Examination of Physiological Function and Biochemical Disorders in a Rat Model of Prolonged Asphyxia-Induced Cardiac Arrest followed by Cardio Pulmonary Bypass Resuscitation

Background

Cardiac arrest induces whole body ischemia, which causes damage to multiple organs particularly the heart and the brain. There is clinical and preclinical evidence that neurological injury is responsible for high mortality and morbidity of patients even after successful cardiopulmonary resuscitation. A better understanding of the metabolic alterations in the brain during ischemia will enable the development of better targeted resuscitation protocols that repair the ischemic damage and minimize the additional damage caused by reperfusion.

Method

A validated whole body model of rodent arrest followed by resuscitation was utilized; animals were randomized into three groups: control, 30 minute asphyxial arrest, or 30 minutes asphyxial arrest followed by 60 min cardiopulmonary bypass (CPB) resuscitation. Blood gases and hemodynamics were monitored during the procedures. An untargeted metabolic survey of heart and brain tissues following cardiac arrest and after CPB resuscitation was conducted to better define the alterations associated with each condition.

Results

After 30 min cardiac arrest and 60 min CPB, the rats exhibited no observable brain function and weakened heart function in a physiological assessment. Heart and brain tissues harvested following 30 min ischemia had significant changes in the concentration of metabolites in lipid and carbohydrate metabolism. In addition, the brain had increased lysophospholipid content. CPB resuscitation significantly normalized metabolite concentrations in the heart tissue, but not in the brain tissue.

Conclusion

The observation that metabolic alterations are seen primarily during cardiac arrest suggests that the events of ischemia are the major cause of neurological damage in our rat model of asphyxia-CPB resuscitation. Impaired glycolysis and increased lysophospholipids observed only in the brain suggest that altered energy metabolism and phospholipid degradation may be a central mechanism in unresuscitatable brain damage.     

Nitrite therapy is neuroprotective and safe in cardiac arrest survivors

Cardiac arrest results in significant mortality after initial resuscitation due in most cases to ischemia-reperfusion induced brain injury and to a lesser degree myocardial dysfunction. Nitrite has previously been shown to protect against reperfusion injury in animal models of focal cerebral and heart ischemia. Nitrite therapy after murine cardiac arrest improved 22 h survival through improvements in myocardial contractility. These improvements accompanied transient mitochondrial inhibition which reduced oxidative injury to the heart. Based on preliminary evidence that nitrite may also protect against ischemic brain injury, we sought to test this hypothesis in a rat model of asphyxia cardiac arrest with prolonged survival (7d). Cardiac arrest resulted in hippocampal CA1 delayed neuronal death well characterized in this and other cardiac arrest models. Nitrite therapy did not alter post-arrest hemodynamics but did result in significant (75%) increases in CA1 neuron survival. This was associated with increases in hippocampal nitrite and S-nitrosothiol levels but not cGMP shortly after therapy. Mitochondrial function 1h after resuscitation trended towards improvement with nitrite therapy. Based on promising preclinical data, the first ever phase I trial of nitrite infusions in human cardiac arrest survivors has been undertaken. We present preliminary data showing low dose nitrite infusion did not result in hypotension or cause methemoglobinemia. Nitrite thus appears safe and effective for clinical translation as a promising therapy against cardiac arrest mediated heart and brain injury.     

Prognosis after cardiac arrest based on age and duration of coma.

In an attempt to determine the relation between duration of coma and neurologic recovery following cardiac resuscitation 163 survivors of cardiac arrest from Winnipeg, Manitoba and Aarhus, Denmark were studied. The age of the patients did not influence the outcome. Of the 153 patients who had awakened from the coma within 24 hours, only 11 suffered brain damage, compared with all of the 10 patients who wakened after 24 hours. The three who wakened after 72 hours had severe brain damage and required permanent care in an institution. It was concluded that recovery of communicative brain function is unlikely if coma persists longer than 72 hours after cardiac arrest and that full recovery cannot be expected after 24 hours of coma.     

Brain nuclear DNA survives cardiac arrest and reperfusion.

