Resuscitated cardiac arrest (CA) is a frequent cause of admission in intensive care unit (ICU). Neurological state of postanoxic comatose patients can evolve either towards the absence of awakening or towards a more or less altered state of consciousness, ranging from the vegetative state to the full recovery of cognitive functions. Most of the deaths result from active withdrawal of life-sustaining treatment, based on poor neurological outcome prediction. Neurological prognostication needs therefore a multimodal approach based on reliable parameters, which should be easy to access and available at the early stage of care, in attempt to limit false poor outcome prognostication and help to not prolong futile active care in patient with severe post anoxic cerebral lesions. Nowadays the prediction of neurological outcome relies on a multimodal strategy including clinical examination, biomarkers and electroencephalography (Guidelines ESICM 2015). Early standard electroencephalography (EEG) is currently recommended and some features, notably absence of reactivity, status epilepticus or burst suppression after rewarming are strongly predictive of poor outcome. But those features require a precise analyze of the EEG usually performed by specialist. EEG patterns can be simplified and classified in five grades according to the Synek classification, ranging from dominant reactive alpha activity (grade 1) to isoelectric encephalogram (grade 5). Grade 1 and two are considering as good prognostic, grade 3 as intermediate and grade 4 to five as poor prognostic. Nevertheless, few data are available on the performance of this classification since generalization of TTM use.

We hypothesize that a multimodal strategy combining clinical examination, NSE concentration and the Synek score would bring a high degree of prediction. We aimed to assess the performances of the combination of clinical examination, NSE analysed at 48-72h and the Synek score to predict hard outcome defined by a cerebral performance category (CPC) 3-5, in postanoxic comatose patients treated with induced hypothermia

Inclusion Criteria:

• patients admitted in our ICU following a resuscitate cardiac arrest between November 1, 2013 to November 1, 2017
• treated with 33° induced hypothermia for 24 h
• having at least one early EEG rated according to Synek score
• available dosage of NSE performed within 48-72h post CA

Exclusion Criteria:

• moribund patients
• awoke within the first 48 hours of admission
• neurologic causes of CA
• patients with missing data

Age ≥ 18 years with inclusion criteria : Will be considered for inclusion patients admitted in our ICU following a resuscitate cardiac arrest between November 1, 2013 to November 1, 2017 , treated with 33° induced hypothermia for 24 h, having at least one early EEG rated according to Synek score and an available dosage of NSE performed within 48-72h post CA.

exclusion criteria: moribund patients, awoke within the first 48 hours of admission, neurologic causes of CA, and patients with missing data