Levy DE, Sidtis JJ, Rottenberg DA, et al. Differences in cerebral blood flow and glucose

utilization in vegetative versus locked-in patients. Ann Neurol 1987;22(6):673–82.

[27] Abbott KH, Sago JG, Breen CM, et al. Families looking back: one year after discussion of

withdrawal or withholding of life-sustaining support. Crit Care Med 2001;29(1):197–201.

[28] Mayer SA, Kossoff SB. Withdrawal of life support in the neurological intensive care unit.

Neurology 1999;52(8):1602–9.

[29] Anonymous. A definition of irreversible coma. Report of the AdHoc Committee of theHar-

vardMedical School to Examine the Definition of Brain Death. JAMA 1968;205(6):337–40.

[30] Van Norman GA. Another matter of life and death: what every anesthesiologist should

know about the ethical, legal, and policy implications of the non-heart-beating cadaver

organ donor. Anesthesiology 2003;98(3):763–73.

[31] Manno EM. Donation in the neurologically devastated patient: non-heart-beating donor.

NeuroCritical Care, in press.

[32] Bernat JL. A defense of the whole-brain concept of death. Hastings Cent Rep 1998;28(2):

14–23.

[33] Powner DJ, Hernandez M, Rives TE. Variability among hospital policies for determining

brain death in adults. Crit Care Med 2004;32(6):1284–8.

[34] Shaner DM, Orr RD, Drought T, et al. Really, most SINCERELY dead: policy and proce-

dure in the diagnosis of death by neurologic criteria. Neurology 2004;62(10):1683–6.

[35] Shemie SD. Variability of brain death practices. Crit Care Med 2004;32(12):2564–5.

[36] Oropello JM. Determination of brain death: theme, variations, and preventable errors.

Crit Care Med 2004;32(6):1417–8.

[37] Anonymous. Practice parameters for determining brain death in adults (summary state-

ment). The Quality Standards Subcommittee of the American Academy of Neurology.

Neurology 1995;45(5):1012–4.

[38] Wijdicks EF. The diagnosis of brain death. N Engl J Med 2001;344(16):1215–21.

[39] Saposnik G, Rizzo G, Vega A, et al. Problems associated with the apnea test in the diagnosis

of brain death. Neurol India 2004;52(3):342–5.


[40] Wijdicks EF. Brain death worldwide: accepted fact but no global consensus in diagnostic

criteria. Neurology 2002;58(1):20–5.

[41] Gallagher C, Wijdicks EFM. Religious and cultural aspects of brain death. In: Wijdicks

EFM, editor. Brain Death. Philadelphia: Lippincott Williams & Wilkens; 2001. p. 135–50.

[42] Bernat JL. Philosophical and ethical aspects of brain death. In: Wijdicks EFM, editor.

Brain Death. Philadelphia: Lippincott Williams & Wilkens; 2001. p. 171–87.

[43] Olick RS. Brain death, religious freedom, and public policy: New Jersey’s landmark legisla-

tive initiative. Kennedy Inst Ethics J 1991;1(4):275–92.

[44] Williams MA, Lipsett PA, Rushton CH, et al. The physician’s role in discussing organ do-

nation with families. Crit Care Med 2003;31(5):1568–73.

[45] DeJong W, Franz HG, Wolfe SM, et al. Requesting organ donation: an interview study of

donor and nondonor families. Am J Crit Care 1998;7(1):13–23.

[46] Gortmaker SL, Beasley CL, Sheehy E, et al. Improving the request process to increase family

consent for organ donation. J Transpl Coord 1998;8(4):210–7.

[47] Anonymous. Medicare and Medicaid programs; hospital conditions of participation; iden-

tification of potential organ, tissue, and eye donors and transplant hospitals’ provision of

transplant-related data–HCFA. Final rule. Fed Regist 1998;63(119):33856–75.

Documents

2006, Vol.24, Issues 1, Brain Injury and Cardiac Arrest