Iron-mediated peroxidation of brain lipids is known to occur during reperfusion following cardiac arrest. Since in vitro damage to DNA is caused by similar iron-dependent peroxidation, we tested whether free radical damage to genomic DNA also develops during reperfusion following cardiac arrest and resuscitation. Genomic DNA was isolated from the cerebral cortex in (i) normal dogs, (ii) dogs subjected to a 20-min cardiac arrest, and (iii) dogs resuscitated from a 20-min cardiac arrest and then allowed to reperfuse for 2 or 8 h. DNA strand nicks were evaluated by in vitro labeling of newly created 3' and 5' termini. DNA base damage was evaluated utilizing reaction with piperidine prior to labeling of 5' termini. The 3' DNA termini were labeled before and after digestion with exonuclease III, and the 5' DNA termini were labeled before and after treatment with piperidine. In vitro experiments with genomic DNA damaged by oxygen radicals verified that these labeling methods identified radical damage. In the experimental animal groups, terminal incorporation and electrophoretic mobility of brain nuclear DNA are not significantly changed either by 20 min of complete brain ischemia or during the first 8 h of reperfusion. We conclude that genomic DNA is not extensively damaged during cardiac arrest and early reperfusion, and therefore such DNA damage does not appear to be an important early aspect of the neurologic injury that accompanies cardiac arrest and resuscitation.     

Evaluation of hospital-based cardiac resuscitation, 1973-77

The resuscitation experience of a large teaching hospital during 1973-77 was reviewed. Resuscitation was attempted on 2091 victims of cardiac arrest; 261 patients (12.5%) survived to be discharged from hospital.Coronary heart disease caused about one half of all the cardiac arrests, but was associated with a better survival rate (14.4%) than the other causes. Cardiac arrest following multiple trauma had the worst prognosis; only 3% of the patients survived to be discharged from hospital. However, the main factor influencing outcome was the site of arrest. The survival rates of patients on whom resuscitation was initiated in the emergency room or an intensive care area were triple and double the rate for patients in hospital wards, although one third of all the cardiac arrests induced by a coronary event and occurring in hospital were on the wards. Patients whose arrest occurred outside hospital, where only basic life support was available, had a survival rate of just 6.3%, whereas those whose arrest occurred in the emergency room had a survival rate of 31.9%. Since these two patient groups were similar in terms of age and diagnosis, we believe that the potential survival rate for victims of cardiac arrest outside of hospital that are optimally treated is close to 30%.These data suggest that increased survival from cardiac arrest can be expected with extension of the resuscitation services both inside and outside of hospital, but particularly with increased emphasis on emergency cardiac care outside of hospital.     

Survey of 3765 cardiopulmonary resuscitations in British hospitals (the BRESUS Study): methods and overall results.

OBJECTIVE--To determine the circumstances, incidence, and outcome of cardiopulmonary resuscitation in British hospitals. DESIGN--Hospitals registered all cardiopulmonary resuscitation attempts for 12 months or longer and followed survival to one year. SETTING--12 metropolitan, provincial, teaching, and non-teaching hospitals across Britain. SUBJECTS--3765 patients in whom a resuscitation attempt was performed, including 927 in whom the onset of arrest was outside the hospital. MAIN OUTCOME MEASURE--Survival after initial resuscitation, at 24 hours, at discharge from hospital, and at one year, calculated by the life table method. RESULTS--There were 417 known survivors at one year, with 214 lost to follow up. By life table analysis for every eight attempted resuscitations there were three immediate survivors, two at 24 hours, 1.5 leaving hospital alive, and one alive at one year. Survival at one year was 12.5% including out of hospital cases and 15.0% not including these cases. Each hospital year averaged 30 survivors at one year: three who had an arrest outside hospital, seven who had one in the accident and emergency department, seven in the cardiac care unit, 10 in the general wards, and three in other, non-ward areas. Within the hospitals survival rates were best in those who had an arrest in the accident and emergency department, the cardiac care unit, or other specialised units. Outcome varied 12-fold in subgroups defined by age, type of arrest, and place of arrest. CONCLUSION--71% of the mortality at one year in patients undergoing attempted resuscitation occurred during the initial arrest. Hospital resuscitation is life saving and cost effective and warrants appropriate attention, training, coordination, and equipment.     

Resuscitation of patients with cardiac arrest by ambulance staff with extended training in West Yorkshire.

