Cardiac arrest (CA) is a major public health problem (1) because of its frequency (30000 to 50000 new cases of CA per year in France), a high mortality 80-90% and a relevant morbidity : 50% of survivors have cerebral sequela. Neurologic prognostication after CA is paramount. The American Academy of Neurology identifies accurate predictors of poor neurologic outcome (2): clinical examination findings such as no pupil response, no corneal reflexes, no motor response to stimulation and early myoclonus status epilepticus; biologic parameters like high neuron specific enolase (NSE) greater than 33 μg/L (3-4) and electrophysiologic results like flat electroencephalogram (EEG) or with burst suppression and Somatosensory evoked potentials (SEP) with no response to N20. Brain imaging may provide additional informations but not recommended yet. However, these pronotic markers were developed before the introduction therapeutic hypothermia (TH) (5-6). TH is now recommended after cardiac arrest by international guidelines, based on demonstration of improved survival and neurological recovery in comatose survivors of CA (7). TH and its associated use of sedative and paralytic agent may delay neurologic recovery and affect the optimal timing of prognostic variables (8).

Many classifications of electroencephalogram (EEG) (9-12) exist but no one is generally accept and recommended and they were created before TH.

In a preliminary study based on a retrospective cohort of 64 patients in CA with initial TH, we developped an electroencephalographic score to predict precocity the neurologic outcome according to the Cerebral performance category (CPC) (13-14).

The purpose of this study is to evaluate a simple and objective electroencephalographic score helpful to predict neurologic outcome after CA. Then, we would like to create a composite score to have a multimodal prognostication.

Patients and methods After CA resuscitation, all patients underwent coronary angioplasty if indicated and then were immediately admitted in emergency room of ICU. All patients were intubated and mechanically ventilated. Sedation was performed with continuous infusion of midazolam 0,15-0,2 mg/kg/h and sufentanil 0,15-0,2 μg/kg/h. All our patients were curarized with cisatracurium 0,2-0,3 mg/kg/h to prevent shivering during TH and took anti-epileptic medication phenobarbital 10-15 mg/ kg two times a day to prevent infra-clinic convulsive seizure.

Mild TH between 33 -34 °C was performed by active cooling with intravascular device (CoolGard® - Zoll, Chelmsford, UK) for 24 hours. Mean arterial pressure was maintained between [75-80] mmHg by titrated norepinephrine with or without inotropics (dobutamine, epinephrine) according cardiac output monitored by echocardiography.

After 24 hours of hypothermia, warming was started at a rate of 0.2 ° C/hr; neuromuscular blocking agents were arrested at 35.5°C and sedation at 36.5°C.

Neurological evaluation performed at 48 hours off sedation included a clinical evaluation (Glasgow score, myoclonus status, response to orders, motor response to stimulation, brain-stem reflexes). NSE was sampled between 48 and 72 H. EEGs were performed at Day-2, Day-3 and Day-5 off sedation and SEP at Day-5 off sedation. All EEGs were interpreted by two certified electroencephalographers blind from clinical neurologic outcome.

Survival and neurologic outcome according to CPC were assessed at ICU discharge and three months later.

Electroencephalographic score

Reactivity Absence 1 Presence 0 Dominant frequency α (8-12 Hz) 0 θ (4-7,5 Hz) 1 δ (< 4 Hz) 2 Amplitude > 10 μV 0 < 10 μV 1 GPDs Absence 0 Presence 1 Burst Suppression Absence 0 Presence 1 Paroxysmal activity Absence 0 Presence 1

TOTAL

Inclusion Criteria:

• Patient > 18 years, hospitalized in resuscitation
• Cardiac arrest in hospital or out
• Collection of the written and informed consent

Exclusion Criteria:

• Cardiorespiratory arrest concomitant with a neurological, traumatic or toxic pathology
• History of cerebrovascular accident or cerebral anoxia
• Patient whose family did not give an informed consent
• Patient < 18 years