OBJECTIVE--To investigate the results of resuscitation of patients with cardiac arrest by ambulance staff with extended training in West Yorkshire. DESIGN--Study of all such attempts at resuscitation over 32 months, based on the standard report form for each call made by the ambulance staff and the electrocardiogram that showed the initial rhythm in each patient. SETTING--Area covered by West Yorkshire ambulance service. SUBJECTS--1196 Patients with cardiac arrests attended by 29 ambulance staff with extended training. MAIN OUTCOME MEASURE--Result of resuscitation. RESULTS--The initial rhythm was asystole or electromechanical dissociation in 740 patients and ventricular fibrillation in 456 patients; overall 65 patients survived to be discharged from hospital. Sixty four of the 456 patients in whom ventricular fibrillation was the initial rhythm recorded, and 46 in whom ventricular fibrillation persisted after the ambulance staff arrived, survived. Only one of the 740 patients who initially had asystole or electromechanical dissociation survived. Factors associated with a greater chance of ventricular fibrillation occurring were: age less than 71, the arrest being witnessed by a bystander, resuscitation by a bystander, the arrest occurring in a public place, and a response time by the ambulance staff of less than six minutes. For patients found in ventricular fibrillation a shorter response time was associated with improved survival but resuscitation by a bystander was not. Additional skills learnt during extended training were used for 51 of the 65 patients who survived. CONCLUSIONS--Ambulance staff with extended training can save the lives of patients with cardiac arrest due to fibrillation, though asystole and electromechanical dissociation have a poor prognosis and should perhaps receive little attention during extended training.     

CARDIAC ARREST DURING ANESTHESIA

In cases of acute cardiac arrest from transient or reversible causes resuscitation is a distinct possibility. Prompt action is of the utmost importance.The sequelae of cardiac arrest will depend on the degree of anoxia that has developed and the amount of irreversible damage to brain tissue. Efficient manual artificial circulation, begun immediately, can prevent such damage.Four cases of cardiac arrest are described. Sudden circulatory collapse, which may or may not imply cardiac arrest, is not uncommon during surgical procedures. Usually the patient recovers quickly when treatment is prompt.     

DNA mismatch repair-dependent activation of c-Abl/p73alpha/GADD45alpha-mediated apoptosis

Cells with functional DNA mismatch repair (MMR) stimulate G(2) cell cycle checkpoint arrest and apoptosis in response to N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). MMR-deficient cells fail to detect MNNG-induced DNA damage, resulting in the survival of "mutator" cells. The retrograde (nucleus-to-cytoplasm) signaling that initiates MMR-dependent G(2) arrest and cell death remains undefined. Since MMR-dependent phosphorylation and stabilization of p53 were noted, we investigated its role(s) in G(2) arrest and apoptosis. Loss of p53 function by E6 expression, dominant-negative p53, or stable p53 knockdown failed to prevent MMR-dependent G(2) arrest, apoptosis, or lethality. MMR-dependent c-Abl-mediated p73alpha and GADD45alpha protein up-regulation after MNNG exposure prompted us to examine c-Abl/p73alpha/GADD45alpha signaling in cell death responses. STI571 (Gleevec, a c-Abl tyrosine kinase inhibitor) and stable c-Abl, p73alpha, and GADD45alpha knockdown prevented MMR-dependent apoptosis. Interestingly, stable p73alpha knockdown blocked MMR-dependent apoptosis, but not G(2) arrest, thereby uncoupling G(2) arrest from lethality. Thus, MMR-dependent intrinsic apoptosis is p53-independent, but stimulated by hMLH1/c-Abl/p73alpha/GADD45alpha retrograde signaling.     

Effects of regional body temperature variation during asphyxial cardiac arrest on mortality and brain damage in a rat model

To date, hypothermia has focused on improving rates of resuscitation to increase survival in patients sustaining cardiac arrest (CA). Towards this end, the role of body temperature in neuronal damage or death during CA needs to be determined. However, few studies have investigated the effect of regional temperature variation on survival rate and neurological outcomes. In this study, adult male rats (12 week-old) were used under the following four conditions: (i) whole-body normothermia (37 ± 0.5 °C) plus (+) no asphyxial CA, (ii) whole-body normothermia + CA, (iii) whole-body hypothermia (33 ± 0.5 °C)+CA, (iv) body hypothermia/brain normothermia + CA, and (v) brain hypothermia/body normothermia + CA. The survival rate after resuscitation was significantly elevated in groups exposed to whole-body hypothermia plus CA and body hypothermia/brain normothermia plus CA, but not in groups exposed to whole-body normothermia combined with CA and brain hypothermia/body normothermia plus CA. However, the group exposed to hypothermia/brain normothermia combined with CA exhibited higher neuroprotective effects against asphyxial CA injury, i.e. improved neurological deficit and neuronal death in the hippocampus compared with those involving whole-body normothermia combined with CA. In addition, neurological deficit and neuronal death in the group of rat exposed to brain hypothermia/body normothermia and CA were similar to those in the rats subjected to whole-body normothermia and CA. In brief, only brain hypothermia during CA was not associated with effective survival rate, neurological function or neuronal protection compared with those under body (but not brain) hypothermia during CA. Our present study suggests that regional temperature in patients during CA significantly affects the outcomes associated with survival rate and neurological recovery.     

Possibility of brain recovery after electrically induced cardiac arrest and reanimation in dogs

The changes of aortic blood pressure (BP), carotid artery flow (CAF), power spectrum of analysed EEG, neurologic deficit and survival rate were determinated in dogs after experimental cardiac arrest of different duration. Following artificially induced ventricular fibrillation of 1, 4, 10, 12 and 15 min duration successful cardiopulmonary resuscitation was performed in 30 experimental animals. Alterations of power spectrum during and after reanimation procedures, severity of the neurologic state and the survival rate deteriorated in parallel with the increasing duration of circulatory stop. Advantageous effect of direct heart massage could be demonstrated by the measuring circulatory parameters. Following a 15 min fibrillation, all animals were lost in a few hours despite the successful restoration of circulation and ventilation. Considering the various experimental and clinical conditions experimental cardiac arrest lasting for 12 min seems to be useful in extrapolating the results to human cases. The suggested model allows to study the brain function recovery after circulatory stop and resuscitation.     

Sudden cardiac arrest on the field of play: turning tragedy into a survivable event

Sudden cardiac arrest remains the leading cause of death in exercising athletes, and recent studies have shown that it occurs more frequently than historical estimates. While out-of-hospital cardiac arrest often proves fatal, advance preparation can improve outcomes and the chance of survival. First responders to a collapsed athlete on the field of play may include team medical personnel, coaches, other athletes, officials, venue staff, emergency medical services personnel, or lay bystanders. Prompt and accurate recognition of sudden cardiac arrest, a comprehensive and rehearsed emergency action plan, early cardiopulmonary resuscitation, and immediate access to and use of an automated external defibrillator are each pivotal links in the chain of survival. This review summarises the components of an effective emergency action plan, highlights the critical role of automated external defibrillators, and reviews the diagnosis and management of sudden cardiac arrest on the field of play.     

Closed-Chest Massage: An Effective Resuscitative Procedure for the Maintenance of the Circulation During Cardiac Arrest

Reports of clinical benefits of closed-chest cardiac resuscitation refute recently published studies contesting its effectiveness.Our experimental investigations demonstrate that closed-chest massage is able to achieve adequate cerebral circulation and oxygenation during cardiac arrest.Clinical studies indicate that a significant number of coronary patients can be saved if a monitor system is used to warn of the onset of ventricular fibrillation or arrest. To be successful, closed-chest resuscitation must be instituted within the four-minute limit after onset of cardiac arrest. When ventricular fibrillation occurs, electrical countershock is usually obligatory. Cardiac pacemaking is a specific necessity for patients with ventricular arrest.     

Brain Stem Death as the Vital Determinant for Resumption of Spontaneous Circulation after Cardiac Arrest in Rats

Background

Spontaneous circulation returns to less than half of adult cardiac arrest victims who received in-hospital resuscitation. One clue for this disheartening outcome arises from the prognosis that asystole invariably takes place, after a time lag, on diagnosis of brain stem death. The designation of brain stem death as the point of no return further suggests that permanent impairment of the brain stem cardiovascular regulatory machinery precedes death. It follows that a crucial determinant for successful revival of an arrested heart is that spontaneous circulation must resume before brain stem death commences. Here, we evaluated the hypothesis that maintained functional integrity of the rostral ventrolateral medulla (RVLM), a neural substrate that is intimately related to brain stem death and central circulatory regulation, holds the key to the vital time-window between cardiac arrest and resumption of spontaneous circulation.

Methodology/Principal Findings

An animal model of brain stem death employing the pesticide mevinphos as the experimental insult in Sprague-Dawley rats was used. Intravenous administration of lethal doses of mevinphos elicited an abrupt cardiac arrest, accompanied by elevated systemic arterial pressure and anoxia, augmented neuronal excitability and enhanced microvascular perfusion in RVLM. This period represents the vital time-window between cardiac arrest and resumption of spontaneous circulation in our experimental model. Animals with restored spontaneous circulation exhibited maintained neuronal functionality in RVLM beyond this critical time-window, alongside resumption of baseline tissue oxygen and enhancement of local blood flow. Intriguingly, animals that subsequently died manifested sustained anoxia, diminished local blood flow, depressed mitochondrial electron transport activities and reduced ATP production, leading to necrotic cell death in RVLM. That amelioration of mitochondrial dysfunction and bioenergetic failure in RVLM by coenzyme Q10, the mobile electron carrier in mitochondrial respiratory chain, or oxygenation restored spontaneous circulation further established a causal relationship between functionality of RVLM and resumed spontaneous circulation after cardiac arrest.

Conclusions/Significance

We conclude that whereas necrotic cell death because of bioenergetic failure triggered by anoxia in RVLM, which precipitates brain stem death, negates resuscitation of an arrested heart, maintained functional integrity of this neural substrate holds the key to resumption of spontaneous circulation after cardiac arrest in rats.     

Management of Cardiac Arrest: A Discussion of Cardiorespiratory Resuscitation

In cases of sudden cardiac arrest the time limit during which anoxic brain damage can be reversed is about four minutes. Therefore, cardiorespiratory resuscitation must be instituted immediately. A simple plan requiring no complicated maneuvers or equipment should be memorized and employed in such cases. The following program is recommended: (1) Institute artificial ventilation. (2) Thump the chest once or twice. (3) Employ closed chest compression. (4) Transfer the patient to hospital if survival appears possible. (5) Obtain an electrocardiogram immediately. (6) Apply specific treatment for cardiac standstill or ventricular fibrillation. (7) Reassess after 30 minutes. Discontinue treatment if there has been no response; if the pupils have remained dilated, fixed and unresponsive to light for over 10 minutes; or if it has been ascertained that the patient has definite serious illness incompatible with continuing life. Such a plan as outlined above should be instituted at once, before attempting to determine the cause of cardiac arrest or planning further special treatment.     

Mitotic death: a mechanism of survival? A review

Mitotic death is a delayed response of p53 mutant tumours that are resistant to genotoxic damage. Questions surround why this response is so delayed and how its mechanisms serve a survival function. After uncoupling apoptosis from G1 and S phase arrests and adapting these checkpoints, p53 mutated tumour cells arrive at the G2 compartment where decisions regarding survival and death are made. Missed or insufficient DNA repair in G1 and S phases after severe genotoxic damage results in cells arriving in G2 with an accumulation of point mutations and chromosome breaks. Double strand breaks can be repaired by homologous recombination during G2 arrest. However, cells with excessive chromosome lesions either directly bypass the G2/M checkpoint, starting endocycles from G2 arrest, or are subsequently detected by the spindle checkpoint and present with the features of mitotic death. These complex features include apoptosis from metaphase and mitosis restitution, the latter of which can also facilitate transient endocycles, producing endopolyploid cells. The ability of cells to initiate endocycles during G2 arrest and mitosis restitution most likely reflects their similar molecular environments, with down-regulated mitosis promoting factor activity. Resulting endocycling cells have the ability to repair damaged DNA, and although mostly reproductively dead, in some cases give rise to mitotic progeny. We conclude that the features of mitotic death do not simply represent aberrations of dying cells but are indicative of a switch to amitotic modes of cell survival that may provide additional mechanisms of genotoxic resistance.     

Serum Neuron Specific Enolase and Malondialdehyde in Patients After Out-Of-Hospital Cardiac Arrest

Background Sudden cardiac arrest (CA) is a leading cause of death in Europe. The victims of CA need immediate cardiopulmonary resuscitation (CPR). Patients resuscitated due to CA have high mortality rate. Prognostic evaluation based on clinical observation is uncertain and would benefit from the use of biochemical markers of hypoxic brain damage. Multiple factors of brain origin can be measured in blood. Objective The purpose of this study was to validate the use of the serum neuron-specific enolase (NSE) and malondialdehyde (MDA) concentrations for predicting in-hospital death, after resuscitation from out-of-hospital CA. Neuronal damage and impairment of the blood–brain-barrier integrity can be detected by the release of NSE into cerebrospinal fluid and eventually into the blood. MDA represents a product of lipid peroxide decomposition reactions. Methods In a prospective study of 35 consecutive survivors of out-of-hospital CA, serum samples were obtained within 24, 72 and 168 h after the CA. NSE and MDA concentrations were measured, relationship between concentration in group of in-hospital death and survived patients were examined. Results There was a significant difference in NSE concentration between survivors and dead group on 1st day of measurement, marginally significant difference on 3rd day and no statistically significant difference in NSE on 7th day of measurement. There was marginally significant difference in MDA levels in both groups in all days of measurements. Conclusion Estimation serum concentrations of NSE but not MDA seems to be a predictor of fate of patients after CA. The exact nature of oxidative stress can only be resolved by further studies.     

Heart Damage Associated with Intracranial Lesions

Focal myocytolysis, a form of myocardial damage, has been found to occur in about 8% of patients dying of intracranial lesions. Electrocardiographic abnormalities in patients with brain damage may be due to this. The cause of focal myocytolysis remains unknown, but if it could be prevented it might avert the patient''s death from cardiac arrest or arrhythmia. Moreover, more than a minimal degree of this type of damage to the heart might make it unsuitable for transplantation.     

Dual Inhibition of PI3K/Akt Signaling and the DNA Damage Checkpoint in p53-Deficient Cells with Strong Survival Signaling: Implications for Cancer Therapy

Natural (intrinsic) resistance of many tumor types to DNA damaging agents is closely associated with their capacity to undergo robust cell cycle arrest in G2/M. G2 arrest is regulated by the DNA damage checkpoint and by survival signaling, with a potential role of PI3K/Akt in checkpoint function. In this work, we wanted to clarify if inhibition of multiple checkpoint/survival pathways may confer better efficacy in the potentiation of genotoxic agents compared to inhibition of either pathway alone. We compared the influence of UCN-01, which affects both the DNA damage checkpoint and PI3K/Akt-mediated survival signaling, with the PI3K inhibitors wortmannin and LY294002 in p53-deficient M1 acute myeloid leukemia cells treated with the DNA damaging agent cisplatin. Our results show that direct inhibition of PI3K/Akt in G2-arrested cells by wortmannin or LY294002 strongly enhanced the cytotoxicity of cisplatin without influencing the G2 checkpoint. Unexpectedly, dual inhibition of both survival and checkpoint signaling by UCN-01, also increased the cytotoxicity of cisplatin, but to a lesser degree than wortmannin or LY294002. The differences in cytotoxicity were accompanied by differences in cell death pathways: direct inhibition of PI3K/Akt was accompanied by rapid apoptotic cell death during G2, whereas cells underwent mitotic transit and cell division followed by cell death during G1 when both checkpoint and survival signaling were inhibited. Our results elucidate a novel function for PI3K/Akt as a survival factor during DNA damage-induced G2 arrest and could have important pharmacological consequences for the application of response modulators in p53-deficient tumors with strong survival signaling.     

Survival of 1476 patients initially resuscitated from out of hospital cardiac arrest.

OBJECTIVES--To determine the short and long term outcome of patients admitted to hospital after initially successful resuscitation from cardiac arrest out of hospital. DESIGN--Review of ambulance and hospital records. Follow up of mortality by "flagging" with the registrar general. Cox proportional hazards analysis of predictors of mortality in patients discharged alive from hospital. SETTING--Scottish Ambulance Service and acute hospitals throughout Scotland. SUBJECTS--1476 patients admitted to a hospital ward, of whom 680 (46%) were discharged alive. MAIN OUTCOME MEASURES--Survival to hospital discharge, neurological status at discharge, time to death, and cause of death after discharge. RESULTS--The median duration of hospital stay was 10 days (interquartile range 8-15) in patients discharged alive and 1 (1-4) day in those dying in hospital. Neurological status at discharge in survivors was normal or mildly impaired in 605 (89%), moderately impaired in 58 (8.5%), and severely impaired in 13 (2%); one patient was comatose. Direct discharge to home occurred in 622 (91%) cases. The 680 discharged survivors were followed up for a median of 25 (range 0-68) months. There were 176 deaths, of which 81 were sudden cardiac deaths, 55 were non-sudden cardiac deaths, and 40 were due to other causes. The product limit estimate of 4 year survival after discharge was 68%. The independent predictors of mortality on follow up were increased age, treatment for heart failure, and cardiac arrest not due to definite myocardial infarction. CONCLUSION--About 40% of initial survivors of resuscitation out of hospital are discharged home without major neurological disability. Patients at high risk of subsequent cardiac death can be identified and may benefit from further cardiological evaluation.